5 Mistakes 90% of Businesses Make Launching Google Ads Themselves
Mistake #1: Broad match without an exclusion list
By default Google sets keywords to broad match — your ad shows on any "matching" query. If you sell "women's running shoes", your ad will also show on "men's work boots", "sport on TV" and sometimes even "socks". Without a list of exclusions (negatives) you pay for clicks completely outside your market.
Fix: change the match type to phrase or exact (mark keywords in "quotes" or [brackets]). Every week review the "search terms" report and add irrelevant queries to the negatives list. That's 15 minutes a week and it can cut costs by 20–40%.Mistake #2: No conversion tracking (or misconfigured)
Without conversions Google Ads doesn't know what's working. The algorithm optimizes for "clicks", not for "customers". Result: ads pull traffic that never buys anything. Classic errors: tracking set up so every page open counts as a conversion, no separation of macro conversions (purchase, form) from micro (scroll, 60-second time on page).
Fix: configure GA4 + Google Tag Manager + Google Ads conversion tags. Use preview mode to check that conversions only count when someone actually bought / submitted the form. Mark conversions as "primary" (for optimization) and "secondary" (observation only).Mistake #3: Performance Max from day one
Google really likes PMax and pushes it on every new client. The problem: PMax is an algorithm that learns from conversions. If your account has zero history and 5 conversions a month, PMax has nothing to learn from — it spends the budget almost at random, mostly on Display and YouTube (where clicks are cheap but quality is low).
Fix: start with a Search campaign with well-chosen keywords and proper conversion tracking. From my experience PMax starts to work sensibly from ~50 value-based conversions per month and 2–3 months of account history (Google no longer publishes a hard threshold, but below this range the algorithm stays in learning mode and resets after every change — earlier than that, PMax just burns budget).Mistake #4: Google auto-recommendations enabled without oversight
Google enables "auto-apply" for recommendations by default: budget increases, adding search partners, expanding match types to broad. All these changes happen automatically and serve to grow spend, not necessarily results. Usually the first thing I see in a new client: 3–5 auto-recommendations applied without their knowledge.
Fix: go to Recommendations → Auto-apply and turn off all the automations. Evaluate every recommendation manually. Some are sensible (e.g. new ad extensions), some are not (expanding targeting = burning budget).Mistake #5: Brand and generic in one campaign
The keywords "your-brand" and the generic "your-industry-city" mixed in the same campaign give a distorted picture of results. CTR on brand reaches 30–40%, on generic 3–7%. Mixed in one campaign they show "average 8% CTR" — brand pulls generic up, and you don't realize the generic campaigns are actually weak.
Fix: split into two separate campaigns. Brand with a separate budget (typically 5–10% of the total), generic with its own (the rest). Then you see the real results of each group and can optimize separately.Note: this mistake applies to accounts where the brand is already searched on Google (customers type your company name). If you're just starting and have no brand traffic yet — skip this point and come back to it after 2–3 months, once you see your company name in the search terms report.
What I see in new clients
An audit of a new account usually reveals:
- Local service — broad match without negatives, no conversions in Google Ads (only in GA4, not passed through). After applying the fixes, customer acquisition cost dropped from 280 PLN to 120 PLN in 2 months.
- Cosmetics e-commerce — PMax from day one, 800 PLN/day budget, ROAS 0.9 (i.e. losing money). After switching to Shopping + Search with normal optimization: ROAS 4.2 at the same budget.
- Diet catering — brand and generic in one campaign, "average CPC 5.80 PLN". Once split it turned out: brand 0.90 PLN, generic 8.40 PLN. The client thought they had control over CPC — they didn't.
How to check whether any mistake is on your account
Five quick checks in 10 minutes:
- Keywords → match type. If you mostly see "broad" — you have Mistake #1.
- Tools → Conversions. If the list is empty or counts "every visit to the site" — you have Mistake #2.
- Campaigns → type. If you have Performance Max and the account has few conversions (roughly below ~50/month) — you have Mistake #3.
- Recommendations → Auto-apply. If enabled — you have Mistake #4.
- Ad groups → keywords. If your company name and generic industry queries are in the same campaign — you have Mistake #5 (assuming you already have brand traffic).
If your budget grew by more than 20% in the last 30 days without your explicit decision — that's also a signal that auto-recommendations are working on you.
About the sources: The article is based on observations from 11 years of running Google Ads campaigns at JP Ads + dozens of audits of accounts taken over after DIY launches. Numbers from my own practice are flagged "from my experience" — distinguish them from hard market research.