Are you paying Google for clicks you already rank for?

Here's a pattern I find in most Google Ads accounts I audit for owner-led companies, and it was the whole story in the last two: the account is buying the searches the website already wins. The cost guides rank on page one for "how much does X cost"; the paid campaign bids on "X cost." The firm is top three for "X lawyer [city]"; the ad shows above its own listing. Every month the invoice says the campaign produced leads, and some of them it did. Some of them would have arrived anyway, by the free route, one inch lower on the page.

Nobody sets out to do this. It happens because the person running the account is paid as a percentage of spend, or because the campaign was built before the SEO started working, or because "we rank for it" and "we bid on it" live in two different reports that nobody puts side by side. This guide puts them side by side, in thirty minutes, and then walks through the decision: what to keep buying, what to stop, and what to do with the difference.

The 30-minute check: the Paid & organic report

Google Ads has a report built for exactly this question, and most accounts have never opened it.

1. Link Search Console to Google Ads. In Google Ads, under linked accounts (Admin → Linked accounts in the current interface), link the Search Console property for your site. (If you don't have Search Console verified for the site, that's step zero, and it's free.)
2. Open the report. Insights and reports → Report editor → Predefined reports → Basic → Paid & organic. Set the date range to the last 90 days.
3. Read the three segments. The report splits each query into Ad shown only, Organic shown only, and Both shown. For each it shows ad clicks, ad impressions, ad cost, organic clicks, organic queries, and the combined totals.
4. Export the "Both shown" rows. That's the overlap: every query where you paid for a click on a page where your own free listing also appeared.
5. Add the organic position. Pull the same queries from Search Console (Performance → Search results → filter by query) to get the average organic position for each. The report tells you the two were on the page together; Search Console tells you how high the free one was.

You'll have a sheet that looks like this (fictional roofer, numbers illustrative):

QUERY                         | ORGANIC POS | AD CLICKS | AD COST | ORGANIC CLICKS | BUCKET            | DECISION
"ridgeline roofing"           | 1           | 212       | $318    | 640            | Brand             | Keep (competitor bidding), cap
"roof replacement cost ny"    | 2           | 94        | $1,692  | 310            | Informational     | Cut; organic page gets a form
"roofer near me"              | 7 (map 3)   | 180       | $3,240  | 45             | Decided buyer     | Keep, tighten geo
"storm damage roof repair"    | 1           | 61        | $1,037  | 150            | Decided buyer     | Test: 3-week geo holdout
"roof leak emergency"         | -           | 140       | $2,380  | 0              | Decided buyer     | Keep; after-hours answer
"how long does a roof last"   | 1           | 38        | $418    | 520            | Informational     | Cut

Thirty minutes in, you can already see it: the two informational queries cost about $2,100 over the period, the site ranks first or second for both, and the organic listing delivered more clicks than the ad on each. The emergency query, by contrast, has no organic presence at all. That's a query the account should own.

Bucket every overlapping query

Don't decide query by query yet. Put each "Both shown" row into one of five buckets; the decision is mostly made by the bucket.

The informational bucket is where the money is. It's also where people get the decision wrong in both directions: cutting a query whose organic page has no way to contact you (the click arrives and leaves), or keeping a query out of habit because "it drives traffic." The test isn't traffic. It's whether a lead arrives, and by which route.

What Google's own research says about overlap

This is the part most agencies won't put in a deck. In 2011 Google ran a large set of "search ads pause" studies and reported that, on average, 89% of paid clicks were incremental, meaning they didn't get replaced by organic clicks when the ads stopped. In March 2012 it published a follow-up that broke the same question down by the advertiser's organic rank, across 390 studies:

Read it the way Google did: when you rank first, about half the clicks you buy on that query would have come free. Two caveats that Google itself attached and that still hold: it's an average across many advertisers, and it measures clicks, not leads. Your number for any given query could be better or worse. Which is why the next step is a test, not a spreadsheet.

The three-week holdout that settles it

Google's recommendation in that same study is to run a controlled experiment rather than trust the averages, and that's still the right answer, especially for brand. The simple version any account can run:

1. Pick the queries to test: the brand terms, plus the three most expensive informational queries where you rank in the top three.
2. Split the geography, not the calendar. If you serve several counties or metros, pause the ads for those queries in half of them and leave the other half running. A calendar split (two weeks on, two weeks off) is easier but confounds the result with seasonality and news.
3. Count leads, not clicks. Calls, forms and bookings from each half, with the source recorded. If calls aren't tracked by source yet, fix that first; most accounts are blind to the phone, and the phone is where the decided buyers are.
4. Run three weeks, four if volume is low. Compare total leads per week in the paused half against its own prior weeks and against the running half.
5. Decide per bucket. Leads held: the organic listing was doing the job; cut the ads and keep the money. Leads fell: the ads were working; keep them and cap the bid so you're paying for the incremental part, not the whole click.

The rule for moving the budget

After the test you'll know, in dollars, what the account was spending on clicks the site already earns. Here's what I do with it, in order:

1. Keep paid on three jobs: brand defence where it's needed, decided-buyer queries, and the problem/stuck cluster. Those are the searches where an ad reaches a buyer organic doesn't reach, or reaches them first.
2. Fix the organic pages before cutting their ads. If the page that ranks for "roof replacement cost" has no form and no number above the fold, the ad was compensating for a bad page. Fix the page, then cut the ad, then watch the organic lead count.
3. Move the freed budget into the organic program rather than back into the auction: the pages for queries you don't rank for yet (usually the problem/stuck cluster and the comparison queries), the metadata rewrite on the pages you do, and the AI-search work that gets you named when the buyer asks an engine instead of a search box. The auction resets to zero every month; a page that ranks keeps paying.
4. Re-run the Paid & organic report quarterly. As the organic side grows, more queries move into the overlap, and the paid account should keep shrinking toward its three jobs.

To size the organic side in your own numbers, the calculator on the SEO page takes your monthly organic visitors, your visitor-to-lead rate, your close rate and what a customer is worth, and turns a realistic lift into a monthly figure. It's the same arithmetic as the sheet above, run forward instead of backward.

A worked example

A firm spends $15,000 a month on Search. The Paid & organic report shows $6,200 of it on queries where the firm ranks first or second organically, most of them informational. The holdout test shows total leads on those queries holding within a few percent when the ads pause. The account drops to $8,800 on brand, decided-buyer and the problem cluster. The $6,200 funds an organic program: the problem-cluster pages that nobody owns, the schema and fact consistency work that gets the firm named by AI engines, and a monthly check that the paid account hasn't crept back onto what organic wins.

Six months later the firm is spending less in total, has more leads from the organic side than it had from the overlapping ads, and has a set of pages that will still be producing when the next agency pitch arrives. That's the shape. The numbers are yours to fill in, and the report above is where they come from.

One more thing worth saying plainly: if your account is managed at a percentage of spend, nobody in that arrangement is paid to find this. That's not a character flaw; it's the contract. Run the report yourself, or have someone run it who gets paid the same either way.

Questions I get about this