A guide for founders · The budget
The startup marketing budget, by quarter.
A startup marketing budget is the money a software company commits to people, programs and tools in order to produce a specific amount of pipeline. The benchmarks put it near 8% of ARR for private B2B software, higher at smaller companies, but the useful number comes from the other direction: the pipeline the company needs, divided by the conversion rates it has, is what marketing must produce, and the budget is what that costs.
This page gives the benchmarks with their sources, a way to size the budget from pipeline instead, a split between people, programs and tools, and a quarter-by-quarter template for the first year. Ranges are shares of ARR or of budget, from named third parties. There are no dollar figures of mine on this page; what I charge is quoted on the first call.
By Neely Thomson·Published
The benchmarks
How much should a startup spend on marketing?
Three surveys answer it three ways, and the spread is the point. Private B2B SaaS companies put a median 8% of ARR into marketing, with selling counted separately. Large companies across all industries average 7.8% of revenue, and half of them run on 6% or less. A broader mix of company sizes shows a median of 5% and an average near 9%, with the smaller the company, the higher the share. A startup at a few million in ARR that spends well under its benchmark is rarely being frugal. It is usually not doing marketing, and the founder is doing it instead, unpaid and unmeasured.
By the numbers
- 8%
of ARR is the median marketing spend at private B2B SaaS companies, with selling costs counted separately at 15%. The two together are roughly a quarter of ARR.
SaaS Capital, Spending Benchmarks for Private B2B SaaS Companies 2026, more than 1,000 companies, March 2026
- 7.8%
of company revenue is the average marketing budget across all industries, flat for three years, and half of CMOs run on 6% or less. The sample skews to large companies.
Gartner 2026 CMO Spend Survey, 401 marketing leaders
- 5%
is the median across a broader mix of company sizes, with an average near 9%. Smaller companies consistently spend a higher share of revenue than larger ones.
The CMO Survey, Spring 2026, Duke Fuqua, Deloitte and the American Marketing Association
- 31.4%
of the marketing budget goes to paid media at large companies in 2026, with 24.5% to labor, 19.4% to technology and 19.2% to agencies. A startup's split looks nothing like this, and should not.
- 15.3%
of marketing budgets is now allocated to AI, while only 30% of marketing organizations say they are ready to scale it. The tool line in a startup budget should be smaller than the process line.
From the pipeline
How do you size the budget from pipeline instead of a percentage?
Five numbers, all in the CRM today. Next year's new revenue target. The average deal. The win rate from opportunity to close. The share of opportunities that should come from marketing rather than sales and referrals. And the cost per marketing-sourced opportunity, from whatever last year produced, or a conservative guess if nothing was measured.
Revenue target divided by average deal is the number of closed deals. Divided by win rate, it is the number of opportunities. Multiplied by marketing's share, it is the opportunities marketing owes. Multiplied by cost per opportunity, it is the program budget. Add the owner's salary and the tools, and that is the marketing budget. Now compare it with 8% of ARR. If it is double, either the conversion rates are the problem, which is cheaper to fix than the budget, or the revenue target assumes a marketing function the company does not have yet. Either way, the comparison told you something a percentage never would.
The split
How should the budget be split between people, programs and tools?
Large companies in 2026 put about a third of the marketing budget into paid media and roughly a fifth each into labor, technology and agencies. That is where a company ends up, not where it starts. At a few million in ARR, people are most of the budget, because one owner with contractors behind them is the whole function. Programs are the next line and should be bets with a date and a kill rule, not standing commitments. Tools are last, and smaller than they look: the tool line should be bought when the person who will run it starts, at the tier the programs justify, and the process around it should cost more thought than the licence.
AI changes the labor line more than the tool line. Drafting, routing, scoring and reporting used to need three people and now need one person and a process legal can sign off on. Budget for the person and the process. The subscriptions are cheap; the usage policy, the approval step and the review loop are where the money and the risk actually sit.
The template
What does the first year look like, quarter by quarter?
Not evenly. Spreading a year's budget across four equal quarters assumes you already know which channels work, and in the first year you do not. The shape below front-loads measurement, spends most in the quarter after the evidence arrives, and finishes with a forecast rather than a plan. The shares are a starting shape, mine, not a survey.
| Share | People | Programs | Tools | What it buys | |
|---|---|---|---|---|---|
| Quarter one | About 20% of the year | The owner. Contractors only where a program needs them. | Small. One or two bets, run as tests with a written kill rule. | The CRM, clean. Nothing else until the owner has chosen it. | A baseline. Five numbers from the CRM, one lead definition signed by sales, the scoreboard, and the first read on which channel produced the last twenty deals. |
| Quarter two | About 25% of the year | The owner, plus the first contractor or agency against a named program. | The first real channel bets, two or three, each with a kill and scale rule and a date to apply it. | What the bets need, bought at the tier the bets justify. | Evidence. Cost per opportunity by channel, for the first time, from attribution that was wired in quarter one. |
| Quarter three | About 30% of the year | Same, unless a channel has proven out and needs an owner of its own. | Scale what returned pipeline. Kill what did not, on the date set in quarter two, not later. | Flat. Consolidate what overlaps. | Pipeline. The quarter where most of the year's marketing-sourced opportunities should land. |
| Quarter four | About 25% of the year | The hiring plan for next year, costed from what the channels now need. | Run the proven channels. No new bets; the data would arrive too late to act on. | Renewals reviewed against use. Cancel what nobody opened. | Next year's budget, sized from real pipeline per dollar by channel, presented to the board with a forecast instead of a plan. |
The rules
What rules keep the budget honest?
Every program has a kill date
Written before the money is spent: what it must produce, by when, or it stops. The most expensive line in most startup budgets is the channel nobody decided to stop.
Reallocate quarterly, not annually
Pipeline per dollar by channel, refreshed each quarter, moves money toward what returns. An annual budget locked in January is a guess defended for twelve months.
Tools follow people
Nothing is bought for a hire who has not started. A stack chosen by someone who left is the first thing the next person replaces.
The board reads the same page
The budget and the five-metric scoreboard are one document. If the board sees spend without pipeline beside it, the budget conversation becomes a cost conversation.
Before the first hire
The budget is the third thing a first marketing hire should inherit.
Without it, every tool and program becomes a negotiation, and the hire learns within a month that asking costs more than it gets. The budget by quarter is deliverable three of The First Hire Plan, sized from your pipeline target rather than a benchmark. The full guide to the first hire is here, and the day-one checklist is here.
If you want it sized with you
Bring the five numbers. The budget takes an afternoon.
Revenue target, average deal, win rate, marketing's share, cost per opportunity. If you have those, the first call can size the budget before anything is signed. If you do not have the last two, that is what a Diagnostic finds in two weeks.
Questions
Questions people ask first.
What percentage of revenue should a startup spend on marketing?
The benchmarks cluster around 8% of ARR for private B2B SaaS, with selling costs counted separately, and the share runs higher the smaller the company. But a percentage is a check, not a plan. Start from the pipeline the company needs next year, work back through today's conversion rates, cost what marketing has to produce, and then compare that number with the benchmark. If the two disagree, the pipeline target or the conversion rates are wrong, and that is worth knowing before the budget is set.
How should a startup split its marketing budget between people, programs and tools?
At the earliest stage, people first: one owner, with contractors behind them, is most of the budget. Programs come second and should be bets with a kill rule, not a standing line. Tools come last and should be bought when the person who will run them starts, not before. The large-company split, roughly a third to paid media and a fifth each to labor, technology and agencies, is a reference point for where the company ends up, not where it starts.
Should a startup budget marketing by quarter or by year?
By quarter, inside an annual number. Quarter one is baseline and build, quarter two is the first channel bets, quarter three scales what worked and kills what did not, and quarter four sizes next year from real pipeline data. Spending a year's budget evenly across four quarters assumes you already know what works, and at this stage you do not.