First 90 days post-launch indexing and growth tracker
A website post-launch checklist should create evidence, not seven rounds of screenshots nobody can compare. Check the same signals on days 0, 3, 7, 14, 30, 60, and 90: index status, sitemap state, indexed pages, impressions, meaningful query positions, conversions, and repeatable AI mentions. Log every intervention beside the result.
Published 10 sources free browser-local tracker
01 / Schedule
What to check on each day
The checkpoints are observation dates, not ranking deadlines. Google says crawling can take a few days to a few weeks and does not guarantee indexing after a request. That is why this plan escalates from discovery checks to pattern diagnosis instead of declaring failure on day three. [2]
-
00Launch baseline
Prove that the measurable system works
- Save the list of canonical URLs intentionally eligible for search.
- Confirm important URLs return the expected public page, canonical, and index controls.
- Submit the sitemap, inspect the homepage and one representative deep page, and save both results.
- Trigger every real lead or sale action and verify the corresponding analytics key event.
- Freeze the AI prompt bank and run the pre-growth baseline before promotion changes exposure.
Fix now: non-200 pages, blocked crawling, noindex on intended pages, wrong canonicals, an unreadable sitemap, or missing conversion events.
-
03Discovery check
Confirm Google can discover and fetch
- Open URL Inspection for the homepage and priority pages; record indexed result and live-test result separately.
- Record the sitemap status, last read date, discovered URL count, and any parsing error.
- Check whether analytics and conversion events have received real or controlled test data.
- Do not interpret an empty Performance report as proof of a technical failure; new properties can take up to a week to show data. [7]
Intervene when: the live test cannot fetch the page, the page is blocked, or the sitemap is not readable. “Not indexed yet” by itself is an observation.
-
07First index read
Compare the indexed version with the live version
- Record indexed priority pages against the intended-indexable denominator.
- For missed pages, save last crawl, Google-selected canonical, referring page, and referring sitemap where available.
- Export the first usable Search Performance window; accept that recent data normally lags by two to three days. [7]
- Re-run the fixed AI prompt bank without changing prompts that returned no mention.
Intervene when: a technical exclusion repeats across important URLs, a wrong canonical is selected, or internal discovery paths are absent.
-
14Pattern check
Diagnose groups, not isolated screenshots
- Group non-indexed URLs by reason, template, directory, canonical target, and link depth.
- Confirm the sitemap contains only canonical, index-intended URLs and uses accurate modification dates.
- Separate zero impressions from zero clicks; a page cannot earn a search click before it earns an impression.
- Record any fix as a dated hypothesis with affected URLs and a future decision date.
Intervene when: priority pages remain unknown or excluded for a shared, fixable reason. Do not request indexing again until the underlying condition changes.
-
30Search baseline
Move from availability to relevance
- Export page-by-query Search Console data for the same complete window and keep the raw file.
- Mark query-page pairs as meaningful only when the query matches the page's job.
- Record impressions, clicks, average position, and conversions without collapsing them into one “SEO score.”
- Repeat the AI bank and score mentions, direct citations, and factual accuracy separately.
Improve carefully: strengthen one weak page around demonstrated intent, internal links, evidence, or snippet clarity. Preserve the before state.
-
60Focused improvement
Test one documented growth hypothesis
- Compare days 31–60 with days 1–30 and annotate launches, campaigns, outages, and major edits.
- Prioritize relevant query-page pairs gaining impressions but still failing to answer the intent completely.
- Trace conversions back to landing pages and channels; confirm event definitions did not change mid-period.
- Review AI citations for source quality and accuracy, not only whether the brand name appeared.
Choose one lever: improve an existing page, add a missing supporting page, clarify entity evidence, or repair a conversion path. Log which signal should move.
-
90Growth decision
Decide what the next quarter is for
- Compare intended and indexed pages, then explain every material gap.
- Identify which query clusters gained relevant visibility and which pages produced meaningful actions.
- Calculate AI mention, citation, and accuracy rates from the unchanged prompt bank.
- Close each experiment as keep, revert, revise, or inconclusive; carry only named hypotheses into the next 90 days.
Make the decision: scale what produced qualified visibility or conversions, repair the proven constraint, or reposition pages whose intended demand never appeared.
02 / Evidence tracker
Build the seven-check evidence log
Entries stay in this browser's local storage. Teqri does not receive the site name, dates, metrics, or notes. Export the CSV into the project's permanent evidence folder.
Set the launch date to calculate all seven checkpoints.
| Checkpoint | URL Inspection | Sitemap | Indexed pages | Impressions | Meaningful queries + position | Conversions | AI mentions / tests | Evidence path + decision note |
|---|---|---|---|---|---|---|---|---|
| Day 0 | ||||||||
| Day 3 | ||||||||
| Day 7 | ||||||||
| Day 14 | ||||||||
| Day 30 | ||||||||
| Day 60 | ||||||||
| Day 90 |
No API connection is implied: enter values from Search Console, analytics, your CRM, and saved AI tests. Clearing browser storage or using another device will not carry the log; export it after every checkpoint.
03 / Metric contract
Define each field before the numbers arrive
A stable definition matters more than a decorative dashboard. Write the definition into the first evidence note and do not change it silently.
- 01URL Inspection status
- Record Google's indexed result and the live-test result separately for the homepage plus a fixed sample of priority templates. Save last crawl, user-declared canonical, Google-selected canonical, referring sitemap, and the stated exclusion reason. A successful live test means the page appears eligible; it does not prove that the page is indexed. [3]
- 02Sitemap state
- Record submitted sitemap URL, status, last read, discovered URLs, and errors. “Success” means the file was fetched and parsed—not that every URL was indexed. Google says submitted URLs can still take time to crawl and some may not be crawled. [4]
- 03Indexed pages
- Use the Page indexing report for the site-level count and URL Inspection for important individual URLs. Track indexed canonical pages divided by intended-indexable canonical pages. Keep redirects, duplicates, private pages, and intentional noindex URLs outside that target. [5]
- 04Impressions
- Export a complete, repeatable date window from Search Performance. Filter the same search type, country, device, and page scope when comparing periods. Search Console defines an impression by whether a link was seen under the rules for that result type; it is not a visit. [6]
- 05Meaningful query positions
- Save query, credited page, impressions, clicks, and average position together. “Meaningful” means the query represents demand the page is designed to satisfy. Position is an average across changing searches—not a fixed rank—and some anonymized or lower-volume queries will not appear in the table. [6] [8]
- 06Conversions
- Name only actions that matter to the business: a qualified form, booked call, purchase, signup, or another agreed outcome. In GA4, create or identify the corresponding event, mark it as a key event, verify it in Realtime, and preserve the counting definition. [9]
- 07AI mention tests
- Use a pre-written prompt bank and fixed platform set. For every output, record exact prompt, platform and surface, date, account or location condition, brand mention, direct citation URL, factual accuracy, and saved evidence. A mention without a citation and an accurate citation without a mention are different outcomes.
04 / AI visibility
A repeatable AI mention protocol
AI answers can vary across time, model, account, location, and product surface. The goal is not to make one answer look favorable; it is to produce a comparable sample whose uncertainty is visible.
- 01
Freeze the prompt bank
Write five to ten prompts for each commercial or informational intent before the baseline. Keep exact spelling and order. Add a prompt only as a new series; never overwrite history.
- 02
Freeze the test conditions
Name each assistant, model or search surface where visible, login state, language, and testing location. Start a fresh conversation for every prompt unless conversational context is the variable being tested.
- 03
Save every result
Record date and time, exact output or screenshot, cited URLs, and whether the business was named accurately. Keep zero-mention runs; removing them invalidates the rate.
- 04
Score separate outcomes
Calculate mention rate, direct-citation rate, and factual-accuracy rate independently. Add a short human review for relevance and sentiment instead of forcing those judgments into one score.
Mention rateoutputs naming the business ÷ valid tests
Direct-citation rateoutputs linking the business site ÷ valid tests
Accuracy rateaccurate scored claims ÷ claims reviewed
05 / Decision rules
Know what each signal permits you to conclude
Thresholds vary by site size, demand, authority, and promotion. These rules identify the kind of evidence needed before acting; they do not promise traffic by a certain day.
| Observed signal | What it supports | Next action | Do not conclude |
|---|---|---|---|
| Live inspection fails | A current fetch or indexability problem exists. | Fix the specific access, render, robots, status, or directive issue; retest live. | That waiting longer will repair a deterministic blocker. |
| Live test passes; URL not indexed | The current page appears eligible, while the indexed state is absent or older. | Check discovery, canonical choice, duplication, usefulness, and last crawl; request indexing once after a material fix. | That the live test guarantees selection for the index. |
| Sitemap is successful; coverage is low | Discovery input is readable, but index selection differs from the intended set. | Group exclusions by reason and template; inspect representative URLs. | That resubmitting the same sitemap is a complete solution. |
| Impressions rise; clicks do not | Search visibility is growing, but result choice or intent fit may be weak. | Inspect page-query pairs, result appearance, title clarity, and whether the page fully answers the query. | That one sitewide CTR benchmark applies to every query. |
| Position changes on few impressions | The sample is volatile and incomplete. | Keep recording; wait for a larger, comparable set before rewriting the page. | That a single average or manual search is a stable rank. |
| Traffic rises; conversions stay flat | More people arrive, but the audience, offer, path, or measurement may be misaligned. | Verify event collection, segment by landing page and source, then inspect the highest-volume qualified path. | That SEO succeeded or failed without a business outcome definition. |
| AI mentions rise; citations do not | The entity may be recalled without the site becoming a direct source. | Review the accuracy of the mention and strengthen first-party, citable evidence for the tested questions. | That an uncited name mention sends traffic or proves source authority. |
06 / Evidence standard
Make the record auditable six months later
A useful log preserves raw evidence, interpretation, and the intervention as three separate things. DarsLab's 90-day checklist is a useful broad sequence; this tracker adds a fixed measurement contract and durable evidence structure. [10]
Every intervention needs seven fields
- Timestamp and owner
- Observed evidence
- Hypothesis
- Exact URL or template scope
- Change made
- Expected signal and decision date
- Keep, revert, revise, or inconclusive
07 / FAQ
New website SEO checklist FAQ
Direct answers for the moments when an early report looks empty, delayed, or contradictory.
How long does it take Google to index a new website?
There is no guaranteed indexing deadline. Google says crawling can take from a few days to a few weeks and that a crawl request does not guarantee indexing. Use the seven checkpoints to distinguish normal delay from a technical blocker.
Should I request indexing every day after launch?
No. Request indexing for a small number of important new or materially changed URLs, then wait and inspect the result. Google states that repeatedly requesting a recrawl for the same URL will not make it crawl faster.
Does a successful sitemap mean every page will be indexed?
No. A successful sitemap means Google could fetch and parse the submitted file. A sitemap helps discovery, but Google does not guarantee that every listed URL will be crawled or indexed.
How many pages should a new website have indexed?
Compare indexed pages with the number of canonical pages that you intentionally want in search, not with every URL the platform can generate. Redirects, duplicates, filtered URLs, noindex pages, and private pages usually should not inflate the target.
When should I act on Search Console query positions?
Act on a query-page pair after it has enough impressions to show a repeatable pattern and the query matches the page's purpose. Average position varies by search context, so one manual search or one early impression is not a reliable trend.
How do I make AI mention tests repeatable?
Freeze a prompt bank before the first test, reuse the exact wording and platform set, record account and location conditions, and score brand mention, direct citation, and factual accuracy separately. Save the output or a screenshot for every run.
This protocol was published September 1, 2026. Search and Analytics definitions were checked against current Google documentation. DarsLab was reviewed as a recent example of a broader post-launch sequence, not as evidence for Google behavior.
The milestone days are a measurement cadence, not a guarantee that a new website will be crawled, indexed, ranked, cited, or converted by a certain date. Site size, demand, quality, links, technical state, promotion, location, and competition change the result.
The embedded tracker reads and writes only local browser storage and creates a client-side CSV. It does not connect to Google Search Console, Google Analytics, an AI platform, or Teqri's servers.
Sources
- Google Search Central: How Google Search works — crawling, indexing, serving, and the absence of an indexing guarantee.
- Google Search Central: Ask Google to recrawl your URLs — recrawl methods, timing, quotas, and sitemap use.
- Search Console Help: URL Inspection tool — indexed-versus-live views, canonicals, discovery fields, and index requests.
- Search Console Help: Sitemaps report — submission state, last read, discovered URLs, and parsing errors.
- Search Console Help: Page indexing report — report scope, sitemap filters, exclusions, and individual-URL diagnosis.
- Search Console Help: Impressions, position, and clicks — metric definitions and average-position behavior.
- Search Console Help: About Search Console data — new-property timing, two-to-three-day reporting lag, time zones, samples, and privacy limits.
- Search Console Help: Performance dimensions and groupings — query anonymization, truncation, and canonical URL aggregation.
- Google Analytics Help: Mark events as key events — defining, marking, and reporting important business actions.
- DarsLab: Your 90-Day Post-Launch Website Checklist — a recent broad post-launch sequence used as contextual comparison.
Keep the evidence moving
Need someone to own the website after launch?
Teqri builds, hosts, maintains, and supports business websites—and keeps the post-launch measurement work tied to real decisions.