The short answer
Start with what one deal is worth. My target is for every shekel spent on ads to come back 8 to 12 times in revenue (a ROAS of 800%–1200%). So for a 200,000 ₪ deal, you can spend roughly 17,000–25,000 ₪ on ads to land it and still make a healthy profit. The budget comes from the deal’s value, not from what feels like a reasonable monthly spend.
How much can you spend to win one deal?
Divide the deal value by 12 and by 8. The first number is your conservative scenario; the second is your ceiling. The budget you can invest to land one deal sits between them.
| Deal value | Budget at 1200% ROAS | Budget at 800% ROAS |
|---|---|---|
| 200,000 ₪ | About 16,700 ₪ | 25,000 ₪ |
| 500,000 ₪ | About 41,700 ₪ | 62,500 ₪ |
This isn’t a monthly budget. It’s a per-deal budget. One deal might come in after a month, another after three. What matters is that by the end of the period, the money coming in is at least 8 times what went out.
The math is also reassuring. A 15,000 ₪ ad bill can be scary. But once you know a single deal covers it 13 times over, the question changes: not how much it costs, but whether the campaign brings in that deal.
Why a 1,000 ₪ lead can be cheap
Because the question isn’t what a lead costs. It’s how many leads you need to close a deal. If 15 leads at 1,000 ₪ each produce one 200,000 ₪ deal, you spent 15,000 and brought in 200,000. That’s within range.
Now compare that to a campaign bringing in 300 leads at 50 ₪ each. Same budget, except now someone has to call 300 people. And if most of them asked a surface-level question and disappeared, the cheap lead is actually the most expensive one, because it also costs your sales team hours. I wrote more about this in the post on cost of sale.
So when you set a budget, stop asking “How many leads will this get us?” and start asking “How many deals, and how much revenue, do I expect this campaign to generate?”
Why your campaign needs time to learn
Google and Meta learn who leaves their details, then look for more people like them. To learn, they need data. Meta, for example, recommends about 50 conversions a week to exit the learning phase.
With a high-ticket product and ten quality leads a month, you won’t hit that number. That means learning is slower, and the first month is almost always weaker than the third. If you stop after a month because “there were no deals,” you may be pulling the plug right when the campaign starts figuring out what it’s looking for.
There’s another reason not to judge too fast: with high-ticket products, the deal usually closes well after the lead comes in. A lead from month one might only turn into a deal in month four. So run the revenue test over a period, not a single month.
What’s the minimum budget I work with?
5,000 ₪ a month, before my fee.
The reason is simple: I run ads on both Google and Meta, because each one reaches the customer at a different moment. 5,000 ₪ a month is about 165 ₪ a day, so a little over 80 ₪ a day per platform. Below that, each platform gets too little budget to reach enough people and learn who’s interested, and the results just aren’t enough. It’s not that you can’t advertise with less. It’s that with less, I can’t promise the revenue test will pass.
And that’s a minimum, not a recommendation. The right budget comes from the math above: what a deal is worth, and how many deals a year you need for the ads to pay for themselves. In a campaign for Y.D. Spectrum, a lighting engineering firm, a modest budget brought in ten leads with a potential value of about 300,000 ₪. It’s too early to know how much of that will close, but it shows how far a modest budget can go when it reaches the right people.
FAQ
What is ROAS?
The ratio between revenue from deals and what you spent on ads. A ROAS of 800% means every dollar spent brought back 8 dollars.
What if you don’t have data on deal value yet?
Start with an estimate. Even if it’s not precise, it beats a budget set by gut feeling. After a few deals, update it with the real numbers.
If I increase the budget, will I get more deals?
Sometimes. But in a small market, like a B2B market with a few tens of thousands of potential customers, you reach a point where you’re spending more money on the same people. So scale up gradually, and check that the revenue test still hits the target.
How long should you wait before judging a campaign?
It depends on the business’s sales cycle. The longer it takes to go from lead to deal, the longer the measurement period needs to be.
