Skip to content

Track your brand
in Google AI Overviews.

Track whether Google shows an Overview, which brands it names and which links it exposes. Hold query, market and device controls fixed across repeated runs.

Best running shoes for beginners?Fixed query · US · desktop
Oct 2Fri
  1. 1Brooks
  2. 2Adidas
  3. 3New Balance
  4. 4PUMA
  5. 5Nike
Oct 3Sat
  1. 1Brooks
  2. 2New Balance
  3. 3Adidas
  4. 4PUMA
  5. 5Nike
Oct 4Sun
  1. 1Brooks
  2. 2Adidas
  3. 3New Balance
  4. 4Nike
  5. 5PUMA
Oct 5Mon

No Overview
Valid observation

Oct 6Tue
  1. 1Brooks
  2. 2Adidas
  3. 3Nike
  4. 4New Balance
  5. 5PUMA
Oct 7Wed
  1. 1Brooks
  2. 2Adidas
  3. 3Nike
  4. 4PUMA
  5. 5New Balance
Oct 8Today
  1. 1Brooks
  2. 2Adidas
  3. 3Nike
  4. 4New Balance
  5. 5PUMA

Google AI Overviews now names you 3rd

Nike rankings over 7 days
2

One query needs more than one check.

The same query can show a different Overview, new sources, or no Overview at all. One spot check can’t show where you stand. A daily record can.

Best running shoes for beginners?Checked three times, Oct 8
Nike 2nd

Most new runners do well in the Brooks Ghost 16. The Nike Pegasus 41 is a close second: cushioned, durable and easy to find.

No Overview

Standard results appear for this query. No AI Overview is shown. Keep the observation in your trigger-rate history; it has no answer or citations to count.

Nike 4th

Try the Brooks Ghost 16, the Adidas Supernova or the New Balance 880. For a firmer ride, the Nike Pegasus 41 works.

Trakkr checks every morning and keeps every result, so your place comes from weeks of observations, not one search.

Google AI Overviews

Best running shoes for flat feet?

For flat feet, most runners do best in a stable shoe with firm arch support. The Brooks Adrenaline GTS 24 is the usual pick. The Nike Structure 25 is a close second: a softer ride, and it comes in wide. For longer runs, try the New Balance 860v14.

runrepeat.com, reddit.com, youtube.com and 1 more

Every answer,
read for you.

Trakkr reads each answer the way you would: where it puts you, what it says about you and which pages it read. Then it keeps that reading, day after day.

See where Google AI Overviews puts you.

Every assistant ranks you its own way, so Trakkr keeps a separate record for each. Today Google AI Overviews is where you’re strongest.

Google AI OverviewsYour best of 8 assistantsYour best of 8
68Score in Google AI Overviews
3rd was 5thof 40 brands
34% +13of answers name you
Your place in Google AI Overviews, by questionToday
best marathon racing shoes1st
best running shoes for flat feet2nd
best running shoes for beginnersup from 5th3rd
best stability running shoePUMA passed you3rd
running shoes for wide feetNot named
Revenue from Google AI Overviews visitors, last 30 daysLast 30 days, 53% of your AI revenue
$6,841

Know why Google AI Overviews
changed its mind.

When an answer moves, Trakkr sets it beside the day before, finds the page Google AI Overviews started reading, and drafts the fix.

See how fixes work
Oct 5Before

Best stability running shoe?

  1. 1Adidas
  2. 2Nike
  3. 3PUMA
  4. 4Brooks
runrepeat.com/guides/stab…
Published Oct 5

The best stability running shoes, tested

  1. PUMA ForeverRun Nitro
  2. Adidas Adistar CS 2
  3. Brooks Adrenaline GTS 24

Doesn’t name Nike

Oct 6After

Best stability running shoe?

  1. 1PUMArunrepeat
  2. 2Adidas
  3. 3Nike
  4. 4Brooks

Get Nike onto runrepeat.com’s guide

Google AI Overviews cites it in 3 of 4 answers. Trakkr drafted the pitch.
Draft outreach

AI Overview tracking: the answer and the evidence

Practical guide and recurring tracker.

AI Overview tracking measures whether an Overview appears for a query, which brands and competitors it includes, which URLs it exposes, and what changes between controlled runs. Google Search Console now adds a Generative AI performance report with combined AI Overview and AI Mode link impressions by page, country, date, and device. Trakkr adds repeat query-level monitoring, answer presence, brand and competitor inclusion, source URLs, and a comparable history.

Setup preview before trial · 14-day free trial · Cancel anytime

What the product records

Clearly labelled sample data.

See the result before you start a trial

Each row keeps the query, controlled run, observed brand position, and exposed source together. The example below uses illustrative Nike data and makes no claim about Nike's real visibility.

Overview presence
44%+9 pts
Brand mention rate
31%+6 pts
Citation rate
21%+5 pts
Competitor presence
38%−3 pts
QueryOverviewBrand and positionExposed source
best running shoes for marathon trainingPresentNike · #1nike.com/running
most cushioned running shoes for beginnersPresentAdidas · #1runnersworld.com
are carbon plate shoes worth itPresentNike · #2nike.com/vaporfly
best shoe brand for wide feetPresentNew Balance · #1review publisher
nike vs adidas for everyday wearPresentNike · #1comparison publisher
durable trail running shoes 2026PresentNot namedirunfar.com
running shoes with removable insolesNot shownNo answer to countNo Overview source set

Coverage rates describe a constructed 30-day sample, not only the displayed rows.

Answer evidence

United States · English · desktop · same query

Best running shoes for marathon training

For marathon training, Nike offers cushioned daily trainers and lighter race-day shoes. Choose a model that fits your mileage and try it on before buying.

Exposed sourcenike.com/running

Read the change

Overview
Present
Nike mention
First named
Direct Nike citation
Present

Keep this dated baseline. A mention and a direct link are separate signals.

Constructed sample, not a live Google answer or customer data. The change does not establish its cause.

Google's report

What it includes and omits.

The Generative AI performance report, precisely

Google says the dedicated report reached worldwide rollout on 31 August 2026. Your site still needs enough impressions for data to appear. It counts organic-link impressions from AI Overviews and AI Mode together, with text and multimodal search filters. It is not a replay of the generated answer.

Dated anomaly: Google reports that a logging error from 13 to 17 August 2026 caused a decrease in Generative AI Search report impressions. Google classifies it as a data logging issue, so annotate that interval instead of treating it as a durable visibility loss. In its 21 August update, Google says the missing data was restored and the report now reflects complete metrics.

Generative AI performance reportSearch Console

It includes

  • Organic-link impressions from AI Overviews and AI Mode combined
  • Page, country, date, and device dimensions
  • Hourly, daily, weekly, and monthly date views
  • Chart and table exports using the report interface

It omits

  • A query dimension in the dedicated report
  • A split between AI Overviews and AI Mode
  • The answer text and whether your brand was named
  • Competitor presence and the full cited-source set
  • A dedicated Generative AI report in the Search Console API, based on Google's current report documentation

Current comparison

Five ways to measure the surface.

Google and Trakkr answer different questions

Search Console tells you what happened to your site's measured Search visibility. Recurring query tracking tells you what appeared in the answer surface. Keep ordinary rank tracking too, because the blue links can move independently.

MeasureGoogle Generative AI reportOrdinary Search ConsoleClassic rank trackerManual checkingTrakkr recurring tracking
Query visibilityNo query dimension in the dedicated reportQueries, but AI traffic is mixed into WebFixed keywords and organic positionsThe query you typeFixed buyer-query set with each run
Cited URLPage receiving an impression, not its role in an answerLanding page, not AI citation contextRanking URL in standard resultsVisible if you save itExposed supporting URL saved per query
Brand mentionNoNoNoYesYes, including uncited mentions
Competitor presenceNoNoOrganic competitors onlyYesYes, inside the observed Overview
HistoryDated impression trend after report availabilityDated Search performance trendOrganic rank historyOnly if you record every checkRecurring answer, mention, source, and competitor history
Market controlCountry and device dimensionsCountry and device filtersUsually country, locale, and deviceHard to hold constantSaved market, language, and device controls
RepeatabilityPassive aggregated loggingPassive aggregated loggingRepeatable for blue-link ranksLow unless tightly documentedScheduled reruns of the same controlled query set
Main limitationCombines AI Overviews and AI Mode; impressions onlyNo separate AI-feature breakoutDoes not capture the generated answerSlow, volatile, and easy to personaliseObserved changes do not prove why Google changed

Sources checked 12 October 2026. Product capabilities and Google reporting can change.

Tracking workflow

Eight visible, repeatable steps.

A complete AI Overview tracking workflow

Start with a small set of buyer queries. Keep the controls fixed, save the sources, and review repeat observations before choosing an action.

  1. Define the query set

    Start with category, comparison, problem, and purchase questions a buyer would genuinely ask. Keep the exact wording in a query inventory.

  2. Fix markets and devices

    Assign a country or market, language, and device to every query. Do not compare runs with different controls.

  3. Capture a baseline

    Run the full set once and date it. A missing AI Overview is a result, not a failed check.

  4. Record Overview presence

    Save whether an Overview appeared and preserve the observed answer or evidence needed for later comparison.

  5. Save mentions and sources

    Record the brand, competitors, exposed URLs, source domains, and their positions in the answer.

  6. Classify gains and losses

    Use a fixed label such as Overview gained, Overview lost, citation gained, citation lost, source switched, brand gained, or competitor gained.

  7. Connect evidence to an action

    Link a material loss to the affected page or third-party source, then assign an owner and a small evidence-led action.

  8. Rerun consistently

    Repeat on the same schedule with the same controls. Review the run history before treating one observation as a trend.

Lost citation diagnosis

Evidence before edits.

A lost citation is a signal, not a diagnosis

The page, result set, retrieval path, answer model, and source mix can all change. A single before-and-after check establishes correlation only. It does not show that a page edit caused the loss.

  1. Confirm the observation

    Match the exact query, market, language, device, and run method. Save the old and new evidence.

  2. Check persistence

    Repeat the run on the planned schedule. Distinguish a one-run change from a stable pattern.

  3. Compare the source set

    Identify which URL replaced yours, whether the answer angle changed, and whether the new source supports a different claim.

  4. Audit the affected page

    Check indexing, snippet eligibility, rendering, changed facts, removals, canonicals, and material edits around the loss.

  5. Look for wider evidence

    Review ordinary rankings, Search Console, other queries, and other cited pages. One query should not carry the whole diagnosis.

  6. Choose the smallest action

    Refresh a weak claim, fix a technical issue, improve a missing comparison, or pursue the third-party source that now shapes the answer.

Google says pages need to be indexed and snippet-eligible, with no extra AI-specific technical requirement. Read Google's AI features documentation and AI search optimization guide before assuming a special file or markup will restore a citation.

Audit template

Free CSV for Sheets, Excel or Numbers.

Run the workflow in a spreadsheet

The CSV opens in Google Sheets, Excel, or Numbers. It includes field definitions plus two clearly labelled illustrative rows, so you can copy the structure without inventing a measurement method.

Query and controls
Exact query, market, language, device, run date, and next run date.
Observed result
Overview present, brand mentioned, brand position, and competitors present.
Source evidence
Cited URL, cited domain, and the saved screenshot or result reference.
Change class
A controlled label for a gain, loss, switch, or stable result.
Action
The page or source action, owner, due date, and notes.
Worked example, illustrative only
RunOverviewBrandSourceChange classNext action
Baseline · 19 AugPresentNike · #1nike.com/runningBaselineRerun unchanged controls
Follow-up · 26 AugPresentNike · #2Comparison publisherCitation lost · source switchedCollect two more runs before editing

Questions,
answered.

The ones we hear most.

Talk through your setup

AI Overview tracking records whether Google shows an AI Overview for a fixed query, which brands it names, which source URLs it exposes, which competitors appear, and how those observations change across repeated runs.

Ask anything elseAsk anything else

The Agent answers from our docs, methods and pricing, and links to the source.

Know what Google AI Overviews
tells your buyers.

Add your brand and the questions they ask. Trakkr asks Google AI Overviews first thing tomorrow, and every morning after.

  • 14-day free trial
  • Cancel anytime
Best running shoes for beginners?Fixed query · US · desktop
Oct 7Wed
  1. 1Brooks
  2. 2Adidas
  3. 3Nike
  4. 4PUMA
  5. 5New Balance
Oct 8Today
  1. 1Brooks
  2. 2Adidas
  3. 3Nike
  4. 4New Balance
  5. 5PUMA
Oct 9Tomorrow
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5