"How much should we spend on Google Ads?" — asked by every founder, answered honestly by almost no one, because the honest answer is a framework, not a number. Here's the framework I use to set budgets for brands from $500/month to $50K/month.
Start from the math, not the wallet
Your minimum viable budget is determined by three numbers you already have (or must estimate):
- Average CPC in your market (Keyword Planner gives ranges; real CPCs usually land near the high end).
- Expected conversion rate — 2–5% for most sites; use 2% until proven otherwise.
- Your unit economics — what a customer is worth (first order, or LTV if you can stomach the payback period).
Worked example: $1.50 CPC ÷ 2% conversion rate = $75 per conversion before optimization. If a customer is worth $60, don't launch — fix the offer, the landing page, or the market first. Ads amplify economics; they can't invent them.
The data-velocity floor
This is the part almost everyone misses: your budget determines how fast you learn. Smart Bidding needs roughly 30–50 conversions per month per campaign to work properly. Below that, you're not really running an optimization system — you're running an expensive coin flip that never converges.
- Minimum learning budget = 30 × expected CPA. At a $20 CPA, that's $600/month for one campaign. At a $75 CPA, it's $2,250.
- Can't afford the floor? Narrow scope until you can: one campaign, one tight theme, one geo. Small budget spread thin learns nothing; small budget concentrated learns fast.
A $1,000 budget on one campaign is an experiment. The same $1,000 across six campaigns is a donation.
The three-phase budget lifecycle
- Phase 1 — Learning (months 1–2): Budget = 30–50 conversions/month, concentrated on your highest-intent segment. Goal is signal, not profit. Judge nothing weekly; judge everything monthly.
- Phase 2 — Efficiency (months 2–4): Same budget, falling CPA. Trim query waste (audit checklist, part 3), tighten targets gradually — 10–15% at a time, never 40%.
- Phase 3 — Scale (month 4+): Raise budget 20% at a time, waiting ~2 weeks between raises so Smart Bidding recalibrates. Watch marginal CPA, not average CPA — when the next dollar converts meaningfully worse than the last, you've hit this structure's ceiling, and more budget needs a bigger architecture, not a bigger number.
Signals it's time to raise the budget
- Search impression share lost to budget (not rank) on converting campaigns — you're literally turning away buyers.
- CPA stable or falling for 4+ weeks at current spend.
- Campaigns hitting daily caps before evening in your best timezone.
Signals more money won't help
- Impression share already >85% on money terms — the demand pool is drained; expansion, not budget, is next.
- CPA rising as spend rises with no structural change — you're buying deeper into worse traffic.
- Tracking gaps >25% — you'd be scaling blind; fix measurement first.
The honest summary: budget isn't a commitment to Google, it's a dial on how fast you buy information — then customers. Set it by CPA math, hold it until the system learns, and raise it only when the data says the ceiling isn't here yet.
Frequently asked questions
What's a good starting budget for Google Ads?
Enough to generate 30–50 conversions a month at your expected CPA. For a $25 CPA, that's roughly $750–1,250/month. Below that floor, narrow your targeting until the budget you have can hit it.
How much should I spend on Google Ads as a percentage of revenue?
Most healthy accounts land between 5–15% of revenue, but this varies hugely by margin and growth stage. Use CPA-times-volume math first; treat the percentage-of-revenue rule as a sanity check, not a starting point.
When should I increase my Google Ads budget?
When you're losing impression share to budget (not rank) on converting campaigns, and CPA has been stable or falling for at least four weeks. Raise it 20% at a time and wait ~2 weeks between raises.
Note that this math changes for longer sales cycles — see how B2B SaaS budgeting handles a 30-50 conversion floor when the real conversion only happens a handful of times a month.