New format
How to choose a paid search contractor
You hand paid search to a contractor so you don’t live in the accounts — and a month later you see “lots of clicks” with no leads, or silence. Without your own metrics and brief it’s hard to tell a normal ramp from a failure.
Below: what to watch in analytics, how to count payback and customer cost, which brief and reporting to set, and which exceptions show up in the first months. Finding a marketer in general is a separate piece; here the focus is controlling paid search itself.
Why you need your own control frame
The contractor manages bids and ads; you know deal economics. Without analytics access and KPIs you judge the work by feelings and pretty account screenshots.
The memo below isn’t a brief replacement — it’s a checklist: traffic → economics → brief/reports → exceptions when numbers aren’t a verdict yet.
Visit and goal dynamics
Check Metrika and Analytics yourself: sessions from paid search, bounce, goals (lead, call, payment). A wide catalog needs a wider keyword set; a narrow service needs phrase precision — not reach for reach’s sake.
Growing visits without growing target actions means check relevance and the landing — not only “tweak bids.”
Economics: ROI, CPA, LTV
Paid search is closer to fast leads than accumulating SEO, but payback still depends on margin and deal cycle. If spend stays above its profit contribution past a reasonable test window — review the channel and the contractor.
ROI is handy as (profit − ad spend) / spend × 100%. Round-number examples illustrate the method — not an industry benchmark.
For repeat purchases compare CPA with LTV, or gross profit per customer over the horizon you actually see. “CPA well below LTV” depends on margin; the “three times” rule from old guides is a rough heuristic, not law.
Minimum numbers in the report:
- spend and clicks/impressions by campaign
- goals and CPA/CPL
- dynamics vs the prior period
- what changed and why
Brief, reporting, and the monthly plan
Without a goal (“leads at CPL up to N,” “sales with attribution”) the contractor optimizes what’s easy to show: cheap clicks or broad reach.
A healthy cycle: brief → monthly work plan → access → regular report with conclusions and the next step. When KPIs rise — scale; when they drop — change hypotheses, not only “need more budget.”
Micromanaging every click hurts. Escalate when there are no reports, no hypotheses, or the account is locked from you.
When weak numbers aren’t a verdict yet
The first month in a narrow niche goes to the keyword set and negatives — expect transparent structure progress, not miracle ROI.
A fuzzy brief (“more sales”) and seasonal dips distort the picture. Lock seasonality and the lead definition early.
A red flag isn’t a one-off loss — it’s no fix plan, refusal to explain numbers, and guarantees with no metrics. Then look for another team.
FAQ
How is this different from finding a marketer?
That piece covers agency vs in-house and skill stacks. Here — how numbers and process show a Direct/Ads contractor runs the channel honestly.
Which systems should I watch?
Yandex Metrika and/or Google Analytics with goals, plus Direct and Ads accounts. Client access is required.
What’s a “normal” ROI?
There’s no universal figure. Watch dynamics vs your margin and deal cycle — not someone else’s case from a pitch deck.
How long should I wait for results?
The first weeks are data and fixes. If after two or three months with a clear brief and budget there’s no movement toward KPI — review hypotheses and the contractor.
Do I need weekly micro-control?
Usually a monthly plan and a regular report are enough. Daily oversight slows work when there’s no incident.
Is weak revenue always the agency’s fault?
No. Season, a fuzzy brief, a weak site or offer, a narrow niche. But no hypotheses and opaque reports are already a red flag.
What should I require in the contract?
Access, scope of work, lead and KPI definitions, report deadlines, budget rules. A ranking guarantee with no metrics is empty noise.
Lots of clicks from the agency — still no clear CPA story?
We’ll lock access, lead definition, and a monthly hypothesis report so you control results — not pretty screenshots.
Discuss the task
