When more budget makes your clicks more expensive
A bigger Google Ads budget lifts average CPC with no change in the auction. What a 14-advertiser study measured, and how to price the marginal click first.
Raising a Google Ads budget can push your average cost per click up without a single change in the auction, because average CPC is a ratio and the budget you add buys the clicks you were not buying before. A study of 14 B2B advertisers published in October 2026 quantified the effect: about 1.4% more per click for every 10% more budget, on top of whatever the market did. The number worth tracking in your own account is not average CPC. It is the price of the marginal click.
That distinction splits one question in two that get argued as if they were the same: did the same click get more expensive, or did we start buying different, pricier clicks? Both are answerable from one export, and most accounts never run it.
Average CPC is a division problem
Google defines average cost per click as the total you paid divided by the clicks you received. It is an average of two aggregates, and it moves whenever either one moves — including when nothing about the auction changed at all.
A campaign that is limited by budget does not buy clicks in a random order. It buys the cheapest ones first and stops when the money runs out. Add budget and it keeps going down the list, into positions it previously declined, searches it previously skipped, and hours where competition is heavier. The keywords are identical. The price paid for them is not.
So the useful measurement is not the average. It is the increment:
- Average CPC — total cost divided by total clicks. Describes the whole period, including clicks you were always going to buy.
- Marginal CPC — the change in cost divided by the change in clicks between two periods. Describes only the clicks the extra money bought.
If cost rises from 4,000 to 4,600 while clicks rise from 1,000 to 1,100, average CPC reads 4.18. Marginal CPC is 6.00 — the price actually paid for the increase, 44% above what the average implies.
What the study actually measured
Grey Matter, a B2B search agency, published "Same Click, Higher Price" in October 2026 using its own client Google Ads data. The design is the reason it is worth quoting: it tracked the same keywords in the same accounts rather than comparing industry averages.
- 14 B2B advertisers, anonymised, from the agency's client base.
- 1,349 non-brand keywords, 5,035 keyword-quarters.
- Six quarters, Q2 2025 through Q3 2026, compared year over year across April to September.
- Search campaigns only. Performance Max, Shopping and Display are excluded.
- Controls applied progressively: campaign spend and impression share lost to budget, then top-of-page position and Quality Score.
- Two exclusions that matter: broad match keywords, because the searches they match drift over time, and Maximize Clicks campaigns, because their CPC reflects how the algorithm spreads a fixed budget rather than what the auction charges.
The headline result is that the same non-brand click cost roughly 21% more year over year after those controls. The step-by-step version is more instructive, because it shows how much of the increase came from advertisers' own budget decisions:
| What the estimate holds constant | Same-keyword CPC change | What it describes |
|---|---|---|
| Nothing | +29% | Raw year over year, including your own budget and keyword changes |
| Campaign spend and share lost to budget | +22% | The same click, with your own budget decisions removed |
| Spend, position and Quality Score | +21% | The market component every advertiser faced |
| Advertisers' actual paid CPC | +36% | What the typical advertiser really paid, own choices included |
The first two rows are the argument: budget changes alone accounted for seven of the 29 points, before any new keyword or bid strategy entered. The gap between the third row and the fourth is everything the advertiser did to itself.
The authors are careful about what this is: correlation, not proven causation. The likely range around 21% is wide, from +1% to +44%. No industries, markets or currencies are disclosed, and because Google Ads does not retain historical budget settings or bid targets, spend and share lost to budget stand in as proxies. Drop the largest spender and the estimate falls to 17%, with a range crossing zero — so even the market figure is not settled.
The elasticity is a power law, not a straight line
The most operationally useful number in the report is buried under the headline: every 10% budget increase added about 1.4% to an advertiser's own CPCs, on top of the market.
That is a ratio relationship, so it compounds. Doubling a budget adds roughly 11%, not 14%, because the percentage applies to a growing base. The third budget increase therefore costs more per click than the first, and the fifth more than the third. If your Q4 plan assumes a linear cost of scaling — spend twice as much, get twice the clicks at the same price — the model is wrong in the direction that flatters the plan.
The report also found that keyword and match-type changes moved individual advertisers' CPCs by up to 17% in either direction, and nearly cancelled out across the sample. Read that carefully, because it hides a trap: at portfolio level the mix effect looks like noise, while inside one account it can be the largest term. One advertiser held budget flat, saw the same keywords rise 20%, then paid 40% more overall after shifting spend into pricier phrase-match keywords. Nothing was wrong with the auction. The buying changed.
Google's own documentation describes the same mechanism
This is not a secret the platforms are keeping. Google's help documentation on the "Limited by budget" status says the condition means your daily budget is below what is needed "to capture all available impressions and clicks for your current settings," and names bid strategy as a cause: "Some automated bid strategies, like target CPA or Target ROAS, might aim for outcomes that require higher bids."
Google also shipped a direct change to how this behaves. From August 17, 2026, its bidding systems were updated to "deliver more consistent and predictable performance for campaigns that are limited by budget," with a warning that Target CPA and Target ROAS campaigns "may cause temporary performance and traffic fluctuations," plus a Bid Target Adjustment Tool available from July 6, 2026. If your account runs target-based bidding and has been beating its number, that is a separate audit and we covered it in what changes when your Target CPA stops being a suggestion.
Now the tension. The same guidance recommends keeping a daily budget "comfortably higher than your average daily spend" so a campaign can "capture as much profitable traffic as possible without restricting your campaign." Both are true, and together close to contradictory: headroom removes the constraint, and headroom is exactly what the study prices at 1.4% per 10%.
The resolution: "comfortably higher" is only sound for a campaign whose marginal cost per acquisition is already below target. For one that is above target, it is an instruction to buy progressively worse clicks. That judgment is the one your Q4 review has to make explicitly.
Run it yourself: the marginal click pull
No vendor dashboard reports marginal CPC. You compute it, in about an hour, from data you already have.
- Open one campaign, not the account. Segment by week, and add Cost, Clicks, Conversions, Search impression share, and Search lost IS (budget) as columns. Restrict to non-brand terms.
- Export twelve weeks. Pair consecutive weeks where the change in clicks is at least 50. Below that, the cost-per-click ratio is arithmetic noise, not a measurement.
- Compute marginal CPC per pair as change in cost divided by change in clicks. Put it next to the average CPC for the same pair.
- Compute your break-even CPC. Target cost per acquisition multiplied by conversion rate gives the click price at which an acquisition still lands on target. That is the line marginal CPC has to clear.
- Gate on the constraint. Only count pairs where Search lost IS (budget) was meaningfully above zero before the increase. If the campaign was never budget-constrained, extra budget cannot raise CPC through this mechanism, and a rise means mix, seasonality, or a bid-strategy change.
- Exclude what the study excluded: broad match, because its matched searches drift, and Maximize Clicks campaigns, because their CPC reflects budget spreading rather than auction price. Both give a confident wrong answer.
The decision rule fits in a sentence: if marginal CPC sits above your break-even, the last budget increase bought clicks you would not have bought one at a time, and the right move is to hold spend and fix the click — not to add more.
Where the relationship breaks
A framework that only works in one direction is a sales pitch, not a model. Three cases where it fails.
Manual CPC with maximum bid caps. When the cap binds, additional budget buys more clicks at roughly the same price until the eligible inventory is exhausted. The elasticity assumes automated bidding can reach for worse inventory; a hard cap stops it.
Accounts where added spend buys cheap inventory. Marginal CPC can fall: long-tail expansion, cheaper placements, or a new geo with less competition pull it below the average. Common enough that it must be tested rather than assumed — the strongest argument for running the pull.
Campaign types outside Search. Performance Max and Shopping buy a different inventory mix the study did not measure. Don't carry the 1.4% into them.
What this does not tell you
It tells you the price of a click. It does not tell you the value of one. Every conclusion above is about cost per click and cost per acquisition against a target you set — none of it is a revenue or conversion claim, and platform-reported conversions are a separate reconciliation problem we covered in the two numbers on Google's invalid-traffic card.
If you take one thing from a study this thin on disclosures, take the shape rather than the number. Fourteen advertisers is a hypothesis with a wide range, and the authors say so. What survives the small sample is the mechanism: average CPC is a ratio, budget buys the next click down the list, and the only honest measure of an increase is what the increment cost.
If your paid search structure needs this kind of measurement rather than another dashboard, that is what our digital ads practice builds — spend decisions priced per increment, not per campaign average. And if you are weighing a fourth budget line against the three you already run, the argument for separating AI Mode spend covers where the money should sit before you scale anything.
Frequently Asked Questions
Does raising a Google Ads budget always raise the cost per click?
Not always, but often — and the reason is mechanical rather than competitive. A campaign that is limited by budget buys the cheapest clicks first. When you add budget, it buys the next tranche of inventory available at the current bid: higher positions, more competitive searches and pricier times of day. Average CPC rises because the mix of clicks changed, even if the auction price for any single keyword did not. Where the campaign was never limited by budget, extra budget cannot move CPC this way at all.
What is marginal CPC and how do I calculate it?
Marginal CPC is the cost of the additional clicks you bought between two periods, not the average of all of them. Calculate it as the change in cost divided by the change in clicks between two comparable weeks. If cost moved from 4,000 to 4,600 and clicks from 1,000 to 1,100, marginal CPC is 6.00 while average CPC is about 4.18. Only the marginal figure tells you what the extra budget actually bought.
How much does CPC rise for each 10% of extra budget?
In the Grey Matter study of 14 B2B advertisers published in October 2026, every 10% budget increase added about 1.4% to an advertiser's own CPCs on top of the market change. The relationship is a power law rather than a straight line, so doubling a budget adds roughly 11% rather than 14%, and later increments cost more per click than earlier ones.



