How Much Should an Established Business Spend on Paid Ads?

Why rules of thumb fail

The most common budgeting method for paid ads is some version of "a share of revenue" or "what we spent last year, plus a bit". Both are easy to approve and both are disconnected from the only thing that matters: whether the money buys customers at a price the business can afford.

A share-of-revenue rule gives a high-margin consulting firm and a low-margin distributor the same proportion, even though one can afford to pay many times more for a customer than the other. Last year's number carries forward last year's mistakes. And copying a competitor's apparent spend assumes they know what they are doing, which is generous.

This article is about the paid media line specifically. For the whole marketing budget (people, brand, web, content and media together), start with how to build a marketing budget around business goals, then use this framework to size the advertising portion within it.

Step one: what a customer is worth to you

Start with gross margin, not revenue. If a new client brings in revenue but most of it goes to materials, subcontractors or delivery staff, only the margin is available to pay for acquiring them.

Then decide on a time horizon. A business with strong repeat purchase or multi-year contracts can reasonably count more than the first sale. A business that sells once and rarely sees the customer again should count only that sale. Be conservative here: an optimistic lifetime value is the easiest way to justify spending money you will not get back.

From that margin, decide what share you are willing to spend to acquire a customer. This is a business decision, not a marketing one. It depends on your cash position, how fast you need to grow and what else competes for the same money. The result is your allowable cost per customer: the most you can pay for a new customer and still be glad you did. Many established businesses have never written this number down, and it changes every conversation about advertising once they do.

Step two: work back through your funnel

Paid ads do not deliver customers. They deliver clicks, which become inquiries, which become qualified leads, which become customers. Each step loses people, and you need to know roughly how many.

Here is an illustrative example. Say a 30-person engineering consultancy in Moncton looks at the last year of inquiries and finds that about one in three was a real, qualified opportunity, and that it won about one in four of those. That means it needs roughly twelve inquiries to win one client. If its allowable cost per customer is a given amount, its target cost per inquiry is one-twelfth of that.

That target cost per inquiry is the number the advertising can actually be managed against. It lets you look at any campaign and say whether it is working, without waiting months for deals to close.

If you do not know your conversion rates, that is the first problem to solve, and it is a measurement problem rather than an advertising one. A rough estimate from the sales team is better than nothing; a few months of tagged inquiries in your CRM is much better. See how to measure lead quality from paid campaigns for how to set that up.

Step three: check against real demand

The economics tell you what you can afford per inquiry. They do not tell you how many inquiries exist. That depends on the market, and in Atlantic Canada the market is often the binding constraint.

• For search ads, demand is the number of people actually searching for what you sell, in the places you serve. Once you are showing for most of those searches, more budget mostly raises what you pay for the same clicks.
• For social and LinkedIn ads, the limit is the size of the audience you can reach and how often you can show them something before they tune it out. A narrow B2B audience in one province can be saturated quickly.
• For seasonal businesses, demand comes in a window. Spending evenly across the year means overspending in the off-season and running out of budget at the peak.

Multiply your target cost per inquiry by the number of inquiries the market can realistically supply at that cost. That is your demand-limited budget. Your working budget is the lower of what the economics allow and what demand supports. If the economics allow more than demand supports, you have room to try a second channel. If demand supports more than the economics allow, the job is improving conversion rates or margin, not spending more.

Step four: separate the test budget from the run budget

A new channel, campaign or market has no history, so the framework above cannot be applied to it with any confidence. Treat it as a test with its own budget.

A sensible test budget is large enough to produce a meaningful number of conversions in a fixed period, and small enough that losing it would not hurt. Too small, and you end the test with a handful of results and no conclusion. Define success before you start: the target cost per inquiry, the minimum number of inquiries, and the date you will decide.

Once a campaign has a track record, it moves to the run budget, which is managed against the target cost per inquiry and reviewed monthly. Keeping the two separate stops experiments from being judged by run standards (and killed too early), and stops proven campaigns from being starved to fund experiments.

The costs outside the media spend

The money paid to Google, Meta or LinkedIn is the visible part of the budget. Several other costs decide whether it works:

• Management. Someone has to review search terms, adjust bids and audiences, refresh creative and report honestly. Whether that is a staff member's time or an outside partner, it is real.
• Creative. Social and LinkedIn ads wear out. A plan that spends steadily on a single set of ads for a year will see results decay.
• Landing pages. A page built for the campaign usually converts better than a general page, and it has to be designed, written and maintained.
• Tracking. Conversion set-up, call tracking and CRM connections take time to build and to keep working as the website changes.
• Sales capacity. More inquiries need someone to answer them quickly. If the sales team is already stretched, the extra leads will cool before anyone calls.

A credible total needs scoping against your goals and your current set-up. Anyone who quotes a media budget without asking about your margins and conversion rates is guessing.

Signs the budget is wrong

Too low

• Campaigns produce so few conversions that nobody can tell what is working.
• Ads stop showing partway through the day or the month while demand is still there.
• Cost per inquiry is comfortably below target and the sales team has capacity, but nobody has asked for more.

Too high

• Cost per inquiry has risen as spend increased, and the latest increase bought fewer inquiries than the one before.
• The same people see the same ads many times, and response is falling.
• Lead quality has dropped: more inquiries, fewer qualified ones.
• Sales cannot keep up with follow-up.

If you are seeing the second list, the answer is usually not a smaller budget everywhere but a closer look at where the last increment went. That is the subject of when to scale an ad budget and when to fix the funnel. And if the spend looks reasonable but you cannot see what it produced, start with the paid media audit checklist.

Putting it on one page

The whole framework fits on one page, and leadership should be able to read it in a few minutes:

1. Gross margin per new customer, over a stated time horizon.
2. The share of that you will pay to acquire one (your allowable cost per customer).
3. Inquiry-to-qualified and qualified-to-customer rates, with the source of each number.
4. Target cost per inquiry.
5. Estimated monthly inquiries available at that cost, by channel and by season.
6. Run budget, test budget and the non-media costs.
7. The date and the criteria for the next review.

When we plan paid advertising for a client, this page comes before any campaign is built, because it is what makes the monthly report mean something. It also makes budget conversations shorter: instead of debating a number, you are debating an assumption, and assumptions can be checked.