How to measure answer engine optimization when there's no last click

Why doesn't AEO show up in your analytics?

Pew Research (March 2025) found an 8% click-through rate to any website when an AI summary was present in search results, versus 15% without one, and just 1% of users clicked through to a cited source specifically. GrowthSRC Media's July 2025 study of more than 200,000 keywords across 30-plus client sites found the same shape at larger scale: average organic click-through rate for Google's #1 position fell 32% year over year (28% to 19%) as AI Overviews rolled out, with the #2 position falling 39%. Two independent, dated studies pointing the same direction is a stronger basis for this argument than either one alone, and it points to a structural gap between where the value gets created and where your analytics stack is built to look for it.

We track a set of bundled AEO and affiliate queries that sit at position 2 across engines, with roughly 170 impressions across 15 query variants over a 90-day window, and zero recorded clicks. On a pure click-attribution model, that program reads as dead. We don't think it is, but it's worth being honest about why: zero clicks at a strong position is consistent with the answer already being resolved upstream, and it's also consistent with a snippet that isn't earning the click or a query set moving through ordinary ranking noise. Search Console data alone can't tell those apart. The only way to know which explanation applies is to directly query the answer engines on these terms and check whether the brand is cited - the audit this piece walks through next. Treat a cluster like this as a strong candidate for that audit, not as confirmed citation activity until you've actually checked.

This shows up across every vertical we work in - insurance, banking, commerce, home services - but the mechanism is the same everywhere. When an AI engine answers a comparison or "best X" query directly and names your brand as part of that answer, the user's job is often done. There's no next step that generates a session in Google Analytics. The value transferred - brand awareness, consideration credit, a name lodged in a buyer's shortlist - and your reporting has no line item for it.

What can you measure in the first 90 days?

Three things are trackable almost immediately, and none of them require waiting on a sales cycle.

First, citation presence. Run your priority queries through AI Overviews, ChatGPT, Perplexity, and Copilot on a weekly cadence and record, query by query, whether your brand is named or cited in the answer. This is binary and simple to log: yes, no, and which engine.

Second, share of voice. Once you're tracking citation presence across a query set, roll it up into a percentage: what share of category answers name you versus your competitors. This is the metric that tells you whether you're winning the category conversation, not just present in it.

Third, citation source mix. When you are cited, what content is doing the citing work - your own site, a partner site, a review property, an editorial piece? This is where a citation audit connects back to your affiliate and partner footprint. Across our own managed programs, we run roughly 550 total partnerships, with about 190 of those productive - generating measurable activity - in the past 30 days, a 34.5% productive rate across our full managed partner base. That figure tells you about partnership productivity, not citation presence; we haven't run a source-mix audit against it, so we won't claim the two automatically overlap. What mapping your own citation source mix will tell you, once you run it, is whether your affiliate program is functioning as an AEO distribution channel - which our own program-management experience suggests is common, since the same content that earns affiliate placement is often built to the sourcing and structure standards AI engines pull from - but treat that as an operational hypothesis worth testing with your own audit, not a conclusion the productivity number proves on its own.

Building the citation query set

None of the above works without a query list, and building that list is the part most teams skip or overcomplicate. Here's how to do it in an afternoon.

Start with Search Console data you already have. Two segments are your natural seed list. The first is striking-distance queries, terms where you're ranking positions 8 through 20, since these are queries where an AI engine is plausibly already considering your content but you haven't earned the citation yet. The second is high-impression, low-click or zero-click queries - the ones GSC shows real search volume against but almost no traffic for. Those are worth including precisely because they're ambiguous: some are genuinely being absorbed by an AI answer, some are just underperforming for ordinary reasons, and the query set - not Search Console alone - is how you find out which is which. That's exactly the profile of the bundled AEO and affiliate cluster referenced above: 170 impressions, 15 variants, zero clicks, sitting at position 2, still unconfirmed until it's actually been checked against the engines.

Second, add category-intent questions that don't originate from your own keyword data at all: "best [category] for X," "[you] vs [competitor]," "how do I choose a [category] provider." These are the question formats AI engines are built to answer directly, and they're often missing from a paid-search or SEO keyword list because they were never worth bidding on or ranking for in the old model.

Third, add the queries your sales team already hears. Every AE and every intake call surfaces a handful of phrasings prospects use when they're comparing options: "why should I use you instead of," "is [category] worth it," "what's the difference between." Sales teams sit on this language and rarely hand it to marketing. Ask for it directly.

On sizing: a query set in the 30 to 60 range is the workable middle ground for most single-category brands. Below 15 queries, single-query noise (an engine flips an answer for one query one week) will read as a program shift when it isn't. Above 100, manual tracking quietly stops happening every week, which defeats the purpose of a leading indicator. Pick the segment of your category that matters most commercially and build the list around that first, then expand.

Refresh the query set on a cadence, not just when someone remembers to. Quarterly is the right default: category questions shift as competitors launch, as your own product line changes, and as AI engines themselves change what they consider a "best X" answer worth generating. Trigger an off-cycle refresh any time a new competitor enters your paid or organic top five, or any time your sales team flags a new comparison question showing up in calls.

Keep the list itself simple: a single tab with columns for the query, its source (GSC striking-distance, GSC zero-click, category question, sales-sourced), and the engine it's tracked against. That tagging matters at quarter-end, when you want to know whether gains are concentrated in queries you already ranked for or in the ones you added deliberately, the difference between winning ground you already had and actually expanding into new territory.

Which demand signals corroborate it?

Citation metrics tell you the input is happening. The next layer tells you whether it's translating into demand, and this shows up on a two-to-four month horizon.

Branded search lift is the cleanest of the three. Pull branded query volume from Search Console and watch it against your citation presence trend. If share of voice is climbing and branded search volume is flat, something's off, either the citations aren't reaching the audience that matters, or the query set doesn't match where your real buyers are searching.

Self-reported attribution is the second signal, and it's underused. Add "how did you hear about us" to intake forms and sales qualification, and add AI tools as an explicit option alongside search, referral, and social. Prospects increasingly answer with some version of "I asked ChatGPT" or "it came up when I was researching," and that's a direct corroboration of citation activity that no analytics platform will hand you on its own.

Third is a quality shift in inbound. Watch for buyers who arrive already mid-shortlist: they know your category positioning, they've already ruled competitors in or out, and they're asking narrower, more informed questions than a cold inbound lead typically asks. That shift in the character of your inbound - not just the volume - is a tell that people are arriving pre-educated by something other than your own site, and AI answers are one of the most likely sources.

When does AEO show up in revenue?

Months three through six is the typical window, and there are two separate mechanics driving that lag, plus the ordinary buying-cycle delay layered on top.

The first mechanic is engine-side: AI systems re-crawl and re-weight sources on their own schedule, not yours. You don't control when a citation you've earned gets reflected in the next answer generation, and that lag varies by engine.

The second is compounding corroboration. Branded search lift, self-reported attribution, and inbound quality shift each build gradually and each reinforces the others; none of them moves in a single clean step the way a paid click does.

On top of both, there's the buying-cycle lag that exists in every considered purchase regardless of channel. A buyer who sees your brand named in an AI answer in month one may not be in-market until month four.

None of this means last-click attribution stops working, it just means it's measuring a narrower slice than it used to. Across our own managed programs in the past 30 days, we tracked 3.4 million clicks against roughly 47,500 tracked orders. Last-click still captures the bottom of the funnel cleanly; it always has. What it can't see is the discovery layer sitting above it, the citation and consideration activity that happens before a buyer ever produces a trackable click.

The 3-tier measurement dashboard

The individual metrics above only work as an operating system if you stack them into tiers with different cadences, because each tier answers a different question and moves on a different clock.

Tier 1 is citation metrics - presence, share of voice, and source mix - updated weekly. This is your leading indicator. It tells you whether the program is doing the thing it's supposed to do, showing up in the answers that matter, before there's any way to know whether that's translating into pipeline.

Tier 2 is corroborating demand signals - branded search lift, self-reported attribution, and inbound quality shift - updated monthly. This is the bridge layer. It won't move every week, and it shouldn't be checked every week, but a month is enough time for a real trend to separate itself from noise.

Tier 3 is business outcomes - pipeline and revenue lift, tied back specifically to the content and query clusters driving citations - updated quarterly. This is the tier that answers the CEO's actual question - did this work - but it's also the slowest tier and the easiest one to misread in isolation.

The reason to run all three tiers together, rather than picking one, is that each tier alone produces a predictable failure mode. Watch only Tier 3 and you'll kill a working program at the 90-day mark because revenue hasn't moved yet, exactly the mistake pure click attribution invites, since rising Tier 1 presence is invisible to someone only checking quarterly revenue. That's the false negative. Watch only Tier 1 and you'll keep funding a program that's present in answers but never translating into demand, mistaking visibility for value. And watch Tier 2 or Tier 3 without Tier 1 as context, and you'll credit AEO for growth that came from somewhere else entirely, a seasonal lift, a new sales rep, a competitor stumble, with no leading indicator to check the timing against. That's the false positive. Three tiers, checked on three clocks, is what lets you tell "not working yet" from "not working."

How do you pay for value the click can't carry?

Performance pricing assumes a click exists to measure and pay against. Citation value structurally doesn't produce one, which means a program priced purely on clicks or conversions will always undercount what AEO is actually contributing, by design, not by accident.

That's why flat and hybrid fee structures now show up in most of our proposals for this work: a flat retainer, or a retainer against a percentage of program performance, whichever is greater. The flat component gets tied to citation KPIs directly - presence across your priority queries, share of voice trend, and whether citation-worthy content is kept fresh enough to maintain those citations as engines re-crawl. That's a different accountability structure than "did this produce a click," but it's a real one, and it's measurable using exactly the Tier 1 metrics described above.

We've written the fuller case for why an affiliate program specifically is the underlying AEO asset, since so much of the content that earns citations is the same content earning affiliate placement, in our pillar piece on affiliate as the AEO asset. If you're building a measurement plan, that's the companion read for the pricing model underneath it.

A realistic budget and resourcing model

The next question every CEO or CMO asks is what this actually costs to run, and who's supposed to run it. Here's the honest answer.

The weekly query-tracking cadence - running your query set through AI Overviews, ChatGPT, Perplexity, and Copilot and logging presence and citation source - is manual work, and a spreadsheet is genuinely fine to start. For a query set in the 30 to 60 range, budget roughly two to three hours a week for someone to run the queries, log results, and roll them into a share-of-voice trend. That's a task, not a role, at that scale; it fits inside an existing analyst's or SEO manager's week without a new hire.

Tooling becomes worth it once you cross a threshold, either your query set grows past 75 to 100 terms, or you're tracking consistently across more than two or three AI engines, or you need historical trend data further back than a spreadsheet comfortably holds. Below that threshold, tooling is often solving a problem you don't have yet. Above it, the manual version quietly stops happening every week, which defeats the point of a leading indicator.

On who owns it: in-house, it sits naturally with a marketing analyst or SEO function, since weekly Tier 1 tracking is close cousin work to rank tracking they may already be doing, and the monthly Tier 2 rollup slots into an existing reporting cadence. Run through an agency, this typically folds into program management rather than billing as a separate line: the same team managing your affiliate or paid channels absorbs the weekly and monthly work, with the quarterly Tier 3 review as a standing agenda item rather than a bespoke project.

That's also the resourcing logic behind the flat and hybrid fee structure discussed above. The flat component isn't priced against clicks, because clicks aren't the unit of work being performed. It's priced against the tracking cadence itself - the weekly query monitoring, the monthly demand-signal rollup, and the quarterly business review - tied to citation KPIs as the accountability mechanism. We keep specific figures out of proposals until we've scoped your query set and engine coverage, since the workload scales with category complexity, not with headcount or revenue size, but the resourcing tiers above are the honest starting framework before any number gets attached to it.

Frequently asked questions

What KPIs measure answer engine optimization? Citation presence, share of voice, and citation source mix are the leading indicators, tracked weekly. Branded search lift, self-reported attribution, and inbound quality shift are the corroborating layer, tracked monthly. Pipeline and revenue lift tied back to citation-driving content is the outcome layer, tracked quarterly. No single KPI answers the question alone.

How long does AEO take to show results? Citation presence is visible within 90 days. Corroborating demand signals typically emerge in months two through four. Pipeline and revenue impact typically shows up in months three through six, driven by engine re-crawl schedules, compounding corroboration, and ordinary buying-cycle lag.

Can you track AI citations automatically? Manually, yes, at small scale: running a query set through AI Overviews, ChatGPT, Perplexity, and Copilot on a weekly cadence and logging results in a spreadsheet works for query sets in the 30 to 60 range. Tooling becomes worth adding once your query set grows past 75 to 100 terms or you need engine coverage and historical trending a spreadsheet can't hold.

Does AEO measurement replace performance measurement? No. Last-click attribution still works for what it was built to measure, bottom-of-funnel conversion. Across our own managed programs we tracked 3.4 million clicks against roughly 47,500 tracked orders in the past 30 days, and that layer stays intact. AEO measurement adds a layer above it - the discovery and consideration activity last-click was never built to see.

How many queries should I track for citation presence? A range of 30 to 60 queries is the workable middle ground for most single-category brands. Fewer than 15 and you don't have enough of a sample to separate a real share-of-voice trend from single-query noise. More than 100 and manual tracking breaks down unless you've already added tooling. Seed the list from Search Console striking-distance and high-impression zero-click queries, add category "best X" and "vs" questions, and add the comparison language your sales team already hears from prospects.

Do I need special software to measure AEO? Not to start. A shared spreadsheet and a weekly hour or two running your priority queries through the major AI engines covers a query set in the 30 to 60 range. Add tooling once your query set outgrows that range, once you're tracking more than two or three engines consistently, or once you need historical trend data further back than a spreadsheet reasonably holds.

How is AEO measurement different from traditional SEO reporting? Traditional SEO reporting assumes a click follows a ranking. AEO measurement assumes the opposite - that a strong citation may produce zero clicks, per Pew Research's finding of 1% click-through to cited sources and GrowthSRC's documented decline in position-1 CTR - and builds its indicators - presence, share of voice - around whether you're named in the answer rather than whether traffic followed, confirmed through direct citation checks rather than click data alone. It's a parallel discipline, run alongside existing SEO reporting rather than instead of it.

Who on my team should own AEO measurement? The weekly citation tracking fits naturally with whoever already owns SEO or marketing analytics, it's close cousin work to rank tracking. The monthly demand-signal rollup belongs wherever branded search and lead-source reporting already live. The quarterly business outcome review should include whoever owns pipeline reporting, since that's the tier that closes the loop back to revenue. If an agency is running your program, this typically folds into existing program management rather than requiring a new internal hire. The one thing to avoid is splitting ownership across three people who never compare notes; the whole point of the three-tier structure is that someone is looking across all three cadences, not just running their own tier in isolation.

If your affiliate program, your content, or your paid media is already generating the citations that show up in AI answers, the question isn't whether AEO is happening. It's whether you're set up to see it, prove it, and fund it properly before a competitor gets there first. That's the conversation worth having with our team.