I had a founder pull up his analytics on a screen-share a while back, scrolling through a dashboard with about fifteen widgets on it. Traffic was up. Rankings looked busy. There was a big green number somewhere near the top. And he asked me, completely sincerely, whether any of it meant SEO was actually working. He genuinely couldn't tell. Neither could the dashboard, the way it was set up.
That's the trap. Modern analytics tools will show you more numbers than you can read, and most of them move for reasons that have nothing to do with whether new customers are finding you. A traffic spike from one off-topic page going semi-viral looks identical to real progress on the chart. It isn't.
So this is the SEO measurement framework I'd hand that founder. It's built on Google's own free tools, because you don't need a paid stack to answer this question honestly. One note before we start: this is about how you read the signals. If the problem is that your reporting comes from someone else and you can't tell which of their changes earned the result, that's a different and very real problem, and I've written about why bundled agency reporting can't tell you what's working separately. This piece stays in the measurement lane.
What actually tells you SEO is working?
Non-branded organic clicks, trending upward over a 90-day window, in Google Search Console's Performance report. That's the single most honest signal for someone who isn't an SEO. If I could only look at one number, it'd be that one.
Here's why it beats the alternatives. A click is someone who saw your link and chose to come to your site. Search Console defines a click as a navigation outside of Google to your page, and re-clicking the same link still counts as one click, so it's a real, deduplicated count of human interest. Now split clicks into branded and non-branded. Branded queries are people typing your company name. Those people already know you. Their clicks measure recall, not discovery. Non-branded queries are people searching for the problem you solve without knowing you exist yet. When they click through, that is SEO doing the actual job: putting you in front of strangers.
You don't have to guess at that split anymore. In November 2025 Google added a native branded and non-branded filter right inside the Performance report. Open the query dimension, add a filter, and choose non-branded queries. Google building that filter is itself a quiet endorsement: the branded-versus-non-branded distinction matters enough that they put it in the product.
Why 90 days and not last week? Because organic search is noisy day to day, and Google's own guidance is that some changes take hours and others take months. A quarter is long enough to see a trend through the weekly wobble. A single up week proves nothing. A line that's higher across three months than it was across the prior three is hard to fake.
Before you read any performance number, confirm your important pages are actually indexed. The Page Indexing report tells you which URLs Google has crawled and indexed. A page that isn't indexed can't rank and can't earn a click, so a strong-looking content effort can be quietly invisible. Indexing is the prerequisite check, not an afterthought.
Which SEO metrics actually matter, and which are vanity?
No official source declares which SEO metrics are "vanity." That label is a judgment call, and I'll mark mine as such. The metrics that matter are the ones documented by Google's own tools: non-branded organic clicks (new people finding you through search), non-branded impressions (whether Google is surfacing your content for discovery queries), indexed pages (the prerequisite before anything else), organic conversions in Google Analytics (the only number that ties SEO to business outcomes), and average position filtered to pages and queries that matter to your business. The metrics that tend toward vanity are reported without context: raw total traffic across all queries, sitewide average position not filtered to target queries, and external rank-tracker positions that may not match what real users see. Domain authority scores don't belong on this list at all. Google has said plainly it doesn't use any third-party authority score. The difference between signal and vanity is almost always segmentation and business context, not the metric itself.
| Metric | What it tells you | Watch out for |
|---|---|---|
| Non-branded organic clicks (documented) | New people finding you through search who didn't start with your name. The clearest "it's working" signal. | Needs the branded/non-branded filter applied. Read it over 90 days, not week to week. |
| Non-branded impressions (documented) | Whether Google is surfacing your content for discovery searches at all. The earliest leading signal. | Impressions can rise because you're showing for irrelevant queries. Pair with clicks and query context. |
| Indexed pages (documented) | Whether your key pages are even eligible to rank. A prerequisite, not a performance metric. | "Crawled but not indexed" on important pages is an early warning before traffic problems show. |
| Organic conversions / key events (documented) | The only metric that ties SEO to business outcomes. The one that ultimately matters. | Requires Analytics with key events configured and Search Console linked. Moves last, so be patient. |
| Average position for target queries (documented) | Whether ranking is improving on the specific pages and queries that matter to the business. | It's an average across all queries a page appears for. It can improve while clicks fall. Filter it. |
| AI Overview impressions (documented, new) | How often your URLs appear inside AI Overviews and AI Mode. Presence in the AI surface. | No click or query data yet. It's a leading signal of presence, not proof of traffic. |
| Raw total traffic (all queries) (judgment) | Useful as context for scale, once it's segmented. | A spike from off-topic queries looks like success and drives no business. Segment or it's vanity. |
| External rank-tracker positions (judgment) | A supplemental view of where you sit for tracked keywords. | Personalized and location-specific; may not match real results. Search Console position is more reliable. |
One row deserves a flag, because it's the one people game. Sitewide average position is the metric most often waved around as proof, and it's the easiest to mislead with. It's a mathematical average across every query a page appears for. A pile of low-value, easy-to-rank queries can pull the sitewide average up while your commercially important pages quietly slide. Always filter average position to specific pages or query clusters. The headline number lies.
And while we're naming traps: domain authority scores are not on this list on purpose. Those are proprietary scores from SEO tool vendors, not Google metrics, and Google has said plainly it doesn't use any third-party authority score. A rising vendor score is not evidence your SEO is working. It's evidence a vendor's model likes your link profile.
What is the difference between leading and lagging SEO indicators?
Leading indicators are the early signals that move first. Lagging indicators are the business outcomes that move last. In SEO, the sequence runs: indexing first (Google has to crawl and index a page before it can rank), then impressions (the page starts appearing in results before earning clicks), then average position (the page competes and its rank stabilizes), then organic clicks (clicks follow once the page reaches positions where click-through is non-trivial), and finally conversions (business outcomes need traffic plus the right intent plus a page that converts). Each step triggers the next, which is why the signals don't move together. The framing of "leading versus lagging" is mine, but each step traces to how Google's own crawl-index-rank-serve process works. The practical payoff is knowing where to look: indexing and impressions are early warnings. Flat impressions at 90 days, not at 60 days, is the signal to actually worry about. Conversions confirm what clicks suggested, but they come last.
The order, fastest-moving to slowest:
- Indexing. The prerequisite. A new or updated page can be indexed within days of a crawl. If it never indexes, nothing downstream happens. This is where pages dropping out of the index shows up first, well before traffic reflects it.
- Impressions. A page starts appearing in results and racking up impressions before it earns meaningful clicks. Impressions for a freshly optimized page often grow for weeks before the clicks follow.
- Average position. As the page competes and accumulates data, its position stabilizes and, ideally, improves.
- Clicks and organic sessions. Clicks follow once rankings reach positions where click-through is non-trivial. This is the mid-lagging signal, and the 90-day non-branded clicks trend lives here.
- Conversions and key events. The slowest mover. Business outcomes need traffic, plus the right intent, plus a page that converts. Conversions come last, always.
The practical payoff: if you're 60 to 90 days into a real SEO effort and impressions are climbing but clicks are still flat, that is normal. That's the sequence working, not a failure. The signal to actually worry about is the opposite. If non-branded impressions are flat or falling past the 90-day mark, something structural is wrong, whether that's an indexing issue, content that doesn't match how people search, or a deeper technical problem. Flat impressions at the start is patience. Flat impressions at 90-plus days is a diagnosis.
Impressions up, clicks flat at 60 days is the sequence working. Impressions flat at 90 days is the sequence telling you to look harder.
How long before you should expect results?
This is where bad advice does the most damage, so I'm going to stick strictly to what Google has actually said and attribute each piece, because the popular "three to six months" figure you'll see everywhere isn't traceable to any primary source. It's a shortened paraphrase that lost its citation somewhere along the way. I'm not going to repeat it.
There are two real timeframes, and they describe two different things.
For assessing whether a specific change worked, Google's written SEO Starter Guide says some changes might take effect in a few hours and others could take several months, and that "in general, you likely want to wait a few weeks to assess whether your work had beneficial effects in Google Search results." It also says, honestly, that "not all changes you make to your website will result in noticeable impact in search results." A few weeks to read one change. That's the tactical clock.
For a full SEO engagement to pay off, the number Google has put on record is longer. In Google's "How to Hire an SEO" video, Maile Ohye, a Developer Programs Tech Lead at Google, said that "in most cases, the SEO will need four months to a year to help your business first implement improvements and then see the potential benefit." Four months to a year. That's the engagement clock.
Don't blend them. Waiting a few weeks to see if one title-tag change moved a page is reasonable. Expecting a whole strategy to remake your organic presence in a few weeks is not, and anyone promising it is selling something Google itself won't promise. While we're here: Google also states flatly that no one can guarantee a number-one ranking, and warns you to beware anyone who claims to. If a pitch comes with a guaranteed ranking attached, that's your answer about the pitch.
Can I measure whether my content shows up in AI search?
Yes, partially, and the tool to do it is new. Measuring modern SEO success includes measuring whether AI answer engines surface your content, because for how people actually search today, ranking in Google and getting cited by an AI answer are the same job. SEO, GEO, and AEO aren't three disciplines. They're one. On June 3, 2026, Google launched generative AI performance reports inside Search Console, covering AI Overviews, AI Mode, and generative AI features in Discover. For the first time you can see how often your URLs appeared in those AI surfaces, broken out by pages, countries, and dates. As of launch the report shows impressions only. There's no click data and no query data, which means you can see you're in the room but can't yet confirm anyone acted on the appearance. Treat AI Overview impressions as a leading indicator of presence, the same way regular impressions are: a useful signal, not yet proof of traffic.
One more limit worth noting: the report is rolling out to a subset of sites first, so you may not have access yet. Worth checking in your Search Console. And when you do get it, worth tracking consistently. Not yet worth treating as a conversion driver.
Honest limit: none of these numbers explain why. They tell you that non-branded clicks rose or that impressions stalled. They don't tell you whether the cause was your content, a competitor's slip, a Google update, or seasonality. The dashboard reports the outcome. Figuring out the cause is a separate, harder problem, and it's usually where a fresh expert eye earns its keep.
Start by confirming your key pages are indexed. Then watch non-branded organic clicks over a rolling 90-day window as your primary "is it working" signal, with non-branded impressions as the earlier read and organic conversions as the lagging proof it reaches the business. Filter average position to the pages and queries that matter; never trust the sitewide average alone.
Expect the signals in order: indexing, impressions, position, clicks, conversions. Give a single change a few weeks to read, and a full engagement four months to a year, in Google's own words. And add the AI surface to the picture, knowing it currently shows presence, not proven traffic.
If you've gone through all of this and the numbers still don't add up, or they technically move but never reach the business, that's not a measurement problem anymore. That's a diagnosis problem. Pinning down why the signals aren't behaving, and what to change, is the kind of work a search and AI visibility engagement is built to do. If you're still deciding what kind of SEO help makes sense for your situation, the consultant vs. agency vs. in-house breakdown covers that decision in full.
Frequently asked questions
Non-branded organic clicks trending upward over a 90-day window in Search Console. Branded clicks come from people who already know you, so they measure brand recall, not discovery. Non-branded clicks are new people finding you through search who didn't start with your name. When that line moves up over a quarter, SEO is expanding your reach. Rankings and total traffic can move for reasons that have nothing to do with new business.
Google's written starter guide says you likely want to wait a few weeks to assess whether a specific change had a beneficial effect, and notes some changes take hours while others take months. For a full SEO engagement to deliver benefit, Google's Maile Ohye has said in most cases an SEO will need four months to a year to implement improvements and then see the potential benefit. Those are two different phases. Assessing one change is a few weeks; the full payoff is four months to a year.
Rankings are a legitimate leading indicator, but they're easy to misread. External rank trackers show personalized, location-specific positions that may not match what real users see. Average position in Search Console is more reliable, but it's a mathematical average across every query a page appears for, so it can improve while clicks fall. Rankings become misleading only when reported without business context. Treat them as a supporting signal, not the headline.
First, confirm your important pages are indexed in the Page Indexing report, because a page that isn't indexed can't rank or earn traffic. Then look at the Performance report and filter to non-branded queries. You want non-branded impressions and clicks trending up over a 90-day window, and stable or improving average position on the specific pages and queries that matter to your business. Flat or declining non-branded impressions past 90 days is a warning sign.
Conversions are the lagging but most meaningful indicator, because they tie SEO to business outcomes. The simplest path for a non-expert is the Google Organic Search Traffic report in Google Analytics, which shows landing pages alongside Search Console metrics and your configured key events. It requires a linked Search Console property and key events set up in Analytics. Conversions appear last, after rankings and traffic, so don't expect them to move first.
Yes, partially, and it's new. In June 2026 Google added generative AI performance reports inside Search Console that show how often your URLs appeared in AI Overviews, AI Mode, and AI features in Discover. As of launch the report shows impressions, pages, countries, and dates, but no click data and no query data, and it's rolling out to a subset of sites first. So treat AI visibility as a leading signal of presence, not proof that the appearance drove traffic.
Sources
- Google Search Console Help: Search performance report
- Google Search Console Help: Performance report metric definitions (clicks, impressions, CTR, average position)
- Google Search Central Blog: Branded queries filter in Search Console (November 2025)
- Google Search Console Help: Page Indexing report
- Google Analytics Help: Google Organic Search Traffic report
- Google Search Central Blog: Generative AI performance reports in Search Console (June 3, 2026)
- Google Search Central: SEO Starter Guide (timeframe for assessing changes)
- Google: "How to Hire an SEO" (Maile Ohye, four months to a year)
- Google Search Central: Do you need an SEO? (no guaranteed rankings)