Half the budget went to the match type doing a third of the work

Last week I pulled a placement report and found the best placement in the account getting a fifth of the budget. This week, same account, different cut of the same data: match type.

This account ran no phrase campaigns, so these two rows are the whole manual budget.

Broad got more of the budget than exact. Not by much, but more. And exact converted at better than twice the rate for a third of the ACOS.

Put it in dollars. Exact returned $4.33 for every dollar spent. Broad returned $1.59.

Nobody sat down and decided to fund broad more heavily than exact. Three things push an account here on their own.

Broad match has far more traffic available to it. One broad keyword can pick up hundreds of variants, so if the bids are anywhere near each other, broad wins more auctions simply because it is eligible for more of them.

Bids usually start identical. Most accounts launch exact, phrase and broad at the same bid on the same keyword list, which is a reasonable thing to do on day one and an unreasonable thing to still be doing on day sixty.

And broad looks fine on the metrics people check. It has impressions, it has clicks, it has orders. The problem only shows up when you divide, and the divide-by-match-type view is not a report Amazon hands you. You have to build it.

This is the part worth being careful about, because the lazy conclusion is "kill broad" and that is wrong.

Broad match is a discovery mechanism. Its job is to find the search terms you did not think of, so you can promote the ones that convert into exact and negate the ones that do not. Judged as a discovery budget, a 63% ACOS might be acceptable, the same way you accept a higher cost on an auto campaign that is doing the same job.

Judged as half your budget, it is not acceptable, because that is not discovery spending any more. That is your main traffic source.

So the test is not "is broad expensive". It is "is broad feeding exact". Pull the search term report and look at what broad has surfaced in the last sixty days that you then promoted into an exact campaign. In this account, that list was short. Broad was not discovering. It was just buying loose traffic at scale.

Pull the Sponsored Products search term report, last 60 days. Campaign Manager, Measurement and reporting, Sponsored ads reports, report type Search term.

Sum spend, sales, clicks and orders grouped by match type. One row per match type. Then divide sales by spend for each, and orders by clicks for each. That is the whole analysis and it takes a few minutes in a pivot table.

Then ask the second question, which matters more than the first: how many of the search terms in your exact campaigns got there because broad found them? If the answer is "a few, a while ago", broad has finished its job at its current size and the budget split should reflect that.

The move in this account was not to pause broad. It was to pull broad's bids down to a level where it keeps discovering, and let exact take the spend that freed up.

Match type is not the only thing varying across those rows. Different keywords sit in different campaigns at different bids, and a broad keyword and an exact keyword with the same text are not competing for identical traffic. The comparison tells you where the money is going and what it is returning. It does not prove that moving a dollar from broad to exact returns $4.33 instead of $1.59. Move the bids in steps and watch what actually happens.

Every match type in your account is doing a different job, and the one with the most traffic available to it will quietly take the largest share of the budget unless something stops it. One row per match type, two divisions. If broad is outspending exact, something is wrong, and it is almost never a decision anyone made.

If you want this done for you, the free audit reads your bulk file and shows you the search terms, keywords and bids costing you money, with the numbers behind each one. No signup, no account, no Amazon login, and nothing is stored.

Thomas, PPC Tools