Can You Estimate a Competitor's Ad Spend?

Meta publishes every active ad in its Ad Library, with creative, format, placements and start date, for every advertiser. It publishes spend ranges and reach only for political and social issue advertising, which is a small fraction of commercial activity.

Google publishes verified advertisers ads in the Ads Transparency Center with date ranges and formats, and again reserves spend disclosure for election advertising. In the EU, transparency rules surface additional reach information for ads served there, which makes European advertisers marginally more visible.

So the honest summary is that you can see what is running and for how long, and you cannot see what it cost. Everything beyond that is inference.

The general method is to estimate impressions, then multiply by an assumed price per thousand impressions or per click. Impressions are inferred from sampled visibility, panel data, or how often the tool own crawlers observed the ad. Prices come from category averages.

Both inputs carry large uncertainty, and the multiplication compounds it. An impression estimate off by forty percent combined with a price assumption off by thirty percent produces a spend figure that can be wrong by a factor of two in either direction, while still being presented as a single number.

Category averages are also the wrong price for any specific advertiser. Auction prices depend on targeting, quality, competition and time of year, so a brand with an unusual audience or a strong quality score will diverge from the average by design.

Active ad count is observable and comparable over time. Twenty active ads versus three is a real difference in programme scale, and a jump from three to twenty in a fortnight is a real change in commitment.

Duration is the best proxy for performance, since nobody keeps paying to run creative that loses money. An ad live for four months is a tested winner, and an ad that vanished in four days probably was not.

Creative velocity is the third. A competitor producing new variants weekly has a team and a testing process, while one running the same two images for six months does not. That tells you more about their operational capacity than any spend figure would.

Teams rarely want the spend number for its own sake. Usually the real question is one of three: are they outspending us, is paid working for them, or should we be worried. Each of those is answerable from observable data.

Are they outspending us: compare active ad counts, placement breadth and creative velocity against your own programme. It is a relative answer rather than an absolute one, which is what the decision needs anyway.

Is paid working for them: look at duration and repetition. A programme that has run continuously for a year with rotating creative is working. One that appears in bursts and disappears is either seasonal or unprofitable.

If you do use a third-party estimate, use it for ranking rather than for planning. Saying that competitor A appears to spend materially more than competitor B is defensible. Putting an estimated monthly figure into a budget case is not, and it will be challenged by anyone who knows how it was produced.

Label the source and the method wherever the number appears. An unlabelled estimate becomes a fact within two slide revisions, and then somebody plans against it.

Oppira tracks the Meta and Google ad libraries daily for the competitors you follow, recording active ad counts, run duration and creative changes over time. It does not invent a spend figure, because the underlying data does not support one.