Three Ways Budgets Get Set, Two Of Them Indefensible
Almost every B2B marketing budget is set by one of three methods, and only the third holds up when someone competent asks how the number was reached.
| Method | How it works | Why it survives or fails |
|---|---|---|
| Percentage of revenue | Last year's revenue, or next year's plan, multiplied by a customary share. | Fails. It makes spend a consequence of past performance rather than a cause of future performance, so it cuts investment exactly when growth stalls. |
| Competitive parity | Estimate what comparable firms spend and match it. | Fails. The inputs are guesses, the comparison set is chosen to justify a conclusion, and matching a competitor assumes their allocation is right and their objectives are yours. |
| Objective and task | State the outcome, decompose it into the activities required, cost each activity, sum, then test the total against affordability. | Survives. Every line traces to an objective, so every cut has a stated consequence and the budget can be defended line by line. |
Objective-and-task budgeting is slow and exposes whoever writes it. That is why it wins in a finance review: it is the only one where challenging the total means challenging a stated assumption.
Working Backwards From The Pipeline Number
Build the budget from the revenue plan, in this order, using your own historical rates rather than published benchmarks.
- Start with new revenue required, split by segment and product.
- Divide by average deal value to get new customers needed.
- Divide by win rate to get opportunities needed.
- Apply pipeline coverage, since not every opportunity closes in the period.
- Divide by each channel's opportunity conversion rate to get the qualified inquiries or meetings required.
- Multiply by the cost per inquiry that channel has actually delivered, not the cost you hope for.
- Add the fixed costs that make any of it possible: people, software, the website, content production.
Two adjustments make this honest rather than theatrical. Offset the plan by cycle length, because spend has to land early enough for the cycle to complete, which means a material share of this year's budget is buying next year's revenue and should be labelled that way. And put a range on every conversion rate rather than a point estimate, showing what the required budget becomes at the pessimistic end. A single-scenario budget is a forecast disguised as a plan.
Fixed, Variable, And Committed Before You Start
Split every line into fixed and variable before you allocate anything. Fixed costs continue whether or not you run a campaign: salaries, agency retainers, software licences, hosting, annual event contracts. Variable costs scale with activity: media, production, event marginal costs, contractor time.
The trap is a budget where four of every five available pounds or dollars is committed to headcount and software before a single campaign brief is written. That shape is an illustration, not a benchmark. Discretionary spend becomes the only place a cut can land, so a modest reduction in the total produces a severe reduction in market activity. The team also spends its time operating the tools it owns rather than deciding what the company should do.
The fix is structural. Set a policy floor for uncommitted spend and defend it in the same conversation where you defend headcount. Put multi-year licences and annual event contracts on a renewal calendar with a named owner and a decision date a quarter ahead of the deadline.
Brand And Activation: What The Evidence Actually Supports
The most-quoted numbers in this argument do not support the weight put on them. The brand-versus-activation split popularised by Les Binet and Peter Field rests on the IPA Databank, a substantial case-based body of work that is largely consumer-weighted. The B2B extension published with the LinkedIn B2B Institute is an extrapolation from a much smaller B2B sample rather than a separate B2B dataset. The commonly cited ratios should not be presented as a measured B2B optimum.
The same care applies to the claim that only a small fraction of business buyers are in market at any moment. That heuristic originates with Professor John Dawes of the Ehrenberg-Bass Institute, published via the LinkedIn B2B Institute in May 2021 and derived from interpurchase intervals. Dawes himself calls it a heuristic rather than a precise rule.
The structural argument survives, and it is enough. If most of your addressable market is not currently buying, spend aimed only at people showing intent addresses a small share of future revenue. Some spend has to reach buyers who are not in a cycle. The proportion is a judgment informed by your own purchase interval, deal value and market size. Make it, record the reasoning, and stop citing a ratio as though it were measured in your category.
Where The Money Actually Hides
Three categories reliably contain more money than the budget line suggests, and all three hide it the same way: the cost was approved once and never re-examined.
- Agency retainers. A retainer buys capacity, not output. Without a documented scope, a monthly record of deliverables against it, and an annual review, retainers drift into paying for account management on work that stopped being useful. Ask what would not happen next month if the retainer stopped. If the answer is a status call and a report, that is what you are buying.
- Marketing software nobody audits. Seat counts grow and never shrink. Run a licence inventory annually listing every tool, its owner, renewal date, seat count, genuinely active users, and the decision that depends on it.
- Trade shows. Stand space is the small part. Freight, drayage, stand build, staff travel, hospitality, hardware rental and post-show follow-up all belong in the true figure. Book them in one place, then compare that total against the pipeline the show produced within a stated window.
None of this needs a consultant. It needs one spreadsheet per category and the willingness to ask what a line item buys.
Talking To Finance In Finance's Language
A budget defence fails on vocabulary more often than on substance. The finance team is evaluating a capital allocation request against other requests, and it has a specific language for that.
Translate accordingly. Talk about payback period rather than campaign performance: how many months from spend to recovered acquisition cost. Talk about contribution and gross margin rather than revenue influenced. Talk about cash timing, because a budget that spends in the first quarter and returns in the fourth has a working-capital shape finance cares about.
Three behaviours earn more credibility than any deck. Bring the number you got wrong last year first, with what you learned and what you changed; nothing else buys as much trust. Present a range with the assumptions named, and say which assumption the total is most sensitive to. And bring a stated marginal case: what the next increment of spend is expected to produce, and what you would stop doing if it did not. Finance is comfortable with uncertainty that is quantified and hostile to certainty that is merely asserted.
What To Cut First, And What It Costs In Six Months
When a reduction is ordered the cuts happen in a predictable order, and the useful discussion is not what goes first but what each cut costs later.
- Sponsorships and hospitality with no measurable pipeline. Cut first. The six-month cost is goodwill in a specific relationship.
- Underperforming trade shows. Cut the ones you cannot connect to pipeline within a stated window. Six-month cost: incidental meetings that never appeared in any report.
- Broad-audience advertising. Usually next, and the expensive one. Six-month cost: unbranded inquiry volume falls, and paid search gets dearer because more of your demand has to be bought rather than arriving because someone already knew the name.
- Content production. Cheap to stop, slow to notice. Organic entrances decay gradually, so the damage is invisible in the quarter it is caused and undeniable a year later.
- Marketing software. Cut duplication rather than capability.
- Headcount. Last, because institutional knowledge does not come back.
Protect three things through any reduction: the website's technical health, measurement integrity, and the ability to answer an inbound inquiry quickly. Those are the cheapest lines to cut and the most expensive to have cut.
Zero-Based Budgeting, Done Seriously
Zero-based budgeting means every line starts at nothing and must be argued for from the objective it serves. Done as theatre, which is common, it produces last year's budget with new labels and a fortnight of wasted effort.
What separates the two is whether the exercise is permitted to reach an uncomfortable answer. Run it every two or three years, or immediately after a strategy change. Require every line, including headcount and every licence, to state the objective it serves, the mechanism by which it serves it, the evidence that the mechanism works in your category, and what would happen without it. Then rank the lines against each other rather than approving them individually, because ranking forces the trade-off into the open. Fund from the top until the money runs out, and look hard at the three lines immediately below it.
The exercise earns its cost through what it kills: the tool bought for a campaign that ended, the event attended because it was attended last year. Set the review date now and keep the ranked list. When the reduction comes, you already have the answer to what goes first and what it costs you.
Frequently Asked Questions
What percentage of revenue should a B2B company spend on marketing?
There is no defensible answer. Published shares blend companies with different cycle lengths, deal values and growth ambitions, so they describe other people's cost structures rather than what your plan requires.
Build the figure instead: work back from new revenue required through deal value, win rate, coverage and channel conversion rates to the demand you need, cost it at rates you have actually achieved, add fixed costs, and test the total against affordability.
How do I justify a marketing budget increase to my CFO?
Present the increment, not the total. State what the additional spend is expected to produce in pipeline and revenue, when the cash comes back, and what you will stop doing if the result is not there by a named date.
Open with a prediction you made last year that was wrong and what you changed as a result. A marketer who audits their own forecasts is far easier to fund.
Is it a mistake to cut brand advertising in a downturn?
It is the cut with the longest-delayed cost, which is why it is usually made first and regretted later. It barely touches this quarter's numbers, because most pipeline closing now was created before the cut. Six to twelve months out, unbranded inquiry volume falls and more demand has to be bought through paid search at auction prices.
Be careful with the evidence. The commonly quoted brand-to-activation ratios derive from largely consumer-weighted data and are not a measured B2B optimum.
How often should we audit marketing software spend?
Once a year as a formal exercise, plus a renewal calendar reviewed quarterly. The annual audit lists every tool with its owner, renewal date, contract term, seat count, genuinely active users, and the specific decision or workflow that depends on it.
The quarterly review exists because the expensive failure is not an unused tool, it is an auto-renewal nobody noticed. Set each decision date ninety days before renewal so cancellation and renegotiation both remain possible.