AUTOMATE PLANET · IMPLEMENTATION NOTE · SEARCH EVIDENCE
What Search Visibility Can Prove After a Website Rebuild
A search report can document visibility and clicks during a stated period. It cannot by itself establish new customers, revenue or which website change caused an improvement. Keep the report’s dates and filters, separate overlapping metrics, show the full trend and connect search activity to verified inquiries before making a business-outcome claim.
Visibility is a starting point. Outcomes need their own evidence.
Included in Web reporting
Separate the work delivered from the result observed
For Family Insurance, Automate Planet rebuilt the website, upgraded its SEO, developed pages for traditional and AI-assisted search, and added local content targeting Minnesota, Utah and Nevada. The owner also described an automated workflow producing an insurance-focused page daily. The site went live roughly three months before the supplied screenshots.
Those are descriptions of implementation scope. The screenshots provide a different kind of evidence: what Search Console displayed during the selected reporting period. They do not isolate the effect of the rebuild, local content, publishing frequency or any individual page. We cannot infer performance in each target state from an overall property total.
The implementation lesson is to keep two records together: a dated change log describing what you shipped, and a measurement record describing what happened afterward. A plausible connection is worth investigating, but timing alone does not establish causation. A daily publishing capability is also not proof that publishing every day improves rankings or earns citations.
Read the snapshot without adding overlapping totals
The client-provided screenshots captured on October 20, 2026 show the following values with a three-month filter selected. These are rounded display values from that historical snapshot, not newly collected performance data.
| Displayed measure | Value | What this supports | What it does not establish |
|---|---|---|---|
| Google Search Web impressions | 35.3K | Reported search exposure during the selected period. | 35,300 unique people, inquiries or customers. |
| Generative-AI-feature impressions | 8.21K | Reported exposure within Google’s generative AI features. | 8,210 separate citations, recommendations or additional visits. |
| Google Search clicks | 88 | Reported clicks from Google Search. | 88 unique leads or sales, or a verified split between AI and other search results. |
| Trend in the supplied charts | Low early visibility; September peak; later decline | Visibility varied across the period. | Continuous growth, a stable current run rate or a proven reason for the decline. |
Google’s AI-feature guidance says AI-feature activity is included in Search Console’s Web reporting. The 8.21K should therefore not be added to 35.3K. That would count overlapping visibility as though the two totals represented separate audiences.
Under Google’s metric definitions, impressions and clicks describe search-result exposure and interaction, with counting rules that depend on the result type. They are not a count of customers. We also cannot assign all 88 clicks to AI features from these screenshots.
A defensible summary is: the supplied reports document search and AI-feature visibility during the period surrounding the website work, with a peak followed by a decline. The reports alone cannot establish a lead increase, a return on investment or the contribution of any particular change.
A peak and a decline belong in the same story
A cumulative total can look encouraging while the latest part of the chart is weaker. In the Family Insurance snapshot, visibility was low early in the reporting window, rose to a September peak and later fell. Describing only the peak would omit information a business owner needs for the next decision.
Do not extend a high day into a monthly forecast. Do not describe “growth from zero” as a percentage when a reliable starting value and comparable periods are unavailable. Here, the approximate launch timing and screenshot labels do not give us a clean, separate before-and-after experiment.
Our proposed review method is to examine the complete time series, then compare specific page groups and types of searches. Record publishing dates, page revisions, campaigns and other known changes alongside the chart. Those annotations generate questions; they do not automatically explain a rise or fall.
For this project, the next useful evidence would be exports with exact dates and filters, page-level and query-level detail where available, a dated implementation log, and verified inquiry records. Without those records, the reason for the later decline remains unresolved. This article does not claim that we diagnosed or fixed it.
Use an evidence worksheet before writing a results claim
Build the evidence before the claim.
This is an original proposed worksheet for a website reporting review. It is not a completed audit of Family Insurance. Fill it out using authorized records, and leave unavailable fields marked “not measured” instead of treating them as zero.
| Record to collect | What to write down | Decision it supports |
|---|---|---|
| Report identity | Property, exact start and end dates, search type, filters and export date. | Whether two reports describe the same scope. |
| Implementation log | Launch date and dated changes to page groups, content, navigation or forms. | Which interventions were present during each period. |
| Visibility detail | Impressions and clicks by relevant page group; query detail where available. | Whether useful service pages gained exposure rather than unrelated content. |
| Comparison period | Equal-length windows with known seasonal, campaign and operating differences noted. | Whether a before-and-after comparison is interpretable. |
| Inquiry definition | What qualifies as an inquiry and how duplicate submissions or repeat contacts are handled. | Whether the reported lead total represents distinct opportunities. |
| Verified outcomes | Confirmed inquiries, appointments or purchases, with recorded source where available. | Which business outcomes can actually be connected to the website. |
| Missing evidence | Unattributed outcomes, unavailable exports, incomplete tracking or unknown changes. | Which conclusions should remain tentative. |
Use aggregate outcome counts in a public report. The worksheet does not require sharing a customer’s name, phone number, email or conversation. Keep detailed records in the business’s approved systems and publish only the evidence the client has authorized.
Before calculating a change, state the unit: impressions, clicks, distinct qualified inquiries or confirmed purchases. Preserve the same definition across both periods. If the reporting setup changed, describe that break instead of presenting incompatible totals as a continuous trend.
Choose the next improvement from the missing step
The following table is a planning method, not a diagnosis of this client’s site. It turns a reporting observation into a focused check while keeping possible explanations open.
| Observation | Useful next check | Evidence needed before a stronger claim |
|---|---|---|
| Impressions rise, clicks do not | Review the pages and searches gaining exposure; compare the offer with the visitor’s likely question. | Matching page and query detail, with dates and filters. |
| Clicks rise, inquiries are unmeasured | Check that the inquiry path works and distinguish a submitted inquiry from a button click. | Confirmed submissions and an agreed source-recording method. |
| A visible peak is followed by a decline | Compare affected page groups and dates with the change log. | Full-period exports and relevant changes; the screenshot alone cannot identify the cause. |
| AI-feature visibility appears | Keep its platform-defined metric separate from visits and business outcomes. | A dated report identifying the feature and metric; explicit outcome data for any sales claim. |
| More content is published | Review whether each page answers a distinct customer question and provides supported detail. | A page inventory and actual usefulness review; publishing frequency is not a result. |
One practical review can end with a short decision record: the observation, the question it raises, the next check, the person responsible and the evidence still missing. For example, “Search exposure increased, but inquiry attribution is unavailable; verify the submission path and define source reporting before making a lead claim.” This is suggested wording, not a finding from a new client test.
For the wider discovery strategy, read our SEO and AI search guide. For outcomes beyond website discovery, use the business-results measurement guide. If you need the site, content and inquiry path implemented together, compare business websites, SEO and AI search services, or review packages and deliverables.
Common questions
Do impressions tell us how many people visited the website?
No. Impressions describe exposure under the search platform’s counting rules. Clicks are a different metric, and neither figure by itself establishes distinct customers, qualified inquiries or sales.
Can we add AI-feature impressions to total Web impressions?
Not for the Google reports discussed here. Google documents AI-feature activity within Web reporting. Adding the supplied 8.21K AI-feature impressions to the 35.3K Web total would double-count overlapping visibility.
Does the Family Insurance report show 8,210 AI citations?
No. The supplied screenshot labels the figure as impressions in generative AI features. It is not a verified count of distinct citations, recommendations or unique people.
Did the website rebuild cause the reported visibility?
The reports document visibility during the period surrounding the website work. They do not isolate the effects of the rebuild, local content or publishing workflow. A separate baseline, exact comparison dates and supporting records were not supplied.
Why include the later decline in a case study?
It is part of the supplied evidence and affects the interpretation of the total. Showing the full trend helps distinguish a temporary peak from sustained performance. The available screenshots do not establish why the decline happened.
What would make the next report more useful?
Exact-date exports, consistent filters, a dated implementation log, page and query detail where available, and verified inquiry or purchase records. Clearly state missing attribution and use the same outcome definitions across comparison periods.
Sources and editorial approach
- Automate Planet project account: Family Insurance website and search visibility
- Google Search Console Help: What are impressions, position, and clicks?
- Google Search Central: AI features and your website
How this page was prepared. AI-assisted writing based on documented Automate Planet work and the sources linked here. This is our implementation account, not an independent evaluation.
Family Insurance figures come from client-provided Search Console screenshots captured October 20, 2026 with the three-month filter selected. Totals are rounded; exact comparison dates, a separate baseline and conversion records were not supplied. This is a historical snapshot, not a current performance audit or proof that one intervention caused the results.
Make the next step work for your business.
Compare the scope and deliverables. Our AI can help you choose a package before you buy.
Find my packageExplore all resourcesYour next chapter.
Starts here.
Tell us what you want to make easier. We’ll help connect the right people, processes, and AI.
Make time for what’s next.
A one-hour consultation to explore your goals, workflows, and opportunities.