Everyone asks me the same question. How much should I spend on ads?

01Why nobody can give you 1 number

Picture 2 service businesses, both ready to start ads. One sells a $400 one off job. The other sells a $6,000 piece of work that renews every year.

The right starting budget for those 2 businesses is not even close. They are not in the same universe. And no article, including this one, knows which of the 2 you are.

Your budget is an answer your business gives you. It is not a number you choose first.

The only honest general thing I can say is this. Your budget sits below what a client is worth to you. Get that number first and the budget stops being a mystery.

02Step 1 of 4. Fix the offer, not the budget

Before any of this matters, the offer has to sell. Meta is a loudspeaker, not a magic wand. If the offer is weak, ads just help more people ignore you faster, and you paid for the privilege of learning that in front of a big crowd.

Here is the test. Have you sold this exact thing, at this exact price, to strangers who did not already know you? If the answer is no, your money is better spent having conversations than running a campaign.

Spending to test an offer nobody has bought yet is the most common way a service business decides that ads do not work.

A lot of businesses under $50k a month have an offer problem wearing an ads costume. The ad account is where the problem becomes visible. It is not where the problem was caused.

03Step 2 of 4. What 1 client is really worth

Not what they pay you the first time. What they are worth over the whole relationship. Renewals, extra work, and referrals if you can honestly point to them.

Most service businesses guess this far too low, which is exactly why they spend far too little. If a client pays $2,000 and stays 2 years at that rate, you are not fighting for a $2,000 customer. You are fighting for a $4,000 one. And you can pay a lot more to win them than the person who only counted the first invoice.

That is the whole reason some competitors seem to outbid you everywhere. They are not braver. They just did the sum properly.

04Step 3 of 4. How fast they pay you back

This is the number that actually decides how fast you can grow, and almost nobody works it out.

Payback is how long it takes a client to return the money you spent getting them. Spend $500 to win a client who pays $250 a month, and you are back to even in 2 months. Everything after that is profit. But for those 2 months, that $500 is gone from your bank.

Now make it bigger. 10 clients a month means $5,000 out and a 2 month wait. 30 clients a month means $15,000 out. The thing holding you back is not whether the ads work. It is whether you can carry the gap. Businesses with working ads run out of cash all the time, and this is why.

Both of those businesses make a profit. Only 1 of them can grow quickly without borrowing. Getting paid back sooner, usually by taking more of the fee upfront, is often a bigger win than cutting the cost of getting a client.

The ads are not the hard part. Carrying the gap while the money is still on its way is.

05Step 4 of 4. Budget for a 10 day test

Now the practical bit. A test has to produce enough results to tell you something, over about 10 days.

Say your real cost per booked call is going to land near $50, and you need 15 to 20 results before a number means anything. Then your 10 day test has to be able to produce that many. A budget that gives you 2 results in 10 days has not told you your ads do not work. It has told you nothing at all, and you will probably read that nothing as a failure.

So the real minimum is not a fixed number in dollars or rupees. It is whatever buys you a readable answer in your market, at your likely cost per result. If you cannot afford that, run fewer angles at once instead of starving all of them.

That number is yours. It came out of your own business, which is why it is worth more than anything I could have put in my first line.

06Month 1, and what happens when it works

Treat month 1 as school fees, not as profit. You are buying information. Which groups of people respond, what your real cost per result is, where people quietly give up on your booking page. That information is what makes month 3 profitable.

So pick an amount you could lose completely without it hurting the business. If no such amount exists, you are not ready for paid ads yet. That is a real answer, not a soft one.

When it does work, everyone wants to know how fast to raise the spend. Honest answer: as fast as your cash and your delivery team allow, which is usually slower than your ambition and faster than your nerve.

Watch 2 things as you grow. Your cost per result will creep up, because more spend means reaching colder and more doubtful people. And your delivery will hit a wall that nobody plans for until it arrives. Doubling your spend into a delivery wall is how good businesses win clients they then let down.

As you reach those colder people, proof matters more than promise. The ads that worked on the people who half knew you will not carry a total stranger. That is a creative problem, not a budget problem. And the stranger will also go and check you out before they book, which is its own job. People now ask a chatbot who to hire, and whether it says your name depends on what you have published, not on what you spend.

07How I keep these numbers in 1 place now

In 2026 I do not keep any of this in my head or in 6 spreadsheets. Every business I work with gets a command center, an AiOS, where the numbers, the clients and the content all sit in 1 place. The cost to win a client and the payback window live there, next to the work.

On top of that sits a company brain, which is just the business remembering things properly. What we sold, at what price, to whom, and what happened next. When the brain holds that, the ad budget question takes 10 minutes instead of a week of arguing.

And I run Telegram assistants over it, so the owner can ask in plain words and get the number back on their phone. Same maths as this article. Less digging.

None of that replaces step 1. A command center cannot save a weak offer either. It just means you stop guessing about your own business, which is where the budget question started.

08Questions, answered