E-ITA.org Recovery Plan

05 Context and How to Read This Plan

This is the plan: 15 steps in 5 phases, and the checks that show whether they're working. The evidence behind each step is in the companion report, E-ITA.org Methodology and Findings (Parts 1 to 5).

What happened, in 5 points

• Search Console clicks fell 88.8%. They averaged 275.5 a day before the drop (2026-07-10 to 08-17) and 30.9 a day after it (2026-08-21 to 09-13) (Part 1.2).
• On 2026-08-21, the final day of the August 2026 spam update, e-ita.org fell from 205 clicks to 39 in one day. Impressions fell from 15,145 to 854 on the same day. Position fell from 9.1 to 18.9 (Part 1.3).
• All 8 top markets and every major content page class broke on the same date. The home page had already dropped once, in early July. Brand rankings held at positions 1.0 to 1.6. The site stayed indexed, so this was an automated demotion with site-wide reach, not deindexing (Parts 1.4, 1.6, and 1.9).
• This collapse was the second break. The first came with the June 2025 core update. Site clicks fell 28.6% across that update, from 500.5 to 357.5 a day. The loss sat mainly on the country pages (Part 2.3d). None of it came back.
• Rank 1, the trigger, is an automated demotion across the whole site. Its timing and breadth are confirmed. A spam classifier is the probable cause. Rank 2 is the mass re-ship of near-identical pages from 2026-07-03 to 07-10. It is probably the page set the update caught. Rank 3 is a thin trust layer on a YMYL topic, which likely caps recovery. Ranks 4 and 5 are standing risks that did not trigger this break. Rank 4 is the alias domains plus one noindexed clone. Rank 5 is link spam (Part 2.5).

Why not roll back

Should the site roll back to an earlier version? No. The archive snapshots from 2024 to 2026 cover the country pages up to 2026-07-09. They show that every earlier version carried the same page template and the same claims. The release history (from 2026-03-28) shows the same for every version since then. Rolling back would not remove what Google demoted. The phased rebuild below is built to remove it. Older code would also reopen security holes that have since been patched: the release log records Statamic security updates on 2026-06-27 and 2026-09-12 (Part 2).

The plan at a glance

Your first 14 days

These steps come from Phase 0 and the start of Phase 1. Each line names who does it and where the detail sits. The 14 days are a recommended order, not a deadline. Start each group of steps when the one before it is done.

First: Phase 0

1. Make the Phase 0 decisions. They cover the country count, the fee line, the headline, the refund rule, and the rest of the list. Who: E-ITA's leadership. (Part 6, top of Phase 0)
2. Freeze the page set. Publish no new country or locale page that the plan does not name. Informational or support pages that fix messaging, such as a guarantees page, are fine. Why: more pages from the same template add to the pattern the August update probably caught (Part 2.4). Who: E-ITA's leadership. (Part 6, Phase 0)
3. Ship the claims fix. Keep a country number E-ITA can back, and drop the words “accept”, “accepted” and “valid”. Publish the exclusion list. Keep the 8-minute line with the condition from the Terms. Say what the higher-fee urgent option adds (Terms, section 6). Make the same edits on the order form at apply.e-ita.org. It serves 2 variants at random, a one-page form and a 5-step form. The AAA/AATA block is optional, for US-focused pages only. The copy pack has the exact text. Why: on 2026-09-29 the homepage contradicted the site's own FAQ, and Perplexity listed the "175+ countries" claim as a warning sign (Part 3.3, Part 5.1). Who: the marketing team writes, the developer ships, and E-ITA's leadership approves. (Phase 0, step 1)
4. Set the measurement rules. Count orders from E-ITA's own order dashboard, which already tracks them. No analytics change is needed. Log every deploy in the change log on the day it ships. Why: 86.5% of GA4 key events since the collapse are not orders, and dated deploys separate site changes from Google updates (Part 5.5). Who: the marketing team, with the developer on the change log. (Phase 0, step 2)
5. Ship 15 redirects now, and only those 15. The other 164 redirects wait: 129 until their target pages exist, and 35 until each locale hub has a row for that country. The 162 removals wait for E-ITA's review. Start with /international-drivers-license-uk/ to /united-kingdom/. Why: that old URL took 3,250 sessions this year, all landing on a 404 (Part 4.1). Who: the developer. (Phase 0, step 3)
6. Review the disavow list by hand, if time allows. It is optional insurance. Upload only the entries that pass. The list holds no university (.edu) domain, and none should be added yet. Who: E-ITA's leadership, or the developer if given owner access to Search Console. (Phase 0, step 4)

Next: Phase 1 begins

1. Name the authors and reviewers. No rewrite ships without them. Why: no page names a person, on a topic where the rater guidelines check who stands behind the advice (Part 3.5). Who: E-ITA's leadership. (Phase 1, step 1)
2. Publish an About page and a real contact page. /about-us/ returns 404 today. Who: the marketing team writes, and the developer publishes. (Phase 1, step 3)
3. Start the reviews program. Confirm the Trustpilot plan first, then turn on automatic invitations. Who: E-ITA's leadership and the marketing team. (Phase 1, step 2)
4. Start the redirect-target pages. Write the 5 regional hubs, the car-rental guide, and the country rows for /es/, /fr/, and /ar/. They unlock the other 164 redirects. Who: the marketing team writes, and the developer builds. (Phase 0, step 3)

Read Search Console movement only after Google marks the September 2026 spam update complete on its status page. It began on 2026-09-24 and can take up to two weeks. Until then, any movement reflects the site before these fixes. The 410 removals start after E-ITA reviews the fold and kill lists. The sitemap, hreflang, and internal-link changes ship in the same deploy.

How to read this document

• Part 6 is the implementation sequence. Work the phases in order. Each step lists its actions, an acceptance check, a finding cross-ref, and the role that does it.
• Part 7 is the measurement framework: 6 readiness gates, the weekly and monthly checks, and the decision rules that declare or deny recovery.
• Finding cross-refs to Parts 1 to 5 point to the companion document, E-ITA.org Methodology and Findings. The verdict sheet and redirect map ship as their own pages and files, because the steps work from them.
• Every threshold in Part 7 was set before any recovery data existed. Do not move them after the fact.

PART 6 Recovery Plan

This part is the implementation guide. It has 15 steps across 4 working phases, plus an ongoing measurement phase. Each step lists its actions, the finding it answers, an acceptance check, and the role that does it. The roles are the ones E-ITA already has:

• E-ITA's leadership decides and approves.
• The developer ships changes through the site's code repository.
• The marketing team writes and edits copy and runs the checks.
• Support handles customer contact.

Work the phases in order. The weeks and dates in this plan are a recommended order, not deadlines. Start each stage when the one before it is done. Small, reversible fixes ship first. Google's September 2026 spam update began 2026-09-24, before any of these fixes shipped, so any change it brings reflects the site as it stands today. The plan is sized for the next named spam or core update after the redirect map is fully working (2026-11-15). The best case needs the 70 rewrites live by early December (Part 7). The 410 removals ship after E-ITA reviews the fold and kill lists, together with the sitemap, hreflang, and internal-link changes that go with them. Day 1 is the day the Phase 0 claims fix goes live. Record that date in the change log. Phase windows count from it. The dated targets in Part 7 are the thresholds the scorecard judges.

Phase 0: First (measurement and safety)

Do these 4 steps first: the later phases build on the low-risk steps 1 to 3, and step 4 is optional. While they run, freeze the page set. Publish no new country or locale page that this plan does not name, and run no bulk automated rewrites. Informational or support pages that fix messaging, such as a guarantees page, are fine. Why: the July re-ship of about 255 near-identical pages is probably the page set the August update caught (Part 2.4). More pages from the same template add to that pattern while the fix is under way. Google's spam policies define scaled content abuse as "many pages are generated for the primary purpose of manipulating search rankings and not helping users". The plan's own new pages are the exception: the 6 redirect targets in step 3, the Phase 1 trust pages, and the gated Phase 2 pages.

Phase 0 decisions for E-ITA's leadership. Make each one before the step that uses it ships.

1. The country count [N]: the distinct destination countries in paid orders, the window, and the date. If the order records hold no destination field, use the live count of country guides.
2. The urgent fee line. The Terms and Conditions, section 6, list a separate, higher fee for urgent processing than for standard (read 2026-10-05). Show both fees, or say "costs extra" and link to them, and say what the urgent option adds.
3. The homepage headline: a translation headline, or the speed headline with the condition from the Terms under it.
4. The refund promise. The order form, the Refund Policy, the Terms, and /help/ stated 4 different rules on 2026-10-05. Pick 1.
5. The "100,000+" line: the customer count from the order dashboard, or remove the line.
6. Any rental company with a written agreement, or drop "preferred provider".
7. Confirm which Cloudways stack runs the site before step 3. The Hybrid stack reads .htaccess, and the Lightning stack does not.
8. Who translates the new copy into Spanish, French, Italian, and Arabic.
9. The date the /go/ ad campaign stopped, for the change log.
10. The Trustpilot plan, paid or free. It decides whether the site may show a rating widget.
11. For Phase 1: the named author and reviewer, the street address, and a support contact.
12. The fold and kill lists in the verdict sheet. Flag any page the business needs before its stage ships. The 162 removals start after this review.

1. Correct the contradicting claims. Keep a country number E-ITA can back, and change the words around it. Why: on 2026-09-29 the homepage said the booklet is accepted in Japan, and the site's own /faq/ said it is not (Part 3.3). By 2026-10-05, E-ITA had replaced "175+" with "Used by drivers in 190 countries" on 207 of the 217 pages checked on 2026-10-05 (the 213 pages in the XML sitemap and 4 more) and dropped Japan from the homepage answer. The rest of this step is the work that was still live on 2026-10-05. The number can stay if E-ITA can back it. The words "accept", "accepted" and "valid" go. Use the Phase 0 copy pack below for the exact new text of every line. Some lines sit in shared templates and some in each page's own content, so search the whole site repository, not only the templates:

   git grep -n -i -E "17[05] ?\+|valid (in|worldwide)|accept(ed|s)? in|recogni[sz]ed|instant|1949|united nations|preferred idl|[0-9]+ ?minut|interchangeable|not valid|the permit|driver'?s permit|100,000|accept|worldwide|[0-9]{3} ?countries|\bIDP\b" -- .

   Run the same search again after the edit. The only expected hits are the exclusion list, the delivery times that carry the Terms condition, the 2-minute application lines, facts about the official IDP, the country-count line E-ITA can back, and a "100,000+" line the order dashboard backs. Then work the list:
   • Take "accept", "accepted" and "valid" out of every country line. On 2026-10-05 these lines were still live:
     • the FAQ question "Which countries accept the IDL?" on the homepage, /faq/, and the pricing page (also in the FAQ structured data). The homepage answer is already fixed: "The IDL is used by drivers in 190 countries, including the USA, UK, Australia, and most of Europe."
     • the link "Check countries that accept the International Driver's License" on the homepage, the apply page, and the country pages
     • the /countries/ heading "List of Countries That Accept an International Drivers License" and its checker line "Ready to check if an IDL is accepted in your destination?"
     • the eligibility checker on the homepage, the apply page, and the country pages. For every destination except Japan, South Korea, and mainland China, it says "[country] accepts the International Driver's License". The text sits in the widget's JSON, so a search of the visible text doesn't find it.
     • the destination checker on /countries/. For every destination, it answers "Yes, you'll need an International Driving Permit" and then offers the booklet at its starting price. The tool must never pair an IDP answer with an offer for the booklet. Why: a buyer told they need an IDP, then shown a price, will read the booklet as that IDP (Part 4.3).
     • the /italy/ FAQ question "Is my International Driver's License valid in Italy?", answered "Yes"
     On 2026-09-14 the "175+" line was on 362 pages. On 2026-10-05 it was on none of the 217 pages checked.
   • Make the same edits on the order form at apply.e-ita.org/package/step-1. It is a separate app, so a template edit doesn't reach it. On 2026-10-05, it still says "Official physical IDP card with security features", "Accepted worldwide at borders and checkpoints", "official IDP translation, instant digital copy, and worldwide acceptance", and "full refund if it's ever declined". The form serves 2 variants at random, a one-page form and a 5-step form, so check the copy of both.
   • Make the same edits in the 4 locales the plan keeps. There, the claims are in the local language, so the English search strings don't find them. By 2026-10-05 the count lines read "Usado por conductores en 190 países" (/es/), "Utilisé par des conducteurs dans 190 pays" (/fr/), "Usata dai conducenti in 190 paesi" (/it/), and "يستخدمها السائقون في 190 دولة" (/ar/), so the "valid" and "accepted" count lines are gone. Still live on 2026-10-05:
     • the FAQ questions "¿En qué países se acepta el PCI?" (/es/), "Dans quels pays le PCI est-il accepté ?" (/fr/), "In quali paesi è accettata la PDI?" (/it/), and "في أي الدول تُقبل الرخصة الدولية؟" (/ar/)
     • the links "Consulta los países que aceptan la Licencia de Conducir Internacional" (/es/) and "Voir les pays qui acceptent le Permis de Conduire International" (/fr/), the same link in /it/, and "اطلع على الدول التي تقبل رخصة القيادة الدولية" (/ar/)
     • the 1949 badge in each locale, for example "Cumple el formato de la Convención de la ONU de 1949" (/es/)
     A native speaker checks each translated line. In the 7 locales the verdict sheet retires, delete the claim lines. Don't translate them.
   • /faq/ says "The Permit is NOT valid only in Mainland China, South Korea, and Japan", and /countries/ says the materials "are not accepted in China, South Korea, and Japan". The U.S. Embassy in Italy tells US drivers to carry an IDP or an official Italian translation, which the booklet is not.
   • Lead with what E-ITA controls, the 12 languages in the booklet, next to the exclusion list. Two numbers work instead. The first is a live count of the country guides ("Driving guides for [N] countries"). The second is a count from E-ITA's own order records, worded "used by travelers in [N] countries", not "valid in". Count the distinct destination countries in paid orders over a stated window, then remove Japan, South Korea, and mainland China. The exclusion list carries the legal limits, and so does the AAA/AATA block where E-ITA uses it. The live line on 2026-10-05, "Used by drivers in 190 countries", follows the second wording. Log the order count, the window, and the date that back 190 in the change log. /countries/ links 196 country guides, so the guide count is a different number.
   • Use one number everywhere. On 2026-10-05 the site did: the pricing page and the rest of the site both said 190 (on 2026-09-29 the pricing page said "170+" and the rest "175+"). Keep the meta descriptions, social descriptions, and FAQ structured data on the same number. The acceptance check reads the page source.
   • In the same pass, replace the lines that imply official status. The FIA says "'International Driving Licences' (IDLs) ... as sold by some internet sites, are not legally recognised and cannot replace an IDP" (FIA, Travel and Tourism). By 2026-10-05, E-ITA had removed "Valid Worldwide", "Instant Approval", and "Globally Recognized & Meets International Standards". The lines still to replace are these:
     • "Follows the 1949 UN Convention booklet format"
     • "We abide by the International Driver's License Booklet specifications detailed in the UN Convention of 1949"
     • the /countries/ line that says the booklet "follows the United Nations guidelines as outlined in the Convention on International Road Traffic of September 19, 1949"
     • "Is the IDL Recognized Worldwide? Yes." and "We are a preferred IDL provider for many rental agencies" (/faq/ and the pricing page), unless E-ITA can name the agreements
     • "100,000+ Travelers Trust E-ITA" and "Trusted by 100,000+", unless the order dashboard shows at least that many customers. If they do, keep the line and log the count and date.
   • Make the refund promise say the same thing everywhere. On 2026-10-05, 4 pages stated 4 rules:
     • Order form: "full refund if it's ever declined".
     • Refund Policy: the application fee is "fully refundable if you attempted to use your documents and were unsuccessful".
     • Terms, section 9: application fees are refundable "if we attempted to produce your documents and were unsuccessful due to issues on our end".
     • /help/: the application fee is "fully refundable if you tried to use your documents and they were not accepted".
     Why: support can apply only one rule. A buyer held to a narrower rule than the checkout promised is a likely complaint. E-ITA's leadership picks 1. Use it word for word on all 4 pages and next to every "Money-Back Guarantee" line. Drop "100%" from those lines. Once documents are processed and shipped, both policies refund only the application fee.
   • Publish 1 exclusion list, word for word the same on every page, in the terms, and in every locale: "The E-ITA booklet is not accepted in Japan, South Korea, or mainland China." On the homepage, /faq/, and /countries/, add: "Other countries set their own rules, so check your destination before you travel." Don't say "only", because other countries, such as Italy for US drivers, also ask for an IDP or an official translation. On 2026-10-05, 5 wordings were live: on /faq/ and the pricing page, on /countries/ (the note and the checker), in the Terms, and on /help/.
   • Make every delivery-time line match the Terms. On 2026-10-05, the Terms (section 8) said digital documents are available "within 8 minutes of placing your application, once your payment is approved and your application is complete and valid", for every order. No page stated a separate standard time any more. The Terms (section 6) still list a higher fee for urgent processing, and /help/ says urgent processing "moves your order to the front of the queue". So the 8-minute line can stay. Wherever it appears, add the condition from the Terms (payment approved, application complete), and say on the same page what the higher-fee urgent option adds. Use 1 name for the urgent option, and drop "instant" and "immediate". Why: the Terms (section 8) promise 8 minutes to every complete order. Section 6 lists a higher fee for urgent processing. So a buyer who reads the 8-minute line needs to see what that fee adds. The FTC tells advertisers to "Place the disclosure as close as possible to the triggering claim" (FTC, .com Disclosures). For public proof, record a real order from submission to the delivered copy. Put the short screen recording on the FAQ, next to the time. No wider data analysis is needed. Rebuilt country pages carry no delivery-time claim (Differentiation Standard, section 13).
   • The 11 /go/ ad landing pages are retired. E-ITA redirected them to the homepage, and that was the right call. They were ad pages selling the same product as the homepage, so the homepage is their closest page. (The step 3 rule against homepage redirects applies to content pages.) They carried the same "175+" claim and a "10 minutes" delivery claim. They took 16% of the site's sessions this year (GA4 landing-page report, 2026-01-01 to 2026-09-26). On 2026-10-05, every /go/ path, including /go/us.html, returned 301 to https://e-ita.org/ in 1 https hop. Log the retirement date in the change log (step 2), so the session drop doesn't read as an organic change.
   • The testimonial block is already off the site. To show the Trustpilot reviews again, use a Trustpilot live widget or exact quotes. Both need a paid Trustpilot plan. A rating widget (TrustScore, stars, and review count) is the simplest. A widget that shows review text also shows claims E-ITA can't make itself, so it belongs on /reviews only. A quoted review needs its date, its star rating, and Trustpilot attribution with a link to the profile. It also needs either the reviewer's documented permission or a fully anonymized quote. Take the quote down if the reviewer edits or deletes the review. Drop any review that claims acceptance, official status, or a turnaround the buyer's processing option doesn't deliver. The 5-step order form still carries a hidden "Rated 4.9 out of 5 on Trustpilot" line (2026-10-05), so it comes down too. Trustpilot shows 4.6 stars (Part 4.2), and its rules require a current, accurate rating. Use a live widget or a plain text link instead.
   • Optional: publish the AAA/AATA disclosure block on US-focused pages only, such as /us/, as 1 shared component with identical wording. Most of E-ITA's clicks come from outside the US (the US share stayed between 14.8% and 19.0%, Part 1.6), and the block names only the US issuers, so pages for other markets skip it. Why it helps on US pages: the FTC's consumer alert tells US buyers that only AAA and AATA issue IDPs, and all 4 tested AI engines called the site not authorized when asked about it (Part 4.3, Part 5.1). Before they pay, the block tells US buyers what the booklet is: a translation, not an IDP. On each US page, place it under the H1. The text: “The E-ITA International Driver's License is a private translation of your license, carried with your license. It is not a driver's license, and no government issues it. It is not an International Driving Permit (IDP). If your destination requires an IDP, get it where your license was issued. For US licenses, only AAA and AATA are authorized by the U.S. Department of State to issue IDPs. For other licenses, the issuer is the body the issuing country authorizes, usually its licensing authority or automobile club.” Rebuilt country pages carry the same facts in their own “Where to get an official IDP” section (Differentiation Standard, section 3).
   • In the same pass, stop calling the product an “IDP”, an “International Driver's Permit”, or “the Permit” anywhere on the site. This includes the /italy/ line that calls IDL and IDP “interchangeable terms for the same booklet”.
   • Fix the FAQ structured data block on every page that carries it. That includes the homepage, /faq/, /help/, the pricing page, and the country pages (checked on /italy/, /spain/, /greece/, and /us/). On 2026-10-05 it began with template code (“<?php $__contextArgs”) where “@context” belongs, so it is not valid schema.org markup, which requires “@context”. Fix the template so the code doesn't print, or remove the block. Why: a template fault prints server code into the page source of the site's main pages.
   On 2026-09-29 the homepage contradicted the FAQ on the same question (Part 3.3), and on 2026-10-05 the exclusion wording still differed from page to page. An AI engine quotes the site against itself (Part 5.1). These claims didn't cause the August drop (Part 2). They are a trust problem that caps recovery.
   SQRG 5.6, p.63 “If you find the information about the website or the content creator to be exaggerated or mildly misleading, the Low rating should be used.”
   SQRG 4.5, p.35 “Multiple or significant factual inaccuracies on an informational page which would cause users to lose trust in the webpage as a reliable source of information”

   Compliance frame: an advertiser needs “a reasonable basis for advertising claims before they are disseminated” (FTC Policy Statement Regarding Advertising Substantiation). E-ITA's own order and customer records are a reasonable basis for a line such as “used by travelers in [N] countries”. They can't back “accepted” or “valid” where a country's official rule says otherwise, as in Japan and, for US drivers, Italy.

   An endorsement “may not convey any express or implied representation that would be deceptive if made directly by the advertiser” (16 CFR 255.1(a)). A reviews page may not imply it shows most or all reviews while it hides some for their rating or negative sentiment (16 CFR 465.7(b)). Trustpilot allows review quotes, star ratings, and widgets on paid plans only. Free plans may link to the profile in plain text (Trustpilot Legal Brand Guidelines, September 2026). The FTC consumer alert names AAA and AATA as the only authorized US issuers.

   Who: the marketing team rewrites the lines, and the block if E-ITA uses it. The developer ships them in one deploy and edits both order-form variants. E-ITA's leadership signs off on the country-count wording, the fee line, and the refund rule.

   Acceptance check: the repository search and a full crawl find no “accepted in” or “valid in” country claim, in any language. This covers page text, meta descriptions, social descriptions, structured data, and the eligibility checker's result text. The exclusion list's “not accepted in” sentence is the one expected hit. A GET of /, /international-drivers-license-apply-online/, /faq/, /countries/, /help/, /italy/, /es/, /fr/, /it/, /ar/, and apply.e-ita.org/package/step-1 confirms it. The source also shows no “10 minutes”, no “Valid Worldwide”, no “Instant Approval”, and no 1949 Convention claim, and neither order-form variant says “Accepted worldwide” or “worldwide acceptance”. Every “8 min” line carries the condition from the Terms. Each page with an 8-minute line also says once what the urgent option adds (copy pack S3). Every known /go/ URL, including /go/us.html, reaches its target in 1 https hop. The exclusion list has identical wording wherever it appears, and the AAA/AATA block does too if E-ITA uses it. The order form shows no typed-in Trustpilot score. The refund rule reads the same on the order form, the Refund Policy, the Terms, and /help/. No page calls the product an IDP, a permit, or interchangeable with one. "100,000+" and "preferred IDL provider" are gone, or the change log records the count or the agreement behind them.

   Finding cross-ref: Part 3.3, Part 4.3, and Part 5.1.

2. Set the measurement rules. Count orders from E-ITA's own order dashboard in every report. It already tracks every order, so no analytics change is needed. Don't report GA4's total "key events". Since the collapse, 86.5% of them are generate_lead, which is not an order. The total therefore hides what orders are doing (Part 5.5). Key events are still fine for one job: picking which pages to rewrite (the verdict rule's 20+ test).
   • Log every deploy in the change log on the day it ships. Give each row the date, step, scope, URLs or files touched, reason, expected signal, GET check result, and who checked it. Why: without a dated record, nobody can tell a drop caused by a deploy from one caused by a Google update (Part 5.5).
   • Treat GA4 figures from 2026-07-02 to 07-27 as void. Collection recorded zero traffic for those 26 days.
   • When a report uses GA4 channels, show orders from the visits GA4 labels "(not set)" as their own line. Those visits started on 2026-07-28 and carry 58.3% of orders since the collapse (Part 5.5). Mixed into other channels, they distort every order trend.

   When	Change-log title	Change-log note
   Claims fix deploy	P0 S1 claims fix: count, times, exclusions	New copy on /, /faq/, /countries/, pricing, order form, country template. No traffic change expected. See change log.
   The date the /go/ ads stopped	P0 S1 /go/ ads stopped	Paid landers now go to the homepage. Sessions and orders fall for a non-organic reason. Do not read as a search change.
   Redirect stage 1	P0 S3 redirects stage 1: 15 x 301	Old 404 URLs now 301 in 1 hop, including -uk/ to /united-kingdom/. Expect fewer 404 landings.
   Redirect stage 4	P0 S3 removals: 162 x 410	Kill URLs return 410. In the recommended order, clicks may dip until 2026-12-06 (Part 7 dip window). Below 15/day for 14 days: audit redirects and tracking.

   Who: the marketing team owns the reports. The developer writes a change-log row for every deploy.

   Acceptance check: every report takes its order count from the order dashboard, and any GA4 channel view shows "(not set)" orders as their own line. Every deploy has a same-day change-log row.

   Finding cross-ref: Part 5.5 and Part 7.

3. Ship 15 redirects now. The rest wait. Only stage 1 ships now: the 15 redirects below, which fix the leftover broken URLs. The other 326 rows ship in 3 later stages, each its own deploy. The 164 redirects ship once their target pages are live. The 35 locale rows among them ship once each hub has its country row. The 162 removals ship once E-ITA has reviewed the fold and kill lists. Why: 184 of 195 English country pages have a near-twin above 0.90 similarity (Part 3.2). Google keeps crawling those near-duplicates until they redirect or return 410 (Part 5.2). Stage 1 also stops visitors landing on dead pages today (Part 4.1).
   • Stage 1, now: ship the 15 redirects whose target page already exists and already covers the old page's topic. Start with the old URLs that still get traffic (about 100 sessions across the 15 in the last 28 days) and return 404. The biggest is /international-drivers-license-uk/, which should go to /united-kingdom/. It had 3,250 sessions and 252 key events in 2026 so far, all landing on a dead page.
   • In the same deploy, fix one post-purchase URL outside the map. https://e-ita.org/thank-you/ returns 404 and still got 51 sessions in the last 28 days (to 2026-10-03). The order app has no live thank-you page: https://order.e-ita.org/thank-you/ shows “404 Page Not Found” in a browser (checked 2026-10-05). All 51 sessions were direct visits, so they come from past buyers with an old link. It is not a search page (0 search clicks), so it is not in the redirect map. Pick one of 2 options:
     • Option A: add a 301 from the old URL to the help center, https://e-ita.org/help/. It answers order questions and links to the customer portal.
     • Option B: publish a real page at /thank-you/. The model thank-you page shows one, built from the Terms and the Help Center. Its notes cover the 3 sibling pages that also return 404.
   • Stage 2: hold the 35 redirects that fold translated country pages into /es/, /fr/, and /ar/. First add a short row for each folded country to that locale hub. Ship each redirect when its row is live. Without the row, Google can treat the redirect as a soft 404.
   • Stage 3: ship the 129 rows that point to the 6 new hub and guide pages below. Ship each row as its target page goes live.
   • Stage 4, after E-ITA reviews the fold and kill lists: ship the 162 rows that return 410. One German page already redirects to its German twin. Both are in stage 4, so delete that redirect when stage 4 ships. Ship them in the same deploy as the sitemap, hreflang, and internal-link changes. In that deploy, delete every link to these URLs. On 2026-10-01, the language switcher on every page linked to pages in the 7 retired locales. The /countries/ page linked to 41 of the 162. Don't hold them for another step. These pages had 4 clicks in total in the 23 days after the drop, at most 1 per page (Search Console, 2026-08-21 to 09-12).
   • Before any 410 ships, the marketing team exports the landing URLs of every live Meta ad, email template, and partner link. Give any mapped URL on that list a 301 to its closest live page, not a 410, or move the ad to a surviving page first. Also confirm /united-arab-emirates/ against the order dashboard. It had 1 search click in the 90 days before the break, and 11 in the 12 months to 2026-10-02.

   How to ship the map on this stack. Add each stage's rows to public/.htaccess in the site repository, above the static-cache rules, when that stage ships. Cloudflare Bulk Redirects can't return 410, so the finished file holds all 341 rows. First confirm that the server runs the Cloudways Hybrid stack, which reads .htaccess. If it runs Lightning, switch the application to Hybrid before stage 1. Lightning's Web Rules are meant for up to about 25 redirect rules. Above that, Cloudways advises the Hybrid stack and .htaccess. Each stage has its own CSV. Stages 1 and 2: redirects-stage-1-ship-now-15.csv and redirects-stage-2-after-es-fr-ar-hub-rows-35.csv. Stages 3 and 4: redirects-stage-3-after-6-new-pages-129.csv and redirects-stage-4-removals-after-list-review-162.csv. The Planned action column is the change to make. Two columns show what the old URL and the target URL return today (checked 2026-10-05). Four more columns show traffic to the old URL: search clicks and GA4 sessions. Each metric appears for 12 months (2025-10-01 to 2026-10-02) and for the last 28 days (2026-09-05 to 2026-10-02). A dead URL gets no search clicks. Its visits come from old links. In stage 1, a 404 on the old URL is expected: that dead page is why the redirect exists. All 15 stage 1 URLs returned 404 on 2026-10-05. In stage 2, the locale hubs already return 200. Each row still waits until its hub has a row for that country. In stage 3, a 404 on the target is expected until that new page goes live.

   • A 301 row reads RewriteRule ^international-drivers-license-uk/?$ https://e-ita.org/united-kingdom/ [R=301,L].
   • A 410 row reads RewriteRule ^monaco/?$ - [G]. Add ErrorDocument 410 /410.html, and build that page with links to the regional hubs.
   • Use the full https target on every row. The /?$ ending catches the URL with and without the slash in 1 hop.
   • Never let a rule match its own target. The /Italy/ row differs from its target /italy/ only by case, so match the exact case only. A case-insensitive rule would send /italy/ into itself. After each stage, GET every target once and confirm it returns 200, not 301.
   • Ship every stage as one deploy with 3 more changes. Remove each redirected or removed URL from the XML sitemap, including its hreflang alternates. In the stage 4 deploy, drop all hreflang links from the sitemap (Phase 2, step 4 removes them anyway). Point every internal link at the final URL, including the country list on /countries/. For a URL that returns 410, delete the link instead, including its entry in the language switcher. Drop each redirected or removed URL from the hreflang sets on the pages that list it.
   • Clear the static cache after each stage. Then run the checks with GET requests.

   Why the kills return 410, not 301. Sending a removed page to the homepage looks safe, but Google can treat a redirect to an unrelated page as a soft 404. Google's site-move guide warns against sending many old URLs "to one irrelevant single URL destination, such as the home page", which "might be treated as a soft 404 error". The homepage is irrelevant to a removed country page. It was relevant to the /go/ ad landers, which sold the same product. Also, visitors who click expecting a specific page land somewhere else. Google treats a 410 the same as a 404 (“All 4xx errors, except 429, are treated the same”, Google, HTTP status codes and Google’s crawlers). Either one takes the URL out of the index. The plan uses 410 because it says the removal is deliberate. A custom 410 page can still tell the visitor the page is retired and point them at the nearest hub. That is why the 162 kill rows return 410.

   • Build the 6 redirect-target pages first: /africa/, /europe/, /americas/, /asia-pacific/, /middle-east/, and /guide/international-driving-permit-car-rental-companies/. They return 404 today. No 301 row may ship before its target returns 200. Aim to get all 6 live by 2026-11-08, a week before the 2026-11-15 gate. Add all 35 country rows to /es/, /fr/, and /ar/ by the same date if possible.
   • Build each regional hub to one spec. Give every folded country 1 row with 3 items: whether a visiting driver needs an IDP, the road-traffic convention the country is party to, and a source link with its accessed date. Use whitelist sources only. Don't copy the old country page text. Make no country count, delivery-time, or acceptance claim for the booklet. Carry the exclusion list, link the surviving country pages in the region, and show a named reviewer and a last-verified date. The same row spec applies to the rows added to /es/, /fr/, and /ar/. The /americas/ hub prototype shows the spec applied to its 30 rows, with notes on how each row is made.
   • The car-rental companies guide replaces the 14 rental-brand pages the site removed. Give each company 1 row with its published rule on licenses, IDPs, and translations, a link to that policy page, and the accessed date. Say that the booklet is accepted only where the company's own policy says so.
   • Map each removed URL to its closest live equivalent. Never map a removed content page to the homepage. The /go/ ad landers in step 1 are the one exception, because they sold the same product as the homepage. Examples from the map:
     • /international-drivers-license-uk/ goes to /united-kingdom/ (a live page, so this row ships in stage 1).
     • /support/solutions/ goes to /help/, the live help center.
     • /belize/ goes to /americas/ and /lesotho/ to /africa/ once those hubs exist.
     • /monaco/ returns 410.
   • Fix the residue: /Italy/ and /Italy3/ must return 301 to /italy/.
   • The map now covers 12 pre-migration legacy rows. GA4's landing-page data shows them still taking traffic. They include the /international-drivers-license-* 404s. The -uk/ path alone took 3,250 sessions and 252 key events in 2026 YTD. They also include the old /support/ help-center paths. The /international-drivers-license-eu/ row ships only after /europe/ returns 200.
   • Keep https on every hop. The 18 historic /translations/ slugs are not in this map. E-ITA's leadership decides their fate before Phase 2 starts. There are 3 options: return 410 (this plan's preferred option), keep the current redirects, or restore the translations with a note on why they exist. The professor links point at translations the site hosted for them. Google's link spam policy lists “Exchanging goods or services for links” as an example, so a reviewer could read them that way. Decide the university (.edu) links at the same time. Until then, make every hop in the existing /translations/ redirects stay on https.
   • After each deploy, resubmit the XML sitemap in Search Console. Search Console limits single-URL requests, so request indexing only for the 10 highest-traffic target pages, such as /united-kingdom/. Don't request indexing for 410 URLs.
   • Script a GET check of every shipped row. Run it after every deploy and cache clear, and once a week. Why: a later deploy or a stale static file can undo a rule without warning (see "How changes go live on this stack"). The full map is green by the Part 7 gate target of 2026-11-15.
   The page-by-page verdict sorts 454 audited URLs into keep 73, rewrite 70, fold 149, kill 162 (Table B.1). "Fold" is the consolidate verdict: the page's content moves into a regional hub or locale hub, and its URL redirects there with a 301. "Kill" is the retire verdict: the page is removed and returns 410. Rewrite means a country page with at least one of these:
   • any paid orders (2026-06-01 to 2026-09-12)
   • 100+ pre-break clicks in 90 days
   • 20+ GA4 key events in 2026 YTD (all channels, 2026-01-01 to 2026-09-26)
   The residue URLs are verified and itemized in Part 3.2.

   Who: the developer ships the map and the 6 target pages. The developer also scripts the map check, because it reads raw status codes. The marketing team writes the hub pages and the locale hub rows, and reads the check results after each run.

   Acceptance check, per stage: every row in the stage returns its mapped status in 1 https hop, and every target returns 200. Stage 1 also shows /international-drivers-license-uk/ at 301 to /united-kingdom/ and /Italy3/ at 301 to /italy/. After each stage, the XML sitemap lists none of that stage's URLs. A crawl finds no internal link or hreflang tag that points at them. After stage 4, the full 341-row check returns the expected status on every row.

   No row targets the homepage.

   Finding cross-ref: Part 4.1, Part 5.2, and Part 3.2.

4. Review, then file, the disavow (optional insurance). The 822-domain disavow list is a separate file: disavow-draft.txt. Review it by hand, then upload only what passes the review. The list is built from the link-vendor and directory inventory: the .store and .shop networks, the fake directories, and the random-burst domains. It lists more domains than the ~600 that the link tool flags as spam. At least 222 entries were therefore not tool-flagged. Keep only clear link-vendor, network, and fake-directory domains. Google's disavow help warns that the tool can harm a site's performance if used incorrectly. It advises the tool mainly for links that caused a manual action or likely will. Search Console's Manual actions and Security issues reports were checked on 2026-09-24, and both read “No issues detected”. This means the disavow is prevention, not remediation. If the review takes longer, finish the other Phase 0 steps first. The list holds no .edu domain. Add none until the /translations/ decision (Part 4.1).
   • Before upload, confirm that every remaining entry is a reviewed link-vendor, network, or fake-directory domain. Confirm that no .edu domain is included.
   • Build the upload file from the reviewed rows. Use a .txt file in UTF-8, with 1 line per domain in the form domain:example.shop. Lines that start with # are comments. Google limits the file to 100,000 lines and 2 MB.
   • Upload through a URL-prefix property for https://e-ita.org/, because the disavow tool doesn't support Domain properties. If only a Domain property exists, add and verify the URL-prefix property first.
   • Log the upload date and the file hash in the change log.

   Who: E-ITA's leadership, or the developer if given owner access to Search Console.

   Acceptance check: if a file is filed, it holds only reviewed entries, every entry traces to the list, and it holds no .edu domain. If no file is filed, the change log says so.

   Finding cross-ref: Part 4.1.

Phase 0 copy pack: exact replacement text for step 1

This pack gives the exact new text for every claims line that step 1 changes. The live text was read on 2026-09-29 and checked again on 2026-10-05. The last column of table C gives the 2026-10-05 state, so rows E-ITA has already fixed are marked done. Where the site uses a long dash, this pack shows a hyphen. Fill every [N] before publishing, from the order records or the live count of country guides (see "How to set [N]"). Pick 1 of the 2 fee options and use it everywhere. Each row changes for a reason given in step 1. Rows 26 to 28 and 49 also drop "drive worldwide", "Avoid fines", and "authorities". Each implies that police accept the booklet, which the FIA quote in step 1 rules out.

A. The standard lines. Use them word for word.

How to set [N]: count the distinct destination countries in paid orders over a stated window. Remove Japan, South Korea, and mainland China. Log the count, the window, and the date. Use the same [N] on every page. If E-ITA picks the live count of country guides instead, S1 becomes "Driving guides for [N] countries". Then drop the "Travelers in [N] countries" sentence from A1 and rows 25 and 32, and use S1 in row 5. The fees in option A are the ones the Terms and Conditions list on 2026-09-29. E-ITA confirms them before publishing.

B. Two FAQ answers used on several pages

C. Page by page

D. Lines that stay

• “International Drivers License Online - Apply in Minutes | E-ITA” (homepage title) and “Complete your application in less than 2 minutes”. They describe the application, which takes about 2 minutes.
• “An International Driver's License is not a replacement for your domestic driver's license or a passport...”. It states a true limit.
• The 1-year, 2-year, and 3-year validity lines. They change with the product in Phase 2, step 2.
• Country FAQ lines about the official IDP, such as a country's own IDP rule. Phase 2 rewrites them with sources.

Phase 1: Days 1 to 30 (trust layer)

The trust layer answers who runs the site and why anyone should believe it. It ships before the content rebuild so Google's re-crawl of the cut page set lands on answered questions.

1. Build the authorship layer. No author or reviewer exists anywhere on the domain today. Work the list:
   • Assign a named author and a named reviewer to every editorial page, including the 73 keep pages.
   • Publish 1 bio page per person. State credentials relevant to driving documentation.
   • Acceptable credentials are travel law, automotive club work, rental operations, or documented travel in the country. E-ITA's leadership can be the named reviewer for facts about the product itself. These include what the booklet contains, how ordering and delivery work, and what customers report at rental desks.
   • Add Article markup with an author field. Emit datePublished and dateModified in ISO 8601 format. If you add reviewer markup, put the reviewer in a reviewedBy field on the page's WebPage node. Schema.org defines reviewedBy for WebPage, not for Article.
   • Until a named person is approved, keep the 73 keep pages on an "EITA editorial team" byline with a named reviewer role. Rewritten country pages wait for real names (Differentiation Standard, section 6).
   • Hold the gate: no rewrite ships without a visible author and reviewer byline and author markup. Readers and raters judge the byline. Google's Article markup page lists author but not reviewedBy, so reviewedBy is optional.
   Authorship scores 2.00 of 10 on the 18 e-ita.org pages of the full 7-criterion pass (Part 3.5). It scores 4.67 across the 445 live pages of the crawl-wide run.
   SQRG 4.5.1, p.36 “YMYL pages or websites that handle sensitive data with absolutely no information about the website or content creator should be rated Lowest.”
   Helpful content guidance “Is this content written or reviewed by an expert or enthusiast who demonstrably knows the topic well?”
   SQRG 3.3.4, p.25 “Educational degrees, peer validation, expert co-authors, and citations can be evidence of positive reputation information for professionals who publish their work.”
   Risk: never invent a profile. Fabricated author pages are a Lowest-rating trigger, not a shortcut.

   SQRG 4.5.3, pp.36-37 “A webpage or website with "fake" owner or content creator profiles. For example, AI generated content with made up "author" profiles (AI generated images or deceptive creator descriptions) in order to make it appear that the content is written by people.”

   Who: E-ITA's leadership names the authors and reviewers. The marketing team writes the bio pages. The developer adds the markup.

   Acceptance check: every rewritten page shows a named author and reviewer with working bio links. Every keep page shows at least the editorial-team byline, linked to the About page. The page markup parses with zero errors.

   Finding cross-ref: Part 3.5.

2. Start the reviews program. The review platform comparison covers the Trustpilot plans, cheaper Google store rating partners, and what happens to the existing reviews. Work the setup, then the reply engine:
   • First, confirm whether the Trustpilot account is on a paid plan. Only a paid plan may show ratings, stars, or quotes on the site. On a free plan, use a plain text link to the profile only (Trustpilot Legal Brand Guidelines, September 2026).
   • Re-check the live Trustpilot profile in a browser. Record score, review count, one-star share, and reply rate. Quote no archived figure before this check.
   • The last browser check (2026-09-22) read 4.6 stars and 1,563 reviews. Some tools still show 4.2 stars and 1,328 reviews from a stale February 2026 copy. That is why the live check comes first.
   • Turn on Trustpilot's automatic review invitations, set to send after the customer has the document, not at payment.
   • Invite every purchaser. Apply no sentiment filter anywhere in the flow. Check the plan's monthly invitation cap first. Trustpilot's pricing page lists 50 a month on Free, 100 on Starter, 300 on Plus, and 1,000 on Premium (read 2026-09-29). E-ITA's order dashboard shows the monthly order count. The Part 7.4 baseline, from GA4, is about 900 a month. If the cap is below the order count, invite by a fixed rule, such as the first orders each day. Never choose by expected rating.
   • Use 1 neutral template. Offer no incentive. Use no star-direction language.
   • Before going live, check the invitation flow with a few internal test orders. Keep a dated log of every invitation sent.
   • Reply to 100% of new negative reviews within 48 hours. Sweep the one-star backlog weekly until cleared.
   • Structure every reply the same way: acknowledge, state the honest product scope, and root-cause the specific claim. Then offer a concrete resolution path and move detail to support.
   • Pre-load the honest-scope paragraph into 3 templates: wrong expectation, delivery or document issue, police or rental refusal.
   Write 1 specific, professional reply to the review that quotes Italian police calling the booklet "a fraud". UNVERIFIED This review appears in the early-September AI-answer test, but no live copy was found on 2026-09-14. The standing reply program covers it either way. Post the reply once, only if the review is visible. Never argue in the thread. Offer no incentive for any review. Refunds to unhappy customers can continue, but never make a refund depend on editing or removing a review. The category leader, the International Drivers Association, replies to every visible review. E-ITA replies to about 5% of its negative ones (Part 4.2).

   Compliance frame: generalized solicitations to all purchasers fall outside the FTC rule's fake-review limits (16 CFR 465.2(d)(1), which excludes them from 465.2(b) and (c)). Sentiment-conditioned incentives are prohibited under 16 CFR 465.4. Review gating can be deceptive under 16 CFR 255.2(e) Example 11, when it makes posted reviews more positive than they would otherwise be. Trustpilot's own rule matches: invitations must be fair, neutral, and unbiased, with no incentives offered (Trustpilot Guidelines for Businesses, June 2026).

   Directional month-3 (Phase 2) targets: 100% of new negative reviews answered, review count up 90 to 150, and one-star share flat to down. About 172 of the 1,563 reviews are one-star (11%). Getting below 7% takes about 900 new reviews if none are one-star, and about 2,100 if 4% are. At the month-3 pace of 30 to 50 a month, even 900 takes 18 to 30 months. The claims fixes, which cut new one-star reviews, are the main lever.

   Who: support writes and posts the replies. The marketing team sets up the invitations. E-ITA's leadership confirms the Trustpilot plan.

   Acceptance check: automated invitations are live. A weekly check shows 100% of new negative reviews answered within 48 hours. If the police review is visible, it carries a visible, professional reply.

   The compliance log shows every send and every reply.

   Finding cross-ref: Part 4.2.

3. Publish the 2 brand-defense pages. Build both to this spec:
   • Publish /reviews as 1 canonical page. Carry dated, attributed, anonymized excerpts, both positive and negative, each with the company reply where one exists. Trustpilot excerpts follow the display rules in Phase 0, step 1, and the page applies no star filter.
   • Add a visible "how we collect reviews" note to /reviews. Prefer a live widget, which updates itself. If the page copies quotes, reflect any edit or deletion within 24 hours, as Trustpilot's Content Refresh Guidelines require. Link out to the live Trustpilot profile as the full record.
   • Publish /is-e-ita-legit. Answer in the first paragraph. Sign it with the entity name, the address, and a visible last-updated date.
   • Answer 4 things directly on /is-e-ita-legit. Say what EITA LLC is and what the product is and is not. Include the refund policy. For US visitors, add the AAA/AATA facts (the optional block).
   • Do not add review markup to either page. Google's review-snippet guidelines cover a business that "controls the reviews about itself". Its pages are "ineligible for star review feature". The guidelines also exclude reviews given for a benefit without clear disclosure (Google Search Central, Review snippet).
   • Also publish an About page and a real contact page. Why: checkout collects card details and license scans, yet /about-us/ returns 404 and the only "Contact Us" link opens the portal login (Part 3.4).
   • Put the physical address, entity details, and support channels on the contact page.
   • Finish the help-center move. On 2026-10-05, every support.e-ita.org address, including the old articles, redirected in 1 hop to the new Help Center at /help/, and /faq/ linked to it. Why: the rater guidelines expect customer service information on a site that takes payments (SQRG 5.5, below). Raters look for it from the homepage first (SQRG 2.5.3, below), so link /help/ from the homepage. Apply the Phase 0 copy pack to /help/. On 2026-10-05 it said the IDL “is used by drivers in 190 countries” and that “Our documents follow the 1949 Convention on Road Traffic (Geneva).” Remove the 1949 Convention claims. Replace each country count with [N], as Phase 0, step 1 does. Keep the support.e-ita.org redirects in place: the old /support/solutions address on that subdomain took 1,370 sessions in 2026 (to 2026-09-26). Check: the homepage links to /help/. A GET of /help/ shows no 1949 Convention claim and no country count other than [N]. Every support.e-ita.org URL still reaches /help/ in 1 hop.
   • Add 1 line to both defense pages: "EITA LLC takes orders only on e-ita.org." This separates the brand from lookalike sites. It matters most if eitalicense.com, which showed "EITA LLC" in its footer, comes back online (it returned 403 on 2026-10-05).
   • Keep the eitalicense.com clone down. It returned 403 on 2026-10-05. If it comes back, 301 each of its pages to the matching e-ita.org page. Why: the clone cannot rank while it is offline or noindexed. It becomes a doorway risk if it comes back without its noindex (Executive Summary, rank 4). It also shows buyers a second EITA LLC site. Who: the developer.
   • Link both defense pages from the footer, sitewide.
   The 2019 and 2022 forum threads cannot be removed, so outrank them (Part 4.2). ChatGPT cites zero brand-owned sources on the legit question today (Part 5.1). The first-party pages become the citable surface for that question.
   SQRG 4.5.1, p.36 “Any site that handles personal, private or sensitive data must provide extensive contact information. This includes sites that ask users to create passwords, share personal information, or conduct financial transactions.”
   SQRG 2.5.3, p.18 “To find contact or customer service information for a website, start with the homepage. Look for a ‘contact us’ or ‘customer service’ link.”
   SQRG 5.5, p.63 “Pages that offer payment functionality or process other types of financial transactions should receive a Low rating if there is an unsatisfying amount of customer service information or contact information.”

   Who: the marketing team writes both pages. E-ITA's leadership signs them. The developer publishes them and the footer links.

   Acceptance check: both pages are live, indexed, and dated. Both reach page 1 for the brand-plus-legit query within 90 days (a Phase 2 target). The contact page shows the physical address and support channels.

   Finding cross-ref: Part 4.2 and Part 5.1.

4. Decide the AI-crawler policy, extend the schema, and optionally publish llms.txt. The WAF (web application firewall: the Cloudflare layer that filters requests) blocks 5 AI data crawlers today (tested 2026-10-05). The AI search crawlers already get through. Work the list:
   • Write the crawler decision down in 1 paragraph. The WAF blocks GPTBot, ClaudeBot, Bytespider, CCBot, and Amazonbot (tested 2026-10-05). These collect training data for AI models. Allow GPTBot and ClaudeBot if E-ITA wants its pages in what those models learn. Keep Bytespider blocked, because no official IP list exists to verify it.
   • Keep the search and user crawlers open: OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, and PerplexityBot. All 5 returned 200 on 2026-09-29.
   • If E-ITA allows GPTBot or ClaudeBot, open AI Crawl Control in Cloudflare. Set that crawler to Allow in the Actions column. Leave Bytespider on Block. Don't turn off "Block AI bots", because that also lets Bytespider through. If the block comes from a WAF custom rule instead, exempt only the allowed bot in that rule. If E-ITA allows neither, change nothing.
   • Keep robots.txt permissive. The blocker today is the WAF, not robots.txt.
   • Match the Content-Signal header to the decision. Cloudflare's markdown copies send "ai-train=yes" by default, although the WAF blocks training crawlers (checked 2026-09-29). If training stays blocked, send "ai-train=no" from the site (Cloudflare changelog).
   • Verify bots by IP address, not by user-agent string alone. Match hits against the published lists: openai.com/gptbot.json, openai.com/searchbot.json, openai.com/chatgpt-user.json, claude.com/crawling/bots.json, and perplexity.com/perplexitybot.json.
   • Probe the 5 search and user crawlers, plus any training crawler the decision allows, in the monthly gate re-probe (Part 7.3). A training crawler E-ITA allows returns 200, ideally by 2026-10-11.
   • If E-ITA wants it, publish /llms.txt as plain text. Link only surviving canonical URLs. PROBABLE llms.txt is a 2024 community proposal (llmstxt.org). No major engine has publicly committed to honoring it. Google says its AI features need no "AI text files" (Google Search Central, AI features). This step is optional. It is 1 small file, so ship it only if the developer has time.
   • Add Organization and WebSite JSON-LD to the homepage, with the legal name. Why: Google says it helps "disambiguate your organization in search results", which matters with a lookalike like e-itca.org (Google Search Central).
   • Add BreadcrumbList markup to country pages as they rewrite.
   • Fix the markdown copies at their source, the page HTML. Cloudflare's Markdown for Agents builds each copy from the HTML on request (Cloudflare changelog). AI tools that ask for markdown read this copy as the page text (Part 5.2). Fetch each copy with curl -s -H "Accept: text/markdown" https://e-ita.org/faq/. The stray "<!doctype html>" line is already gone (checked 2026-09-29). Add 1 H1 to the /faq/ HTML, which has none today. Output the meta description apostrophe as a plain character, not "&#039;".
   CONFIRMED on 2026-09-17: 3 crawlers were blocked, llms.txt returned 404, and the markup was FAQPage-only (Part 5.2). Note: structured data is not a ranking factor. Google's documentation does not list markup as one. The markup fixes how machines read the entity.

   Who: the developer, in Cloudflare and the site templates.

   Acceptance check: a probe with each of the 5 search and user crawlers returns 200, and so does each training crawler the decision allows. If E-ITA chose llms.txt, /llms.txt returns 200 as plain text. The homepage parses with Organization and WebSite nodes. / and /faq/ return markdown with no doctype leak, and the /faq/ twin carries exactly 1 level-1 heading.

   Finding cross-ref: Part 5.2.

Phase 2: Days 30 to 90 (content rebuild)

The content rebuild replaces the templated country pages with pages a rater can tell apart. The Differentiation Standard sets the rules, and 2 model pages apply them: /italy/ (English) and /es/estados-unidos/ (Spanish). Both carry full content with every fact sourced. E-ITA adds the author and reviewer names before publishing. Clone the model pages, and do not invent a new shape.

1. Rewrite the country-template program. Every rewrite-verdict page must meet the Differentiation Standard. Start with the order-bearing pages. Per page, require all of the following:
   • At least 10 facts that appear on no other country page.
   • At least 6 primary sources, each with an accessed date. Include one source for each source type in the coverage rule (see “Citation mechanics” below).
   • 1 filled data table, such as issuing body and convention status.
   • Enough country-specific text to carry the facts above. Plan for about 800 to 1,500 words, or 2,500+ for priority markets. This number sizes the research. It is not a pass mark, because Google has no preferred word count. Never pad a page to reach it.
   • Below 30% shared n-gram overlap with every sibling page.
   • A named author and reviewer.
   • One H1. No heading string repeats within a page, and country-specific headings do not repeat on a sibling page.
   • 3 visible dates: published, reviewed, and last verified.
   Then run the production workflow:
   • Clone the 2 model pages. Carry their 12-section anatomy in the same order on every page.
   • Build each page through 6 steps. The steps are research pass, draft, SME review, linguist pass for translated pages, quality gates, and publish and annotate.
   • Run the research pass against the whitelist below. Record every URL, its status, and its accessed date in the page working file. Target 15 candidate facts per page.
   • Ship the 25 order-bearing pages first. All order share sits on English pages. Then rank the rest by click loss.
   • The first batch of 10: /italy/, /spain/, /us/, /united-kingdom/, /greece/, /kuwait/, /egypt/, /france/, /thailand/, /canada/. Plan these for about 2,500 words each. The unique-fact, source, and overlap gates decide whether a page ships, not its length.
   • Hold 3 hard gates before publish. Overlap must read below 30% against every sibling. Every numeric claim carries a cited source. No contradiction survives against the FAQ, the homepage, and every locale.
   • Where automation contributed to a page, add a per-page disclosure block. It answers who was involved, how automation was used, and why. A site-wide sentence with no per-page indication fails the standard.
   • The volume is 70 rewrites, with 10 in the first batch. The target is 2027-01-31. Why: it gives the pages about 2 months to be crawled before a likely spring update. Google ran a core update in late March 2026 (Search Status Dashboard). A 2027 date is a forecast, not a promise. The best-case path in Part 7 (back to about 270 clicks/day by late December) needs all 70 live by early December. If the claims fix goes live in the first week of October, Phase 2 starts in early November. That is about 15 pages a week. Meeting only the 2027-01-31 target needs about 6 a week. That target, not day 90, is the end of the rewrite work. Time the first 2 pages from research to publish. Then set the weekly number from that time and the names in Phase 1, step 1. If the measured rate cannot reach early December, plan to 2027-01-31. Then expect the December window in Part 7 to pass without a lift.
   184 of 195 English country pages have a sibling above 0.90 similarity today (Part 3.2).
   Spam policies: scaled content abuse “Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users. This abusive practice is typically focused on creating large amounts of unoriginal content that provides little to no value to users, no matter how it's created.”
   SQRG 4.6.5, p.42 “Pages and websites made up of content created at scale with no original content or added value for users, should be rated Lowest, no matter how they are created.”
   SQRG 3.2, p.21 “Accuracy: For informational pages, consider the extent to which the content is factually accurate. For pages on YMYL topics, consider the extent to which the content is accurate and consistent with well-established expert consensus.”
   SQRG 4.7, p.46 “The website terms of use states that "some articles" are generated by artificial intelligence and may have errors or be out of date; there is no indication to which pages this statement applies. The information in this article is not trustworthy and is Lowest E-E-A-T”

   Who: the marketing team writes each page to the Differentiation Standard. A named reviewer signs it off. The translator chosen in Phase 0 reviews the translated pages. The developer publishes.

   Acceptance check: every page's claims carry their cited sources, with no open rows in the page working file. A diff against every sibling reads below 30% overlap. The author and reviewer byline and the author markup are live on the page. Heading uniqueness is checked in the page and against sibling pages.

   The change log records the gate results per page.

   Finding cross-ref: Part 3.2 and Part 3.5.

2. Replace multi-year SKUs with renewal plans. Work the list:
   • Retire the 2-year and 3-year booklet options.
   • Sell a 1-year product as the flagship.
   • Offer a prepaid renewal plan at the same total price, so 3-year revenue holds.
   • State on the checkout page, in plain text: the product is a private translation booklet, not a government permit.
   • Carry the same statement in the FAQ, the terms pages, and every locale.
   • Where US buyers are addressed, quote the official 1-year rule. USA.gov states: "An IDP is valid for one year and cannot be renewed." (USA.gov, International Driving Permit, last updated 2026-04-14).

   Why change it: the booklet is a translation, not an IDP. Selling it with a 2- or 3-year "validity" reads like an official permit term. A US-issued IDP lasts 1 year (USA.gov), so a 3-year term signals a product that is not one. Perplexity listed the 1/2/3-year options among its warning signs about the site (Part 5.1). The product and price decision stays with E-ITA. The renewal plan is one way to keep the revenue without the validity claim.

   PROBABLE No private IDP-vendor renewal SKU was verified. Treat the renewal-plan shape as the honest-reframe path, not a copied precedent.

   SQRG 5.6, p.63 “If you find the information about the website or the content creator to be exaggerated or mildly misleading, the Low rating should be used.”

   Who: E-ITA's leadership decides the product and renewal terms. The developer changes the checkout.

   Acceptance check: no page sells a validity above 1 year as permit validity. A renewal plan replaces the multi-year options. The checkout states what the product is and is not.

   Finding cross-ref: Part 5.1.

3. Build the country-guide program on original data. Work the list:
   • Some countries have enough first-party data (below) to say something the rewritten country page doesn't. For each of these, add 1 "Do I need an IDP in X" guide in the existing guide collection. Otherwise, keep that answer on the country page, as the /italy/ model does. This way, no country gets a second page that answers the same question.
   • Insert 3 first-party proof classes: anonymized orders per destination, reported rental-desk issues, and refund reasons.
   • Suppress any country count below 10. Round counts from 10 to 99 down to the nearest 10, and larger counts down to the nearest 50. Label every figure with its window, for example "Jun-Sep 2026". Why: no figure can point to one customer or show exact sales.
   • Publish under the same authorship, sourcing, and citation rules as the country pages.
   • Mine support tickets monthly. Add the top 5 to 10 questions per country to the page FAQ. Answer each through the citation gate.
   • Log rental-desk reports in a country-keyed sheet: document type, printed or digital, company, country, date.
   • Update a country's "Rental desk reality" section after 2 consistent reports. Date the update.
   • Never invent a report or a count to fill a section.
   Anonymize every figure. Originality scores 2.22 of 10 on the 18 pages of the full pass (Part 3.5). It scores 2.18 across the 445 live pages of the crawl-wide run. Original information is the first helpful-content question.
   Helpful content guidance “Does the content provide original information, reporting, research, or analysis?”
   Helpful content guidance “Does your content clearly demonstrate first-hand expertise and a depth of knowledge (for example, expertise that comes from having actually used a product or service, or visiting a place)?”

   Who: the marketing team builds the guides. Support supplies the monthly ticket questions and rental-desk reports.

   Acceptance check: every country guide names its sources and dates. Every figure traces to the order system. The authorship block is live.

   Finding cross-ref: Part 3.5.

   Cadence: run this as a repeatable monthly cycle, not a one-off project. The cycle is research, a scored list of ideas, drafting, an editor pass, an audit and fix pass, and localization where needed. The Content Skill Walkthrough deck shows the cycle end to end. The gates in the deck match the gates in this plan. Every page needs a named author and a named reviewer, as in Phase 1, step 1. A page ships only when a reader cannot get the same thing elsewhere.

4. Fix the localization gaps. Patch the locale template once, then go locale by locale:
   • Localize every locale title tag, meta description, and H1. Ship es, fr, it, and ar first. These 4 locales hold 124 pages in the verdict sheet. Of these, 78 stay after the plan (56 keep, 22 rewrite).
   • The 124-page template-level pass is scoped per locale. PROBABLE The scope is an estimate, not a measured value.
   • Add the missing H1 to the 16 keep pages without one (per the 2026-09-13 crawl). The other 21 pages without an H1 are kill pages, so skip them. Every locale FAQ carries exactly 1 H1.
   • Remove the doubled brand suffix on locale FAQ and legal pages.
   • Remove the duplicate H2 string on eligibility from the source of the page template. It repeats in source on 313 of the 445 live pages, and one copy sits inside a widget payload. The rendered pages do not show a duplicated H2 today (live check 2026-09-21). This is cleanup, so do it while the template is open for the fixes above.
   • Hold 1 register per page. Spanish currently mixes tu and usted.
   • Pick 1 product term per locale. Spanish holds one of IDL, PIC, or Permiso de Conducir Internacional.
   • Ban English title case in non-English headings.
   • Have a native speaker translate or edit every kept page in each locale (es, fr, it, ar), not just review it. A paid human translation service also works, for example TheWordPoint. It costs about 8 cents a word. Standard delivery, in 24 to 48 hours, is in that price. Faster delivery costs extra (a rush fee). Credit the translator by name on each locale page. Why: E-ITA used AI for the translations. The text reads fluently but mixes registers and product terms, and no page names a translator today (Part 3.3). Carry the same citation set as the English master.
   • Fix hreflang among surviving locales only. Emit an alternate only when the localized page exists, returns 200, and links back. Today the hreflang sets break two of the three conditions: 30 non-reciprocal edges and 33 targets at redirected URLs.
   • Use one hreflang method only. Remove the hreflang links from the XML sitemap and keep the on-page tags. The sitemap uses codes such as "en_US" and "es_ES" today. Google's format joins codes with a hyphen, and using two methods adds no benefit (Google Search Central, localized versions). The on-page tags use "en" and "es".
   • Localize the language-switcher chrome.
   • The verdict sheet retires 7 locale hubs: /de/, /el/, /hi/, /pt/, /ru/, /th/, /tr/. Apply this step to the 4 survivors: /es/, /fr/, /it/, /ar/.
   8 of 11 locale hubs carry English title tags today (Part 3.3). The translated text itself reads fluently in every sampled locale. These fixes are metadata, headings, and consistency, plus a named reviewer on each page.

   Who: the developer patches the locale template once. A named native-speaker translator, in-house or from a paid service, edits each surviving locale.

   Acceptance check: a crawl of the 4 surviving locale hubs shows zero English title tags. Every locale FAQ has exactly 1 H1. Each locale page names its reviewer. A crawl finds zero non-reciprocal hreflang pairs and zero hreflang targets that return 301 or 410. The XML sitemap carries no hreflang links.

   Finding cross-ref: Part 3.3.

The citation standard: whitelist and blacklist

Every rebuilt page stands on citations. The IDP topic is YMYL (Your Money or Your Life: topics where wrong content can harm finances or safety), so the scrutiny standard is high.

The rule of thumb: the official source wins. When sources conflict, the page states the conflict and defers to the more authoritative source.

(a) The whitelist: approved sources by claim type

(b) The blacklist: source types to never cite

(c) Citation mechanics

• Carry at least 6 primary outbound citations per country page, one for each source type in the coverage rule. The types are convention status, issuer of record, national authority, rental company, the consumer-protection frame, plus 1 context source. That includes at least 1 government or club authority and 1 rental company source.
• Verify officialness before citing. The URL must sit on a state-controlled domain, or the ministry's own portal must link to it. The depositary record (UN Treaty Collection, UNECE) also qualifies.
• Treat a mirror as citable only when the official page is unreachable or unreadable to machines. Label it as a mirror in the sources list.
• Check machine readability. Major AI crawlers do not run JavaScript (Vercel crawler study, 2024-12-17), so they cannot see a page that is blank in View Source. Test the cited page with View Source, including the site's own pages.
• Weight by readership. A source nobody reads carries little corroboration. Prefer sources that rank or get cited for the topic.
• Link the official source, never a mirror, when both are readable.
• Cite the specific statute or page, not the domain home page. Examples: Codice della Strada art. 135, 7 FAM 1430, and service-public.fr F11534.
• Keep the citation next to the claim it supports, in the same passage. Engines cite passages, not pages.
• Date-stamp every legal claim. Every source line carries an accessed date in YYYY-MM-DD form.
• Re-verify the 10 first-batch pages (the priority markets) monthly and all other country pages quarterly. Why: fines and permit rules change, and raters check YMYL pages for accuracy (SQRG 3.2, p.21). Reflect any source change within 30 days, with a change-log entry.
• Trace every acceptance claim to convention status or an issuer of record, never to marketing copy.
• Present the conflict on the page when sources disagree, and let the more authoritative source win.

Sources: the officialness test follows SQRG 3.4 (quoted above). The View Source test follows the Vercel crawler study. Cite statute text from official registers.

Who: the marketing team applies the standard to every page it writes. The named reviewer checks it before publish.

Acceptance check: every rebuilt page carries at least 6 primary outbound citations, one per source type. They include at least 1 from a government or club authority and 1 from a rental-company source. Every cited URL passes the View Source check. Every source line carries an accessed date in YYYY-MM-DD form.

Finding cross-ref: Part 3.5.

Phase 3: Days 90 to 180 (entity building)

Entity building grows who e-ita.org is off-site. It starts after the on-site trust layer is live, because every new mention then points back to a site that can hold it.

1. Open the affiliate program. Work the list:
   • Stand up tracking before recruiting anyone. Run the program on an affiliate platform of the Impact, PartnerStack, or Tapfiliate class. The platform needs 30-day cookies, sub-ID parameters, coupon codes, and custom landing pages.
   • Write the program spec. Target the niche band of 30% to 50% commission per completed order, the published category precedent. Publish the band only after a unit-economics check confirms fulfillment can carry it.
   • Recruit rental agencies first, because the buyer's moment of need is the rental counter. Then recruit driving schools, then insurers and travel-insurance checkouts.
   • Vet every partner before approval for a real business and real traffic. Also check that the partner's site makes no claims that violate the exclusion list or the 1-year framing.
   • Review partner quality monthly through the sub-ID ledger. A partner program without sub-IDs cannot be audited per placement.
   • Add an agent or reseller tier only after the claims pass is documented. A reseller repeating an unsafe claim is your claim under the FTC frame.
   • Ship a claims-approved product paragraph in the partner kit. Partners use it verbatim or write inside the same rules.
   • Put the disclosure clause in partner terms from day one: partners must disclose the paid connection (16 CFR 255.5(a)). Disclosure language ships in the kit.
   • Require partners, in the partner terms, to mark every affiliate link to e-ita.org rel="sponsored".
   • Never let partners buy placements for ranking credit. A partner's paid link is your liability.
   • Never buy links for ranking. Paid placements that pass ranking credit violate the link spam policy even when disclosed.
   • Never buy marketplace or directory listings as links. Never run press-release link campaigns. Never place guest posts to borrow host authority.
   • Why affiliates: from 2026-10-05, Google's government documents and services ads policy limits those ads to authorized providers. Trade press reads it as covering this category. Microsoft Advertising has required preapproval since 2026-08-19, and its rule is already in force. Affiliates, Meta paid social, and organic remain open.
   Spam policies: link spam “Google does understand that buying and selling links is a normal part of the economy of the web for advertising and sponsorship purposes. It's not a violation of our policies to have such links as long as they are qualified with a rel="nofollow" or rel="sponsored" attribute value to the <a> tag.”

   External sources: the 30% to 50% commission band, the 30-day cookie, and the partner profile come from a rival's affiliate terms. Source: idaoffice.org, Travel Affiliate Program, read 2026-09-29. That read is market research, not a page-citation source. Google Ads, Government documents and services policy update, announced 2026-08-04, effective 2026-10-05. Microsoft Advertising Government Services Policy, effective 2026-08-19 (PPC News Feed, 2026-08-26); Microsoft Advertising preapproval page. FTC endorsement rule, 16 CFR 255.5(a).

   Who: E-ITA's leadership sets the terms and approves each partner. The marketing team recruits and runs the monthly audit. The developer sets up tracking.

   Acceptance check: the spec is written. Tracking works before the first recruitment email. Partner terms carry the disclosure clause.

   A sample crawl shows rel="sponsored" on every partner link. The sub-ID ledger shows traffic per partner from the start.

   Finding cross-ref: Part 4.3 and Part 5.7(c).

2. Reclaim the reclaimable links. Work the list:
   • Work from the dead URLs listed below. Sort them by the authority of the dead page, not by link count.
   • Live-test every candidate with a GET request that prints the headers, such as curl -s -o /dev/null -D - URL. Never trust a tool's "lost link" status alone.
   • 301 every dead URL that still holds external links to the live page that now carries the linked content. Bulk redirects to the homepage discard the topical signal that made the link worth having.
   • Add the 2 dead 2021 header images under /wp-content/uploads/2021/ (page authority 35 and 33) to the redirect map. 301 each to the live image that replaced it.
   • These are already rows in the stage 1 redirects. Here, only confirm that they held:
     • /Italy/ and /Italy3/ (3 followed links from travel blogs about Italy)
     • /wikiprocedure-idl-apply/, the dead URL that holds 473 nofollow links from wikiprocedure.com (the gain is the visitors, not link value)
     • /how-to-print-your-international-driving-document/, the old print-guide URL that travelynnfamily.com links to
   • Ship single-hop 301s. Verify every chain ends in 200.
   • Request indexing for each fixed target page in Search Console the same day, within the daily limit. Re-check the full map at 30 and 90 days.
   • Watch the education-domain referring pages quarterly. A link inside a retired course page can vanish at any CMS cleanup.
   • Run unlinked-mention reclamation on coverage that names the brand without a link. Warm requests convert at 15% to 40%. Convert only high-authority, topically relevant mentions, and leave low-value mentions alone.
   • Where a redirect cannot fix the source page, run reclamation outreach to the linking site.
   • The eitalicense.com clone is handled in Phase 1, step 3.
   • Check the 4 alias domains for their own history: past content and outside links. An alias with neither can keep its domain-level 301 to the homepage. An alias with its own content or links gets page-to-page redirects onto matching live pages. Never republish an alias's old content, which is the pattern the expired domain abuse policy below targets (Part 5.7(a)).
   Part 4.1 maps the full list. The redirect items are recoverability 1.0: no outreach, and no one can say no. PROBABLE The 15% to 40% warm-reclamation band is vendor-reported. Treat it as a planning band, not a promise.
   Spam policies: expired domain abuse “Expired domain abuse is where an expired domain name is purchased and repurposed primarily to manipulate search rankings by hosting content that provides little to no value to users.”

   External sources: the reclamation method and the sort-by-dead-page-authority rule follow the standard lost-equity recovery practice for migrated sites. The 15% to 40% warm-reclamation band is vendor-reported (Reporter Outreach, May 2026).

   Who: the developer ships the redirects. The marketing team runs the outreach.

   Acceptance check: every listed URL returns 1 hop, 301, to a live equivalent and ends in 200. None targets the homepage. The 30-day and 90-day re-checks are logged.

   Finding cross-ref: Part 4.1.

3. Start the editorial layer. Work the list:
   • Open a news-and-guides section under named authors, with the sourcing rules and citation standard above.
   • Lead with original data: permit statistics, destination rule changes, fee updates.
   • Ship 1 maintained dataset per quarter, built on official sources. The first candidate is a country-by-country table of what a visiting driver needs. It covers whether an IDP is required, the road-traffic convention, and the maximum stay on a foreign license. Build it from the whitelist sources, and refresh it on every rule change. Time the pitch before the summer travel peak.
   • Structure every asset for extraction with an answer-first opening and a strict heading hierarchy. Add 1 quotable stat statement per section, a methodology block, and tables. PROBABLE A 2026 preprint found structure-only changes lifted citation rates 17.3% across 6 engines (GEO-SFE, arXiv 2603.29979). It is a single study, not yet replicated.
   • Add 1 change-log entry per substantive update on each page.
   • Never restamp a date without a substantive change.
   • Build 2 target lists before pitching. The link list holds major editorial publications. The citation list holds the roughly 20 outlets that AI engines actually cite for this category. Take them from the monthly check of what the engines cite (Part 5.6 lists today's). The 2 lists barely overlap, so don't reuse one for the other.
   • Seed the citation list with the travel-media slots engines already read: wise.com, remitly.com, and thelocal.it class sources.
   • Pitch original data only. 88% of journalists delete off-beat pitches (Muck Rack, 2026). Cap each campaign at about 100 personalized sends. The yield sweet spot is 51 to 100 sends. Campaigns of 1,001 to 5,000 sends earned 70% fewer links (BuzzStream, 31-million-email analysis).
   • Skip wire press releases. PROBABLE Press releases appear in about 1% of AI citations (Muck Rack, What is AI Reading?, updated 2026-02-16).
   • Optional, lowest priority: open a YouTube education channel. No engine in the Part 5.6 sweep cited a video, so start it only after the steps above run. Ship short explainers: what the document is, where it is not accepted, and what it does not do. Post transcripts. Description links carry no ranking equity. The channel is a citation and brand-defense surface.
   • Keep the entity facts identical everywhere: name, address, and product description on the site, Trustpilot, any video channel, and partner pages.
   • Stay out of Reddit while the 2022 scam thread stands. Participation now reads as astroturfing. Reddit's own systems stop about 25,000 net new spammy posts and comments a day (Reddit, 2026-07-06). Seeded campaigns are the category's named failure pattern (Kotaku on the Trap Plan case, 2025-11-10, and FTC 16 CFR Part 465).
   • After the claims fix is verified and the thread's premise is weakened, start the real-person answering program. Use 1 named account with 30+ days of genuine participation before any brand mention. Put a disclosure line in every brand-adjacent answer. Post mentions, not links.
   • Never draft AI comments at volume. Never purchase aged accounts. Never let colleagues upvote.
   • Measure monthly: the money-query prompt set, plus the Bing Webmaster Tools AI Performance report for citation counts. PROBABLE The Bing report launched 2026-02, and its per-property counts are unstable month to month. Read the counts by quarter, not by month.
   • Run a quarterly audit: every numeric claim cited, every date backed by a logged change.
   Why: for "best online IDP" and "IDA vs AAA" questions, AI engines cite editorial pages and travel media. E-ITA has no page that can take those slots (Part 5.6, displacement map), and no blog or editorial section (Part 3.4).
   Risk: never restamp a date without a substantive change. Freshness signals are significance-based, so cosmetic date edits add risk, not benefit.

   Helpful content guidance “Are you changing the date of pages to make them seem fresh when the content has not substantially changed?”

   External sources: Muck Rack, 2026 State of Journalism (88% delete off-beat pitches; syndicated release, the globenewswire.com original was unreachable on 2026-09-22). BuzzStream spray-and-pray analysis, 2025 (51 to 100 sends as the yield sweet spot; 1,001 to 5,000 sends earned 70% fewer links). Reboot Online JRPass FOI case study (travel-category data-PR precedent, 100+ links, vendor-reported). GEO-SFE, arXiv 2603.29979. Reddit, 2026-07-06.

   Who: the marketing team, under named authors.

   Acceptance check: the section is live with named authors. Both target lists exist before the first pitch. A quarterly audit finds every numeric claim cited and every date backed by a logged change.

   Finding cross-ref: Part 3.4.

Phase 4: Ongoing (measurement cadence)

Who: the marketing team runs the weekly and monthly checks. E-ITA's leadership makes the call at each dated decision review.

Recovery is declared by data, not by effort. Part 7 sets the scorecard in advance, with 6 readiness checks and the weekly numbers. It also sets the rules for when to call a recovery and when to change course. Read Part 7 before the removals start. The expected dip starts as soon as the removed pages go.

PART 7 Measurement and Monitoring

This part is the KPI framework. Every threshold below was set before any recovery data existed. The weekly and monthly reads use Search Console, GA4, E-ITA's order dashboard, and the Trustpilot business account. The gate probes in 7.1 request pages as named crawlers and read raw status codes (410, 301, 403). For this reason, the developer runs them from a command line.

7.0 Terms

• GSC means Google Search Console.
• The probe set means the 257 queries that still showed impressions after the collapse. The list ships with the verdict sheet as its own tab, so every monthly read uses the same queries.
• Weighted position means the average position, weighted by impressions.
• Rolling mean means the average of the last N days.
• A key event means a GA4 count of a completed action.
• The plateau path means the 2026 pre-collapse level of about 270 clicks/day, the recovery target (not the 2025 peak).
• Kill and fold are verdict-sheet labels. A kill URL is retired and returns 410 (gone). A fold URL is merged into its closest surviving page with a 301 redirect.
• The WAF is the web application firewall, the Cloudflare layer in front of the site that filters requests.
• A markdown twin is the plain-text copy of a page the site serves to tools that ask for markdown.
• “(not set)” is the block of GA4 visits with no recorded source. It holds 58.3% of post-collapse orders (Part 5.5).
• A named update is a Google core or spam update listed on the Google Search Status Dashboard.

7.1 The 6 readiness gates

These gates measure deploy state, not traffic. Each is a yes-or-no probe. None was green on 2026-09-17, the pre-remediation state. Gate 1 is partial. Gates 5 and 6 carry the recovery work: the page cut and the claims fix. Gate 3 tells Google and AI tools who runs the site. Gates 1, 2, and 4 cover AI tools only (Part 5.2).

7.2 Weekly checks

Run these every Monday. GSC data inside the last 72 hours is partial. Read through the latest complete day only.

• Clicks/day vs the 30.9 baseline. GSC, Performance, Date view. Sum the last 7 complete days. Divide by 7. Compare to 30.9, the post-collapse floor, and to the 270 plateau path.
• Index coverage. GSC, Index, Pages. Record total indexed. In the recommended order, expect the count to fall toward the surviving page set during 2026-10-19 to 2026-11-15. A rise in "Crawled - currently not indexed" among kept or rewritten pages is a stop signal. Check it by filtering the report to the submitted sitemap. Why: none of the 438 monitored URLs sat there on 2026-09-29, so a kept page landing there means Google read it and refused it. If it happens, pause the rewrite schedule and inspect the first affected URLs.
• Review velocity. Trustpilot business account. Record new reviews, the 1-star share (baseline 11%), and the reply rate on negatives (target 100% within 48 hours).
• Order tracking. E-ITA's order dashboard, daily. Zero on any day is a tracking alarm, not a sales event. In GA4, record the "(not set)" share of orders. It is 58.3% today (Part 5.5). Above 65% is amber, because attribution is getting worse. A falling share means attribution is recovering.

7.3 Monthly checks

• AI-engine re-sweep. On the first Monday of each month, ask ChatGPT, Gemini, Perplexity, and Claude the early-September questions again. Include "Is e-ita.org legit?" and "best place to get an international driving permit online 2026?". Log 3 things per engine: the verdict, the sources it cites, and whether any first-party page is cited. The early-September baseline is 4 of 4 negative, anchored on the FTC alert, the forum threads, and the site's own claims.
• Gate re-probe. Re-run the 6 gate probes. A gate that was green and turns red is a deploy regression. Fix it before any new work ships.
• Position read. GSC, query export, last 28 days. Compute the impression-weighted position on the probe set. Compare to 33.60, the collapse-window value, and to 9.40, the pre-break reference.

7.4 Scorecard

The dates in this table follow the recommended order. They are not deadlines. A later date moves the read, not the plan.

7.5 Decision rules

These rules decide what the numbers mean. They are fixed in advance so no one re-reads them after the fact.

• Lift candidate. The 7-day rolling clicks/day rises more than 20% above the trailing 28-day mean. It also stays above the 30.9 floor. It holds 7 or more days, confirmed only after a named update completes. A rebound from the expected dip alone is not a lift. This rule is the only authority for declaring recovery. Why 20% and 7 days: the June 2026 update cut clicks 22.4% over its 3-day window, then rebounded within 4 days (Part 2.3c). After that rebound, the 7-day average rose at most about 6% above its trailing 28-day mean (2026-07-05), so it could not trip this rule.
• Expected dip. It starts with the stage 4 deploy and lasts about 7 weeks. In the recommended order that is 2026-10-19 to 2026-12-06. Clicks below the 30.9 floor are expected in that window while kills and folds leave the index. It is amber if the 7-day rolling value falls below 15 for 7 straight days. It is red if it stays below 15 for 14 days. In that case, audit redirects and tracking first. Why 15: it is below the worst single day since the collapse (18, Part 5.9), and the removed pages carried only about 2 clicks/day.
• Suppression-lift detector. All 4 conditions must hold at once, on the probe set:
  • Position is 5 positions better than the trailing 28-day mean.
  • That holds for 14 consecutive days.
  • Probe impressions stay at or below 1,269 per day on each of those days.
  • No named spam or core update is mid-rollout in the window.
  On trip, add content first to the queries already ranking in positions 4 to 20. Do this before any click lift is visible.
• Why 5 positions and 14 days: in the June 2026 spam update, position moved only from 8.76 to 8.59 (Part 2.3). A 5-position move held for 2 full weeks is far outside that noise.
• Why 1,269: it is half the probe set's pre-break impressions (96,447 over 38 days, 2,538/day). Probe impressions run about 87/day today.
• Brand guardrail. Brand weighted position beyond 10.8 sustained 14 days is amber, and beyond 15 is red. 10.8 is 3 positions worse than brand's post-collapse 7.84 (Part 1). A brand slide re-opens the cause assessment and pauses all recovery reads.
• Pivot trigger. It is evaluated 28 days after the first named update that completes 28 or more days past a fully green redirect map, and again at the 2027-03-31 final read (7.4). At the final read, a 28-day average below 100 clicks/day trips it on its own. Otherwise, trip on either condition. Condition 1: generic recovery is below half of brand recovery. Condition 2: rewrite-cohort pages earn 10 or more generic clicks/day while keep-cohort pages earn under 5. That would show rebuilt pages ranking where kept pages do not. The 10 and 5 are this plan's planning cut-offs, not Google numbers. On trip, assemble the pivot packet: generic ratio, brand ratio, both cohort reads, and gate history. A pivot moves effort from keeping and consolidating existing pages to replacing them with pages rebuilt to the Differentiation Standard. The decision is yours, and the numbers above are the whole case.

7.6 Failure states

B Appendix: Verdict and Redirect Tables

The two working sheets behind the verdicts (Part 3) and the plan (Part 6) are on their own pages, in full. Table B.1 gives a class to every URL in the 454-URL audit inventory. Table B.2 maps the redirects and removals for the URLs that do not stay.

B.1 Page-class verdict sheet: 454 URLs

One row per URL. Verdicts: keep 73, rewrite 70, fold 149, kill 162. The four classes are keep, rewrite, consolidate, and retire. Fold means consolidate (a 301 into a hub). Kill means retire (a 410). The verdict rules sort pages into these four buckets. Keep = homepage, EN core/info pages, hubs and legal/info pages of the four locales with demand (es, fr, it, ar). Keep pages need no writing. Add the author line, contact link, and title fixes. Rewrite = a country page with any paid orders (2026-06-01 to 2026-09-12) or 100+ pre-break clicks in 90 days. A country page with 20+ GA4 key events in 2026 YTD (all channels, 2026-01-01 to 2026-09-26) is also Rewrite. Rewrite is the only bucket that needs writing. Fold = every other country page with 5 to 99 pre-break clicks. Each one gets a 301 to its regional hub (EN, 5 hubs) or its locale hub (es, fr, ar). Fold needs no rewrite of the old page. The hub needs a short, sourced row for each folded country. Kill = under 5 pre-break clicks, no orders, and under 20 GA4 key events. The 162 pages carry 37 key events between them in 9 months. Kill also covers every URL in the seven retired locales (ru, pt, de, el, tr, hi, th). Together they had 283 clicks in the 90 days before the drop and no orders. They had 4 clicks in the 23 days after it, 1 each on 4 pages. Serve a 410 and remove each one from the sitemap, hreflang, and internal links, including the language switcher and /countries/. Score columns use the E-E-A-T rubric (1 to 10) described in Part 3. Blank cells mean the metric does not apply. The exceptions are nine rows with no scores and five country rows with no cosine. Five of the nine are pages added to the inventory outside the 2026-09-13 crawl (live-verified 2026-09-21). The other 4 are legacy locale slugs that already 301 to their translated pages. The Δ clicks/day % column is blank where pre-break clicks are zero. Cosine and boilerplate exist only for pages with a sibling comparison. The GA columns are the GA4 landing-page export (all users, all channels, 2026-01-01 to 2026-09-26). Key events are mostly generate_lead, which is not an order. Read them as a relative signal. The landing-page report groups each page path across e-ita.org and the apply and order subdomains, so its key events combine all three. The reasons for the old pre-migration URLs in the redirect map count e-ita.org only, because a redirect catches only e-ita.org visits. Pre-break clicks count the exact e-ita.org URL.

Open the full verdict sheet (also ships as an Excel workbook).

B.2 Consolidation and redirect map: 341 rows

The plan redirects 179 rows (301) to the closest matching page and removes 162 rows (410). Ship the map in 4 stages (Part 6, step 3). Stage 1: 15 redirects now. Stage 2: 35 as the locale hubs gain rows. Stage 3: 129 as the 6 new pages go live. Stage 4: the 162 removals, after E-ITA reviews the fold and kill lists. The Planned action column is the deploy plan, not the current state. The live columns show what each old URL and target returned on 2026-10-05. This map covers all 311 fold and kill URLs from Table B.1 plus 30 URLs that have no verdict row. The 30 are 14 rental-brand guide redirects, 4 legacy aliases, and 12 pre-migration paths from GA4 landing-page data. Those paths sent thousands of sessions into 404s. The -uk/ path had 3,250 sessions and 252 key events in 2026 YTD. The 18 retired /translations/ pages wait on the decision about the university (.edu) links that point at them (Part 4). The 5 regional hub targets and the car-rental guide target are new pages. They returned 404 on 2026-10-05. Build them before the rows that point at them ship. 15 rows can ship now, including 11 of the 12 pre-migration rows (-uk/ to /united-kingdom/ among them). The 35 rows that fold translated country pages into /es/, /fr/, and /ar/ ship as each locale hub gains a row for that country. Only the /international-drivers-license-eu/ pre-migration row waits for /europe/.

Open the full redirect map (also ships as a CSV file).