The standard answer to this question is a percentage of revenue — commonly somewhere between five and ten percent, higher if you are growing fast, lower if you are established with a strong referral base. It is a reasonable sanity check and a poor decision-making tool.
The reason it is poor is that it tells you how much to spend without telling you whether spending it works. Two companies at 8% of revenue can be running entirely different businesses: one is buying jobs profitably and should spend more, the other is subsidising a channel that does not convert and should stop today.
Start from cost per job, not percentage of revenue
The number that actually drives the decision is what it costs you to acquire one job, compared with what that job is worth. Everything else is commentary.
- 01Total marketing spend for a period
Everything: ads, the LSA invoice, the website, the van wraps amortised, the sponsorship, the person who runs it. Owners routinely undercount by half by forgetting labour and fixed costs.
- 02Divide by jobs won in that period
Attributable jobs. If you cannot attribute them, that is the first problem to fix, and it is worth more than any budget decision.
- 03Compare to gross profit per job
If a job produces $400 of gross profit and costs $120 to acquire, the channel works. If it costs $380, you are working for the privilege.
- 04Then do it again with lifetime value
For any business with repeat work, a first job that barely breaks even can be excellent if the customer returns for years — but only if you actually retain them, which is a different problem.
Fix the leaks before increasing the spend
This is the part that saves the most money and is the least satisfying advice. Marketing spend multiplies whatever your conversion process does — and most service businesses have a conversion process with large, cheap-to-fix holes in it.
- Unanswered calls. Every missed call is a lead you already paid for, handed to a competitor. Fixing this raises the return on every channel simultaneously and costs almost nothing.
- Unfollowed quotes. Roughly half of home-service estimates never get a single follow-up. Spending more to generate quotes you will not chase is the most expensive mistake in this list.
- Untracked sources. If you cannot say which channel produced which job, you are allocating by anecdote, and anecdote consistently over-credits whatever is most visible.
- No repeat contact. Past customers are the cheapest demand available and most companies contact them never.
A business that fixes those four typically finds it can grow for a year on the same spend — which is a better outcome than any budget increase, and it makes the eventual increase far more productive.
Where the first dollars should go
Roughly in this order, because it is roughly the order of return. The paid channels come last deliberately.
- Response capability. Missed-call text-back and a queue where leads cannot be lost. It is not marketing, and it raises marketing returns more than anything else you can buy.
- Google Business Profile and reviews. Free, compounding, and the highest-intent local traffic there is. A steady review trickle beats any campaign.
- Past customers. A list you already own, a message that costs nothing, and the highest close rate in the business.
- A referral programme. Cheap, high-quality, and almost always under-run relative to how well it works.
- Paid acquisition. Local Services Ads, search ads, paid social — genuinely useful once the four above are working, and an expensive way to hide their absence if they are not.
TARGETS
Seasonality changes the shape, not the size
Spending evenly across the year is a common and costly default in seasonal trades. The demand curve is not flat, so the spend should not be either — and the instinct to advertise hardest when you are busiest is usually backwards.
Two adjustments are worth making. Spend ahead of the season rather than during it, because the customer deciding in February is the one you want in April and the competition for attention is lower. And in the peak, when you are turning work away, the better use of money is often retention and price rather than acquisition — there is no value in buying leads you cannot serve.
What to look at monthly
| Number | What it answers | When to act |
|---|---|---|
| Cost per lead, by source | Which channel is getting expensive | When one drifts up while others hold |
| Lead-to-job conversion, by source | Which channel sends real buyers | When a cheap source stops converting — cheap leads are not cheap |
| Cost per job acquired | Whether the whole thing works | When it approaches gross profit per job |
| Share of jobs from past customers and referrals | Whether you are building or renting demand | When it falls — you are becoming dependent on paid |
The last row is the strategic one. A business where most work comes from referrals and past customers has an asset. A business where most work comes from paid channels has a subscription, and the price of that subscription is set by somebody else.
A budget is not a number you pick. It is what your cost per job tells you that you can afford.
Common questions
What percentage of revenue should a service business spend on marketing?
Five to ten percent is the common benchmark — higher when growing fast, lower with a strong referral base — but it is a sanity check rather than a decision. Two companies spending the same percentage can be running opposite businesses: one buying jobs profitably and one subsidising a channel that does not convert.
How do I calculate cost per job acquired?
Total marketing spend for a period — including labour, website, and fixed costs that owners usually forget — divided by attributable jobs won in that period. Compare it to gross profit per job, then repeat the comparison against realistic lifetime value if your business has genuine repeat work.
Where should a service business spend its first marketing dollars?
On response capability first — missed-call text-back and a lead queue — because it raises the return on every channel at once. Then Google Business Profile and reviews, past customers, and a referral programme. Paid acquisition comes last: it works once the others do and is an expensive way to hide their absence if they do not.
Should marketing spend be even across the year?
Not in a seasonal trade. Spend ahead of the season rather than during it, since the customer deciding in February is the one you want in April and attention is cheaper then. At peak, when you are turning work away, retention and pricing usually beat acquisition — there is no value in buying leads you cannot serve.