Nine in ten ad accounts we audit leak money through the same handful of mistakes. Check yours against this list: each item includes the symptom, the cost, and the fix.
Match types set to spray
Broad match without proper conversion signals and guardrails buys adjacent-but-wrong intent at scale. Symptom: search terms report full of queries you would never bid on. Fix: tighten to phrase and exact for core terms, let broad run only where conversion tracking is clean and the algorithm has enough data to learn from, and review search terms weekly, without exception.
The empty negative list
Negatives are half the targeting. Accounts without a maintained negative library pay for "free", "jobs", "how to become", competitor names they cannot convert, and every informational modifier in the language. Build negatives at account level for universals and campaign level for specifics; seed from the search terms report, then keep pruning weekly.
Bidding against your own data
- Smart bidding on broken or thin conversion tracking: the algorithm optimizes toward noise
- Target CPA set to a wish instead of history, strangling delivery
- Maximize clicks left running long after lead quality data existed
- Every conversion counted equally: a newsletter signup training the system like a purchase
Sending traffic to the homepage
The click is half the job; the page is the other half. Generic homepages force visitors to re-find what the ad promised. Message-matched landing pages, same offer, same language, one clear action, routinely double conversion from identical traffic. If you can only fix one thing this month, fix where the clicks land.
Reporting on activity instead of profit
Impressions, clicks and even raw ROAS mislead: blended ROAS hides that branded terms subsidize losing campaigns. Segment brand from non-brand, track cost per qualified lead (not per form-fill), and push toward profit-aware targets. What you measure is what the account optimizes toward, choose numbers that pay bills.
Key takeaways
- Weekly search-term review plus a living negative list is the cheapest performance lift in paid media.
- Smart bidding is only as good as conversion data, fix tracking before trusting automation.
- Message-matched landing pages double outcomes from the same spend.
- Segment brand vs non-brand or your ROAS is fiction.
Frequently asked questions
How much should I spend to test Google Ads properly?
Enough for statistical signal: typically fifty-plus clicks per ad group decision. In most Indian service markets that means at least ₹30k–₹50k monthly for a meaningful test; less just buys noise.
Should I fire my agency if these mistakes exist in my account?
Ask them to walk you through search terms, negatives and conversion setup first, competence shows fast. Silence or dashboards-instead-of-answers tells you what you need to know.
Is Performance Max a mistake?
It is a tool: powerful with strong conversion data and assets, a black-hole budget drain without. We typically run it alongside, never instead of, well-structured search.