Google Ads management doesn't waste money loudly. It wastes it through structure: brand and generic terms sharing a budget, search terms nobody reviews, automated campaigns grading their own homework. The account looks fine at the top line and leaks at every layer below. This is the paid search pillar; Performance Max and budgeting each get their own piece.
The single most distorting habit in small accounts is letting brand search sit inside general campaigns. Brand clicks are cheap and convert well because the buyer already chose you. Blended together, brand performance hides generic underperformance.
A worked example. An account reports a blended 6x return. Split out, brand runs at 18x on a small spend and generic runs at 1.4x on the large one. Same account, completely different decisions. Until the split exists, every number the account produces is a blend of two unrelated businesses.
| Campaign | Job | Budget logic |
|---|---|---|
| Brand search | Defend your name, capture decided buyers | Small, capped, fully served |
| Generic search | Win non-brand demand by theme | The real acquisition budget, judged on its own return |
| Shopping or PMax | Product feed coverage | Separate, with brand traffic excluded where possible |
| Remarketing | Recover known visitors | Small, frequency-capped |
Inside generic search, group by intent theme, not by your internal catalogue logic. Tight themes keep ads relevant and quality scores honest.
The search terms report shows what people actually typed before clicking. Reviewing it is the core maintenance loop:
Accounts that skip this for six months routinely find a meaningful slice of spend going to searches that could never convert. The report is free. Reading it is the discipline.
Smart bidding works when conversion tracking is accurate and volume is sufficient. Both conditions fail quietly in small accounts. Before trusting target-return bidding: confirm purchase conversions track once per order with correct values, and expect instability below roughly 30 to 50 conversions a month per campaign. Under that, maximise conversions with a sensible budget cap, or manual control on small defended campaigns, behaves more predictably. Bid strategy changes also trigger learning periods, so change them rarely and judge them over weeks.
Fifteen minutes, same order: spend pacing against the month, search terms scan, conversion tracking sanity check against platform orders, any disapprovals, then one structural improvement. That cadence is the difference between running the channel and renting it, the same standard we apply in the Facebook ads piece.
Enough to generate conversion data you can decide with, sustained for at least two to three months. The budget piece covers the sizing logic; spreading a tiny budget across many campaigns guarantees nothing learns.
Terms you tell Google never to show your ads for. They're the main defence against paying for searches that can't convert, and the search terms report tells you which to add.
Always. Brand converts at a different rate for different reasons. Blending it with generic spend hides the true cost of acquiring new customers.
Qwrki is the operating layer that runs paid media delivery for small businesses, so the weekly routine actually happens: the search terms scan, the pacing check, the structural fix. We keep brand, generic and shopping accountable to their own budgets instead of one blended number that hides the leaks. Book a call if you want a read-through of how your account is structured today.
What Performance Max actually does, the visibility you trade away, the brand-traffic problem, and how to run PMax so it earns its automation.
A practical method for setting a paid media budget: work back from contribution margin, size spend for learning, and govern with MER instead of platform ROAS.
Creative is the targeting now. A practical testing framework for Meta ads: concepts before variations, a sustainable production cadence, and fatigue signals.
Google Ads management doesn't waste money loudly. It wastes it through structure: brand and generic terms sharing a budget, search terms nobody reviews, automated campaigns grading their own homework. The account looks fine at the top line and leaks at every layer below. This is the paid search pillar; Performance Max and budgeting each get their own piece.
The single most distorting habit in small accounts is letting brand search sit inside general campaigns. Brand clicks are cheap and convert well because the buyer already chose you. Blended together, brand performance hides generic underperformance.
A worked example. An account reports a blended 6x return. Split out, brand runs at 18x on a small spend and generic runs at 1.4x on the large one. Same account, completely different decisions. Until the split exists, every number the account produces is a blend of two unrelated businesses.
| Campaign | Job | Budget logic |
|---|---|---|
| Brand search | Defend your name, capture decided buyers | Small, capped, fully served |
| Generic search | Win non-brand demand by theme | The real acquisition budget, judged on its own return |
| Shopping or PMax | Product feed coverage | Separate, with brand traffic excluded where possible |
| Remarketing | Recover known visitors | Small, frequency-capped |
Inside generic search, group by intent theme, not by your internal catalogue logic. Tight themes keep ads relevant and quality scores honest.
The search terms report shows what people actually typed before clicking. Reviewing it is the core maintenance loop:
Accounts that skip this for six months routinely find a meaningful slice of spend going to searches that could never convert. The report is free. Reading it is the discipline.
Smart bidding works when conversion tracking is accurate and volume is sufficient. Both conditions fail quietly in small accounts. Before trusting target-return bidding: confirm purchase conversions track once per order with correct values, and expect instability below roughly 30 to 50 conversions a month per campaign. Under that, maximise conversions with a sensible budget cap, or manual control on small defended campaigns, behaves more predictably. Bid strategy changes also trigger learning periods, so change them rarely and judge them over weeks.
Fifteen minutes, same order: spend pacing against the month, search terms scan, conversion tracking sanity check against platform orders, any disapprovals, then one structural improvement. That cadence is the difference between running the channel and renting it, the same standard we apply in the Facebook ads piece.
Enough to generate conversion data you can decide with, sustained for at least two to three months. The budget piece covers the sizing logic; spreading a tiny budget across many campaigns guarantees nothing learns.
Terms you tell Google never to show your ads for. They're the main defence against paying for searches that can't convert, and the search terms report tells you which to add.
Always. Brand converts at a different rate for different reasons. Blending it with generic spend hides the true cost of acquiring new customers.
Qwrki is the operating layer that runs paid media delivery for small businesses, so the weekly routine actually happens: the search terms scan, the pacing check, the structural fix. We keep brand, generic and shopping accountable to their own budgets instead of one blended number that hides the leaks. Book a call if you want a read-through of how your account is structured today.
What Performance Max actually does, the visibility you trade away, the brand-traffic problem, and how to run PMax so it earns its automation.
A practical method for setting a paid media budget: work back from contribution margin, size spend for learning, and govern with MER instead of platform ROAS.
Creative is the targeting now. A practical testing framework for Meta ads: concepts before variations, a sustainable production cadence, and fatigue signals.
