How much to spend is the question everybody asks first. It is also the easier half, and a business can get the percentage right and still have a budget that collapses the moment demand drops.
What follows is about the structure rather than the size: how it is split, what belongs in it, and what happens to it in February.
Put your hours in it
The largest line in most small marketing budgets is invisible, because it does not leave the bank account.
If you spend six hours a month writing posts, chasing reviews, answering platform enquiries, and updating listings, and your billable rate is $80, that is $480 a month of marketing cost that appears nowhere.
Count it. Not because you will pay yourself, but because it changes decisions. A channel costing $50 a month and eating eight hours is more expensive than one costing $300 and running itself, and no budget that ignores hours will ever show you that.
Three buckets, not one pot
Split the money by what it does, because the three behave completely differently and are judged on different timescales.
Capture. Spending that reaches people who already have the problem. Local Services Ads, search ads, the work behind your Google Business Profile. Judged on cost per booked job, within weeks.
Compounding. Things that build an asset you keep. Content, service area pages, review generation, photography. Judged on impressions and enquiries over quarters, never on this month.
Awareness. Spending that reaches people before they need you. Streaming ads, direct mail, sponsorship, vehicle wraps. Judged over seasons, and honestly not very measurable at all.
A workable starting split for a business with an established profile is roughly half capture, a third compounding, and the remainder awareness. If your profile is incomplete or your reviews are thin, the compounding share should be larger, because awareness spending against a weak foundation is the most reliable way to waste money in this industry.
Respect the floors
Several channels have a minimum below which they do not work at all, and spending under it buys nothing rather than buying less.
Search ads need roughly $1,000 a month before an account gathers enough conversion data to optimise against. Radio needs meaningful frequency, which starts near $900 a week in a small market. Streaming television needs about ninety days and a real monthly number to mean anything.
So a $600 monthly budget spread across search, radio, and streaming produces three channels that each do nothing. The same $600 in one place, above its floor, can work.
The rule that follows: fewer channels, funded properly. If the budget will not clear a channel's floor, it does not belong in the plan yet.
Shape it to your year, not to the calendar
A fixed monthly spend is the most common structure and it is wrong at both ends.
Home service demand swings 40 to 60 percent month to month in a typical trade, with peak-to-valley variation reaching 250 to 600 percent in some. Against that curve, flat spending under-invests when demand exists and over-invests when it does not. One analysis puts the cost at leaving roughly 40% of peak revenue unclaimed while overspending around 60% in a quiet month.
The correction is to shift capture spending toward your peak and compounding spending toward your trough, offset by each channel's lag. If you want to own July, the content work happens in March and the ad budget rises in June.
Build the reserve in deliberately
This is the part that makes a budget survive a slow month rather than being abandoned during one.
Decide in advance what you will cut and what you will protect. Without that decision, the thing that gets cut in a bad February is whatever is easiest to cancel, which is almost always the compounding work, which is exactly the thing that would have produced March.
A defensible order: awareness goes first, capture reduces in proportion to demand, compounding is protected. That is the reverse of instinct, because compounding is the spending with no visible immediate return, which is precisely why it feels cuttable.
Hold back roughly ten percent unallocated. Something will need it, and a budget with no slack turns every opportunity into a trade-off.
What is not a marketing budget item
Two things get miscounted constantly.
Field service software. Jobber, Housecall Pro, ServiceTitan and similar are operations costs. They make you better at delivering work, not at getting it. Putting them in the marketing budget inflates the number and makes your marketing look more expensive than it is.
Payment processing. A percentage of revenue, and a real cost, but it is a cost of taking money rather than of finding customers.
Conversely, one thing that does belong and usually is not there: the time and cost of following up on quotes. That is a marketing function in everything but name, and it is frequently the highest-return activity available.
Review it on the right clock
Quarterly, not monthly. Monthly reviews punish anything with a lag, and channels report at wildly different speeds.
At each review, ask three things. What did each channel cost per booked job. Which channel is worst, and is it worse than you assumed. And is anything still running only because cancelling it never reached the top of the list.
That last question retires more wasted spending than any optimisation.
The reframe
A marketing budget is not a number, it is a set of decisions about what to fund, in what order, and what to protect when things get tight.
Most small businesses have the first part and none of the rest, which is why a slow month tends to produce a panic cut rather than a plan. Decide the order now, while nothing is going wrong, because the version of you in February will make a worse decision than the version reading this.