How much should I spend on Google Ads? The maths, not a range
Every agency gives you a range. A range is useless without your numbers. Here is how to work out your own floor in about ten minutes.
Premkumar Ragav
"What budget do I need?" is the most common question I get, and the most common answer — "it depends" — is true and completely useless. So here is the actual calculation.
Start from what the algorithm needs, not what you can afford
Smart Bidding is a machine-learning system, and like any such system it needs a minimum volume of examples before its predictions beat guesswork. Google's own guidance has long pointed at roughly 30 conversions in 30 days per campaign.
Below that, three things happen, all bad: bids stay volatile, cost per acquisition swings week to week, and you cannot tell a real trend from noise — so you start changing things at random, which resets the learning again.
The calculation
You need three numbers. Two you can look up; one you estimate.
- Cost per click for your main keywords. Use Keyword Planner, and take the top of the range rather than the average.
- Landing page conversion rate. If you do not know it, assume 3% for an unoptimised page and 8–10% for a purpose-built one.
- Target conversions: 30.
Monthly floor = CPC ÷ conversion rate × 30
A plumber at £4.50 a click with a 6% conversion rate: 4.50 ÷ 0.06 × 30 = £2,250 a month. A dental implant clinic at £12 a click with 5%: 12 ÷ 0.05 × 30 = £7,200.
Worth knowing: this is a floor, not a recommendation. It is the point below which you are paying for learning you never finish, not the point at which the account is doing well.
Now check it against what a customer is worth
The floor tells you what the platform needs. This tells you whether it is worth paying.
Break-even cost per lead = customer value × close rate
If a customer is worth £2,000 in gross profit and you close one in four leads, you can pay up to £500 a lead before losing money. If your calculated CPL is £75, you have a very healthy business case. If it comes out at £600, the channel does not work at your current close rate — and the fix is the sales process, not the ad account.
| Business | CPC | CVR | Monthly floor | Resulting CPL |
|---|---|---|---|---|
| Local plumber | £4.50 | 6% | £2,250 | £75 |
| Dental implants | £12.00 | 5% | £7,200 | £240 |
| B2B software | £8.00 | 3% | £8,000 | £267 |
| Home services (AU) | A$6.00 | 7% | A$2,570 | A$86 |
If the floor is more than you have
This is common, and it is not the end of the conversation. You have four honest options.
- Narrow the geography. One city instead of a region concentrates the same budget into enough volume to learn.
- Narrow the service. Advertise your highest-margin service only. Fewer keywords, higher intent, cheaper clicks.
- Fix the conversion rate first. Doubling a 3% page to 6% halves the floor. This is the cheapest lever available and almost nobody pulls it first.
- Use Meta instead. Cheaper traffic, lower intent — but at small budgets a well-built Meta campaign often beats an under-fed Search campaign.
What not to do
Do not split a small budget across Google and Meta so that neither reaches its learning threshold. Two half-fed campaigns perform worse than one properly fed one, every time. Pick the channel that fits your intent profile, get it working, then expand.
And do not judge the account in week one. Even a correctly budgeted campaign needs four to six weeks before the data is worth reading.
Frequently asked
Can I start with a smaller budget and scale up?
You can, provided you accept the first month is a data-gathering exercise rather than a performance test. What does not work is starting small, judging it on week two, and concluding the channel does not work — that is the most common and most expensive mistake.
Does the 30-conversions rule apply per campaign or per account?
Per campaign, and more precisely per bid strategy. That is exactly why running six campaigns on a small budget performs so much worse than running one: you divide your conversion data into six piles, none of them big enough to learn from.
Should I count leads or sales as conversions?
Start with leads because the volume is high enough to learn from. Once you have consistent volume, import offline conversions from your CRM so the algorithm optimises toward leads that actually closed, not just leads that submitted a form.