Ask five agencies how much you should spend on Google Ads and you'll get five different rules of thumb — usually some version of "5-10% of revenue." That's not wrong, exactly, but it's not useful either. It doesn't account for how competitive your keywords are, what a job is actually worth to you, or how good your team is at closing the leads that come in.
Here's a better way to think about it.
Start from your close rate and average job value, not your revenue
The real question isn't "what percentage of revenue is normal" — it's "what does a booked job cost me to acquire, and can I afford that at volume." Work backward:
- Average job value. What's a typical job worth, after any recurring or repeat-business value you can reasonably attribute? A plumbing repair might be $350. A dental implant case might be $4,000+. A signed personal injury case might be worth tens of thousands over the life of the matter.
- Close rate. Of the leads or calls that come in, what percentage actually become paying customers? Most local service businesses close somewhere between 20-40% of qualified inbound leads, though this varies enormously by industry and how fast you follow up.
- Target cost-per-acquisition (CPA). A reasonable target is to keep your cost-per-booked-job at 10-20% of the job's value for one-off, lower-consideration services, and lower than that for big-ticket or recurring work where margins are healthier. A $350 job with a 15% CPA target means you can afford roughly $50 per booked job.
- Divide by close rate to get cost-per-lead. If you close 30% of leads and can afford $50 per booked job, you can afford up to $15 per lead.
That number — your affordable cost-per-lead — is what actually determines your budget, once you know what leads cost in your market.
What drives cost-per-click (and therefore cost-per-lead)
Google Ads pricing is an auction, and some industries are simply more expensive than others because more businesses are bidding for the same searches. As a general pattern (not a guarantee — your local market will vary):
- Legal is consistently among the most expensive verticals on Google Ads. High-value practice areas like personal injury and criminal defense routinely see cost-per-click well into double digits, sometimes $50-100+ for the most competitive terms, because a single signed case can be worth a great deal.
- Home services (HVAC, plumbing, electrical) tend to sit in a wide but more moderate range, often single digits to low double-digit CPCs, with emergency/urgent terms costing more than routine maintenance searches.
- Dental and medical fall somewhere in between, with cosmetic and high-value procedures (implants, Invisalign) costing more to bid on than routine cleanings.
- Auto dealers often run lower CPCs per click but need more volume, since a single click is further from a purchase decision.
Multiply your target cost-per-lead by roughly how many clicks it typically takes to generate one lead (your conversion rate on the landing page matters a lot here — a page that converts at 3% needs three times the clicks of one that converts at 9% for the same number of leads) and you'll land on a defensible monthly budget instead of a guess.
The budget floor: why "just start small" often doesn't work
There's a practical floor below which Google Ads struggles to work well in competitive verticals: if your daily budget can't sustain enough clicks to get through the auction's learning phase and generate a handful of conversions a week, the algorithm doesn't have enough signal to optimize, and you end up paying premium rates for erratic results. For most local service categories, that floor tends to land somewhere between $1,500 and $3,000/month — below that, dial back scope (fewer keywords, a tighter service radius) rather than spreading a too-small budget across too many terms.
What to do with this
Run the math above with your own numbers before setting a budget, and revisit it every quarter as your close rate and average job value change. If a channel isn't hitting your target cost-per-booked-job after a genuine testing window (usually 60-90 days, long enough for the algorithm to optimize and for you to have a real sample size), that's a signal to adjust targeting or creative — not necessarily to spend more.