How we measure results against a baseline
Draft of 21 October 2026, prepared by AnswerShare. Every figure below gives its date and the stored record it comes from. Where we have no measurement yet, we say "not yet measured". A missing figure is never shown as zero.
The measures
| Measure | What it is |
|---|---|
| GEO score | How accurately and consistently AI engines describe and recommend Texas First Rentals, scored 0 to 100 by each of five engines (ChatGPT, Claude, Gemini, Perplexity and Grok) on the same rubric. We report the median, with an outlier dropped when one engine is far from the rest. |
| SEN | Brand sentiment: how favourably the same five engines describe Texas First Rentals. We report the median. |
| RR, Relevance Ratio | The share of the visible text on your home page that is page content rather than navigation and footer. Higher is better. |
| RTC, Retrieval Token Cost | What a page costs an AI crawler to read: the page's size in tokens, multiplied by the seconds to receive it, divided by the characters of useful content. Lower is better. |
| AI crawler visits | Requests from each AI crawler on the rule's list, by day and by crawler. |
| U/T, user-intent share of AI crawler traffic | The share of AI visits made while an assistant is looking for information to answer a person's question. We identify these visits by the type of AI crawler. It is evidence that AI is coming to your site for information while preparing a response. It does not measure whether that response names you or cites your website. |
| Answer share | For a fixed, versioned set of questions your customers ask, the share of AI engine answers that name Texas First Rentals or cite your site. |
| AI-referred sessions | Visits to your site from AI assistants (for example chatgpt.com, perplexity.ai, claude.ai, gemini.google.com and copilot.microsoft.com), from your Google Analytics 4 property. |
Today's baseline
Read on 21 October 2026, read-only, from AnswerShare's measurement database. The most recent full measurement of your site is from 6 October 2026. Every one of these runs measured your site as it is today, with no AI layer in place.
| Measure | Value | Measured | Source record |
|---|---|---|---|
| GEO score | 49.5 | 2026-10-06 23:07 UTC | comparison_scores row 0d5e398e-82ef-4d4f-a334-9ac99bf0f223, receipt 1fdfb57b-8655-4ea7-a2ad-fa8f116e29cf. Engines: Grok 49, Perplexity 47, Claude 31, Gemini 50, OpenAI 55 (Claude's score dropped as an outlier before the median). Prompt set 5.5.1, judge panel 2026-10-04.claude-sonnet-5-5, 5 of 5 engines valid. |
| SEN | 50 | 2026-10-06 23:07 UTC | Same row. Engines: Grok 65, Perplexity 0, Claude 5, Gemini 50, OpenAI 50. |
| RR | 0.5511 | 2026-10-06 23:07 UTC | Same row. Home page: 1,267 characters of content out of 2,299 visible. |
| RTC | 2.7807 | 2026-10-06 23:07 UTC | Same row. Home page: 67,753 tokens, 0.052 seconds to last byte, 1,267 useful characters. |
| Time to first byte, home page | 52 ms | 2026-10-06 23:07 UTC | Same row, median of a rapid series. |
| Pages in your sitemap | 497 | 2026-10-06 23:07 UTC | Same row. |
| Answer share | Not yet measured | n/a | No answer-share results and no prompt set exist for Texas First Rentals yet (checked 21 October 2026). The prompt set is built from your data and reviewed with you before the first run. |
| AI crawler visits | Not yet measured | n/a | No crawler records exist for your site (checked 21 October 2026). We have no view of your traffic until the rule is live. The only "before" source is your Vercel request counts, if your plan keeps them. |
| U/T | Not yet measured | n/a | Depends on AI crawler visits. |
| AI-referred sessions | Not yet measured | n/a | Needs Viewer access to your Google Analytics 4 property. Google Analytics keeps history, so the months before launch can be read once access exists. |
Every measurement of your site so far
Results move from run to run even when nothing on your site changes, because AI engines do not answer identically every time and response times vary. These are all five runs on prompt set 5.5.1. We read any later change against this spread, not against a single number.
| Measured (UTC) | Source row | Judge panel | GEO | SEN | RR | RTC | TTFB |
|---|---|---|---|---|---|---|---|
| 2026-09-26 18:41 | 27c2a91d | 2026-09-15.pplx-agent-sonar | 47 | 60 | 0.5513 | 2.8316 | 52 ms |
| 2026-09-29 15:43 | 4cdc0d13 | 2026-09-15.pplx-agent-sonar | 45 | 15 | 0.5513 | 2.8850 | 53 ms |
| 2026-10-01 19:07 | 428b88dd | 2026-09-15.pplx-agent-sonar | 49 | 55 | 0.5513 | 3.7398 | 68 ms |
| 2026-10-03 20:29 | 93f2ede1 | 2026-09-15.pplx-agent-sonar | 49 | 50 | 0.5513 | 3.4193 | 63 ms |
| 2026-10-06 23:07 | 0d5e398e | 2026-10-04.claude-sonnet-5-5 | 49.5 | 50 | 0.5511 | 2.7807 | 52 ms |
Across these runs GEO ranged from 45 to 49.5, SEN from 15 to 60 and RTC from 2.78 to 3.74. The 26 September run is the baseline we recorded when the work began. The "Now" figures in our 19 October email (RTC 3.4, about 60 ms) come from the 3 October run.
Before launch: completing the baseline
- One more full run in the preview week, on the instrument versions current then. We fix those versions (prompt set and judge panel) for every later run in the test.
- The first answer-share run, on a prompt set built from your Google Analytics and Search Console data, any keyword exports you share, and real customer questions, and reviewed with your lead. The prompt set gets a version number and does not change during the test.
- RR and RTC for the five side-by-side pages, as AI crawlers receive them today, using the same metric code as the home page.
- Google Analytics and Search Console history, read once access exists, back 13 months by month and 90 days by day.
- Your Vercel request counts by AI crawler for the last 30 days, if your plan keeps them. If not, crawler visits are reported from launch onward, and we say so in every readout. Our question about your Vercel log retention is in the plan.
After launch: the same measures, the same instruments
| Measure | How it is taken after launch | First reported |
|---|---|---|
| GEO and SEN | Same prompt set and same judge panel version as the pre-launch run. If an instrument has to change during the test, we re-run the baseline on the new version and compare like with like, never across versions. | 30 days after enable |
| RR and RTC | The home page and the five side-by-side pages, fetched through the live rule as an AI crawler receives them, with the same metric code. | Handover call |
| AI crawler visits by type | Counted at AnswerShare for every request the rule sends us, by crawler and by day. Your Vercel request counts give an independent check. | Handover call, then daily on your dashboard |
| U/T | User-triggered AI visits divided by all AI crawler visits, with the crawler types listed on the rule page. | Handover call |
| Answer share | Same prompt set version, same engines, same scoring. | 30 days after enable |
| AI-referred sessions | Google Analytics 4, aggregate reports only, by assistant, compared with the same weeks before launch and the same period a year earlier. | 30 days after enable |
| Google-side search | Search Console clicks and impressions. The rule does not touch Google Search, so this series shows change that has other causes. | 30 days after enable |
Readouts are at the handover call after the 7-day soak, and at 30 and 90 days after enable [CONFIRM: cadence].
Separating our effect from other changes
- Crawler-side measures are direct. AI crawler visits, U/T, RR and RTC of the AI version exist only on requests the rule sends us, so they measure the rule itself.
- Outcome measures can move for other reasons. Answer share, GEO, SEN and AI-referred sessions can also move because of your own site changes, engine updates or the season. We keep a dated log of every such change we know of: your deploys (from the webhook events), the rule's enable date, the FAQ's publication, and engine updates. Each readout reads results against that log.
- We compare against the spread, not one number. A change smaller than the run-to-run spread above is not reported as an effect.
- Google's side is a comparison series. Google Search, AI Overviews and AI Mode keep reading your normal site, so their trend shows what happens without the AI layer.
How results are recorded
Every run keeps the complete response from each engine, with its run and attempt IDs, timestamps, instrument versions and a SHA-256 hash. A retry never overwrites an earlier attempt or inflates a count. If a capture fails, the readout says the run is incomplete and shows what was recovered; it never fills the gap with zero. We read your analytics with read-only access and aggregate reports only: no revenue, transactions, conversions, funnels or user-level data.