The Best Google Ads Optimization I've Done This Year Was Deleting Things

Every ads blog says the same thing: add more creative, add more keywords, add more budget. Turn the dials up. So here is the most useful thing I can tell you after a year of rescuing ad accounts — the highest-ROI work is almost always deletion, not creation.

The account that prompted this post was bleeding. Nonbrand cost-per-acquisition had climbed roughly 50%. Weekly nonbrand revenue had fallen about 55%. The instinct in the room was to build: new campaigns, fresh creative, a bigger test budget. We did the opposite. We audited what the account was already paying for, and we started removing things. Revenue went up.

We cut spend and revenue went up. That is not magic. That is a measure of how much waste was in the account.

Leak 1: You are paying for searches that will never convert

Broad and phrase match quietly pull in queries that have nothing to do with intent to buy. Someone searching for a free version, a competitor's warranty policy, or a job at your company is a click you paid for and will never recover. The fix is unglamorous: pull the search-terms report, sort by spend, and build a negative-keyword list from the queries that burned money without producing revenue. This one move often reclaims a double-digit percentage of nonbrand spend.

Leak 2: Competitor-conquest terms that never pencil out

Bidding on competitor brand names feels aggressive and strategic. It rarely is. Those searchers have high intent for a different brand, so you pay premium clicks to convert at a fraction of your own branded rate. Some conquest works. Most of it is a slow leak dressed up as a growth tactic. Score each conquest campaign on its own CPA, and pause the ones that only survive because nobody looks at them in isolation.

Leak 3: The account structure fighting itself

This was the strangest finding in the rescue, and it is more common than you would think. The brand's best-converting search terms were being suppressed — not by the algorithm, not by budget, but by the account's own exclusion lists. Somewhere along the way, negatives had been added that blocked the exact queries driving the most efficient conversions. The brand was paying to hide its winners.

Before you add anything, read your negatives. Cross-reference your negative-keyword and excluded-audience lists against your top converters. If a term that converts beautifully appears anywhere in an exclusion, you have found free money.

Leak 4: Landing pages quietly rerouting your best traffic

Great ad, terrible destination. When a chunk of paid traffic lands on a page that converts three to five times worse than your proven page, no amount of bid tuning saves it. Map each campaign to the page it actually sends people to, compare conversion rates across those pages, and reroute traffic off the losers. The ad is doing its job; the funnel is dropping the ball.

What not to touch during a rescue

Subtraction has a discipline to it. When an account is bleeding, the temptation is to change everything at once — which makes it impossible to know what worked. Resist it. In a rescue, I leave the levers that carry real risk of breaking things alone until the bleeding stops: I do not slash prices, I do not swap out hero images or proven top-of-funnel creative, and I do not restructure campaigns that are actually profitable just because the account 'feels' messy. The point of a rescue is to remove waste, not to introduce new variables. Cut the leaks, watch the numbers stabilize for a week or two, and only then start testing additions on a clean baseline.

There is also a lesson here that lives outside the ad platform entirely. On one account, the single biggest swing in performance had nothing to do with bids or keywords — it was an out-of-stock product that came back into stock. No amount of clever media buying can sell inventory that isn't there. Sometimes the best optimization is a conversation with your supply chain, not your ad manager. Before you blame the campaign, make sure the thing it is advertising is actually available to buy.

How to run the subtraction audit

You do not need new budget to do this. You need an afternoon and the discipline to look at what you are already buying:

1. Pull the search-terms report, sort by spend, and negate the queries that spent without converting.

2. Isolate every competitor-conquest campaign and judge each on its own CPA. Pause what does not pay.

3. Read your negative and exclusion lists against your top-converting terms. Unblock any winners you are suppressing.

4. Map campaigns to landing pages, compare conversion rates, and reroute traffic off the underperforming pages.

Here is the reframe that makes all of this land. Most brands do not have a traffic problem. They have a leak problem. Buying more water for a bucket full of holes just gets you a more expensive puddle. Patch the holes first. The account almost always has more to give than the media plan assumes — you just have to stop paying to work against yourself.

If your CPA is climbing and your instinct is to add, resist it for one afternoon. Audit what you already own. The best optimization I did all year was pressing delete.