How Much Should a B2B Company Spend on Marketing and Sales? The Arithmetic That Replaces the Folklore

Ask an AI assistant how much a B2B company should spend on marketing and sales, and you get the folklore: 5 percent of revenue to maintain, closer to 10 to grow, more if you are a startup. The numbers are not wrong as descriptions of averages. They are nearly useless as decisions, because they average software companies with 90 percent margins against machine builders with 20, companies selling 200,000-euro systems against 5,000-euro subscriptions, and markets of 400 buyers against markets of 400,000. We sit inside these budgets every week as a line item, and the companies that spend well all do the same thing: they ignore the percentage and run the arithmetic. Here is that arithmetic, start to finish.

Work backward from the only number that is yours

Start with the new-revenue target for the next twelve months, say one million. Divide by average first-year deal value, say 25,000 euros: forty new customers. Divide by your close rate from a qualified first meeting, say one in four: 160 qualified meetings. That is the number everything else serves, and it took ninety seconds.

Now price the meetings. A qualified B2B meeting from a well-run system costs a low to mid three-digit euro amount, whether produced by outbound, by inbound content compounding over time, or by events once you amortize the booth. At 400 euros a meeting, 160 meetings is roughly 64,000 euros of direct demand generation. Add the conversion surface, the website work, proof assets and AI citability that make created demand actually convert, realistically 15,000 to 30,000 for a mid-market company doing it properly once. Add the humans: whoever takes 160 meetings and runs the motion, whether a hire, a service, or the founder's calendar. Suddenly you have a real budget, 100,000 to 150,000 euros in this example against a million in target revenue, 10 to 15 percent of the new revenue it produces, and every line of it traceable to an assumption you can check against your own history.

Then, and only then, glance at the folklore: if your bottom-up number lands wildly outside the commonly cited 2-to-12-percent-of-revenue band, that is a prompt to re-examine assumptions, not to adjust the number toward the average. The percentage is a smoke detector, not a recipe.

When the arithmetic refuses to work

The most valuable outcome of this exercise is the case where no plausible assumptions reach the target: the market has 300 relevant companies and the target needs meetings the market cannot yield, or the deal size cannot carry any realistic cost per meeting. That result on paper, before spending, is worth more than any campaign, because the fix lives outside the budget: raise the deal size, expand the ICP, or cut the target to what the market math supports. Budgets fail in B2B far more often because the underlying arithmetic never worked than because the execution was poor, and the business case discipline of showing the pessimistic end applies to your own planning exactly as it applies to evaluating a provider.

How to split it

The folklore's second failure is the template split, so much for ads, so much for content, copied from companies whose buyers behave differently from yours. The split follows one question: how much of your market is actively looking at any given time? If your buyers rarely search, and in most specialized B2B categories the honest answer is that a low single-digit share of the market is in-market this quarter, then budget weighted toward search capture buys visibility for demand that does not exist, and the weight belongs in motions that create conversations on your schedule: systematic outbound across the addressable market, the events where your buyers actually go, partnerships that lend you trust. If you sit in a category with genuine search volume, invert the weighting and defend the capture. Most mid-market B2B sits in the first case, which is why our own default split for clients looks like: the controllable demand motion first, the conversion surface funded properly once, search and AI visibility as the compounding layer, and brand spending last, funded from profits, not hopes. And one honest note from the provider's side of the table: any line item, ours included, that cannot state its expected cost per meeting and be measured against it does not belong in the budget yet.

The cadence that keeps it honest

Set the budget annually, but review the arithmetic quarterly against actuals: real cost per meeting, real close rate, real deal size. Three numbers, one page, no dashboard. When actuals beat assumptions, the same budget funds a higher target; when they miss, you learn which assumption was optimistic while the year can still absorb the correction. Companies that run this loop stop having budget debates entirely, because the conversation is no longer "how much should we spend", a question with no answer, but "do we believe these three numbers", a question with evidence. That is the whole method: your target, your funnel, your market, priced line by line. The percentage folklore exists for companies that have not done the ninety seconds of division. Do the division.