The short version
The most-quoted benchmark says companies spend about 8% of revenue on marketing. That figure comes mostly from companies earning over $1 billion. At $1M to $20M, you'll usually need a higher share.
A better method is to work backwards: from your revenue target, to the pipeline you need, to what that pipeline costs.
Every paid channel has a minimum spend below which it can't produce enough leads to judge. Fund that floor or don't run the channel.
Start with tests, then raise spend in steps while your cost per qualified lead holds. Going all in on day one and starting too low are both expensive mistakes.
Every founder asks this question eventually, usually right after a board member, an investor or a spreadsheet asks it first.
The internet's answer is a percentage of revenue. It's a reasonable place to start and a bad place to stop. Here's where the percentage comes from, why it misleads at your size, and how to build a budget you can actually defend.
The percentage-of-revenue rule, and where it comes from
The most widely cited figure is Gartner's annual survey of marketing leaders. In 2026, marketing budgets held effectively flat at 7.8% of company revenue, up only slightly from 7.7% in 2025.
That sounds like a clean answer: take your revenue, multiply by 8%, done. But look at who was surveyed. The 401 marketing leaders were mostly from companies reporting over $1 billion in annual revenue.
That matters, because a large share of marketing cost is fixed. You need a minimum spend to test any channel, a minimum level of tools, and someone to run it all. A $2 billion company spreads those costs over enormous revenue. A $3 million company can't.
Here's what 7.8% actually buys at your size:
Annual revenue | 7.8% of revenue | Per month |
|---|---|---|
$1M | $78,000 | $6,500 |
$5M | $390,000 | $32,500 |
$10M | $780,000 | $65,000 |
$20M | $1,560,000 | $130,000 |
At $1M, $6,500 a month has to cover tools, any outside help, and ad spend. That's enough to run one paid channel properly, not three. Use the percentage as a sanity check on whatever number you arrive at, not as the number itself.
A starting range by stage
Here are the ranges we use as a starting point with clients. They're our working figures, built from what it costs to run channels at a meaningful level, not a survey average. The share falls as revenue grows because the fixed costs spread further.
Stage (annual revenue) | Starting range | What it typically funds |
|---|---|---|
$1M to $3M | 10% to 15% | Founder-led content, one paid channel at a proper test budget, part-time or outsourced help |
$3M to $8M | 8% to 12% | Two paid channels, a content programme, a marketing lead or an agency |
$8M to $15M | 7% to 10% | Three or more channels, a small team or a full agency, proper tracking |
$15M to $20M+ | 6% to 9% | A team with specialists, brand investment, testing new channels |
Go higher if you're growing fast, entering a new market, or replacing referrals as your main source of clients. Go lower if most of your growth comes from existing clients expanding.
But the most reliable number doesn't come from a range at all. It comes from your pipeline.
Work back from your pipeline target
Instead of asking "what percentage should we spend?", ask "what does it cost to create the pipeline our revenue target needs?"
Here's the method, with an example:
New revenue target: $2M in new annual contracts next year.
Deals needed: at a $40,000 average deal, that's 50 new clients.
Qualified opportunities needed: at a 20% win rate, 250 opportunities. If you're unsure of your real win rate, our guide on how much pipeline you need to hit your number shows how to get an honest one.
Marketing's share: if referrals and outbound will produce 40% of those, marketing needs to create 150.
Cost per qualified opportunity: if yours is $1,500, that's $225,000 in programme spend.
Add the fixed costs: tools, people and agency fees on top.
The answer is a budget tied directly to a revenue goal, which is far easier to defend than a percentage borrowed from someone else.
The weak point is step 5. If you don't know your cost per qualified opportunity yet, you can't trust the final number. That's not a reason to guess. It's the reason to test first.
How to split the budget
Every marketing budget goes into four buckets:
People: salaries for in-house marketing staff, fully loaded with benefits and taxes.
Agency and outside help: agencies, freelancers, a fractional marketing leader.
Tools: CRM, analytics, SEO, design and automation software.
Media: ad spend, sponsorships and events.
There's no universal split, but three rules hold at every stage.
Fund the media before you fund the management. An agency fee spent managing $800 a month of ads is mostly wasted. Every channel needs enough spend to produce results worth managing.
Keep tools in proportion. Software subscriptions creep. If tools take more than about 15% of the budget at under $10M revenue, audit what's actually being used.
Count everything. A budget that leaves out salaries or agency fees makes marketing look cheaper than it is, and makes every return figure look better than it is. Our guide on how to calculate marketing ROI explains why full costs matter.
As an example, a $5M company spending $40,000 a month might put $18,000 into media, $12,000 into an agency, $7,000 toward a marketing hire's time, and $3,000 into tools.
How fast to ramp: test first, then scale in steps
This is where most budgets go wrong, in one of two opposite directions.
The first mistake is going all in. A founder approves a large budget, spreads it across four channels at once, and finds out three months later that two of them don't work. The money is gone and nobody learned which part failed.
The second mistake is starting too low. A founder approves a small budget, splits it across several channels, and none of them produces enough leads to judge. Each channel looks like a failure, gets cut one by one, and the founder concludes "paid marketing doesn't work for us." It never had a chance.
The way to avoid both is to work out each channel's minimum spend before you start.
Cost per lead = Cost per click ÷ Conversion rate
If clicks cost $35 and 4 to 5% of visitors become leads, each lead costs roughly $700 to $800. A single lead tells you nothing about whether a channel works. You need several, and you need to see how many become real opportunities.
That's why, in our experience, $1,000 to $2,000 a month is the practical floor for a paid channel like Google Ads or Meta, and more for expensive channels like LinkedIn. Below that, you're paying for data you can't read.
Then follow three steps:
Test. Run one or two channels at or above their floor for 60 to 90 days. Measure cost per lead and, more importantly, how many leads become qualified opportunities.
Scale in steps. Raise spend on the channels that work by around 20 to 25% at a time, and only while your cost per qualified opportunity holds steady.
Stop when it bends. When an increase no longer produces proportionally more pipeline, you've found the channel's limit. That's the moment to test a new channel, not to keep pushing money into this one.
// The Rule
Better to run one channel properly than three channels too thinly to learn anything.
Signs you're underspending or overspending
You're probably underspending if:
Your pipeline regularly falls short of what your revenue target needs.
Your channels never produce enough leads in a month to tell you whether they're working.
Your CAC payback period is very short, under about six months, and your sales team has spare capacity. You're likely leaving profitable growth on the table.
You're probably overspending if:
Each budget increase buys less pipeline than the one before. That's channel saturation.
CAC payback is stretching past 18 months. Our guide on what a healthy B2B CAC looks like covers where that line sits.
Leads are arriving faster than your sales team can follow them up.
Review the budget every quarter against these signs, not once a year. A budget set in January and left alone is guaranteed to be wrong by June.
Not sure what your budget should be? In a 30-minute Growth Diagnostic, we'll work back from your revenue target to a budget, a channel plan and the minimum spend each channel needs.
Book a 30-min Growth Diagnostic
Frequently asked questions
What percentage of revenue should a B2B company spend on marketing?
Gartner's 2026 survey puts the average at 7.8% of revenue, but that figure comes mostly from companies with over $1 billion in revenue. Smaller B2B companies between $1M and $20M usually need 6 to 15%, with the higher end for younger or faster-growing firms.
How much should a $5M B2B company spend on marketing?
Our working range for a company at that stage is 8 to 12% of revenue, or roughly $400,000 to $600,000 a year. The more reliable figure comes from working back from the pipeline needed to hit the company's revenue target.
Should a marketing budget include salaries?
Yes. In-house marketing salaries are part of the true cost of marketing and should be included. Sales salaries are usually kept separate and counted in customer acquisition cost instead.
How should a founder split budget between agency fees and ad spend?
Every channel should have enough ad spend to produce readable results before management fees are added. If agency fees are larger than the media they manage for more than a short test period, the channel is probably underfunded.
How quickly should a marketing budget increase?
In steps, typically 20 to 25% at a time, and only while the cost per qualified opportunity stays steady. When an increase stops producing proportionally more pipeline, the next dollar belongs in a new channel.