You have $4,000 in the bank, six months of runway, and a co-founder who keeps asking when you're going to "start doing marketing." That question has a nasty habit of arriving before you have any idea what marketing should cost — or which dollar does the most work.
The instinct is to open a spreadsheet, list every channel you've heard of, and split the money evenly. I did that with my first product. I set aside $1,200 for a "marketing budget," spent $700 of it on LinkedIn ads over five weeks, generated three leads, and closed none. The spreadsheet looked balanced. The business didn't move.
What actually works at the early stage is less about which channels you pick and more about sequencing: you spend the smallest possible amount to learn whether anyone wants the thing, then spend more only on evidence. A marketing budget for an early stage startup isn't a plan for growth yet. It's a plan for buying information cheaply.
Key takeaways
- Build the budget from your runway and your revenue target, not from a channel list. Channels come last.
- Budget in three horizons: money to learn, money to double down, money to scale.
- Keep a contingency of roughly 10–20% — early tests fail more often than they work.
- Allocation rules like 70/20/10 and 50/30/20 are useful as sanity checks, not as laws.
- Track cost per qualified conversation, not impressions. Vanity metrics eat runway silently.
How to build a marketing budget that survives contact with reality
Start with the number you cannot cross. If you have $180,000 in the bank and need 14 months of runway to hit a fundable milestone, your total monthly burn is roughly $12,800. Marketing should not push that over. Most founders I've worked with land marketing at 5–15% of total monthly spend when there's no revenue and 8–20% of revenue once money is coming in — though the second number swings wildly depending on whether you sell to businesses or consumers.
From there, the budget has to answer a single question: what do you need to learn this quarter?
Step one: write the learning goal
Not "get more customers." Something like: "Confirm that operations managers at mid-size logistics firms will book a demo after seeing a 90-second product video." That sentence tells you what to spend on and when to stop.
I keep this on one line at the top of the budget sheet. When a founder sends me a budget with "brand awareness" in row one, I know the next three months will be expensive and quiet.
Step two: set the total from runway
Work backwards. Take your monthly burn, add the marketing line, and check it against your runway in months. If marketing pushes runway below the point where you can raise or reach profitability, cut the marketing number — not the runway. This sounds obvious. I've watched two companies do the opposite and both had to cut their teams within a year.
Step three: allocate across three horizons
Split the total into learning, doubling down, and scaling:
- Learning (50–60% at pre-revenue): two or three cheap experiments across different channels. Small budgets, fast readouts.
- Doubling down (30–40%): more money into whatever showed a pulse.
- Scaling (10%): almost nothing at first. This line grows only when a channel has proven unit economics.
At pre-revenue, the learning slice should dominate. By the time you have paying customers and a repeatable acquisition path, the scaling slice takes over. Most startups get this backwards — they scale before they've learned, and the money disappears into a channel that never worked.
The allocation rules, and when each one earns its place
Founders keep asking me about the percentage rules they've read somewhere. They're handy shortcuts, but each assumes something about your stage. Apply the wrong one and you'll starve the part of the business that needs feeding.
What is the 70/20/10 rule for marketing budget?
The 70/20/10 rule splits your marketing spend into 70% on proven channels, 20% on promising new bets, and 10% on experimental ideas. It's a rule for companies that already know what works — the 70% protects the revenue engine, the 20% funds the next one, and the 10% buys lottery tickets.
For an early stage startup with no proven channel, 70% on "proven" is meaningless, because nothing is proven yet. I'd invert it: 70% on experiments, 20% on the strongest signal so far, 10% held back. Once a channel returns more than it costs for two consecutive months, flip the ratio.
What is the 3-3-3 rule in marketing?
The 3-3-3 rule is a discipline for testing three channels, with three variants each, over three months before deciding where to concentrate spend. The idea is to avoid declaring a channel dead after one bad week, and equally to avoid sinking six months into a channel that never had a chance.
Three months is the part people skip. A single month of paid search data is noise — you'll see swings of 40% in cost per lead that mean nothing. Give each test a full quarter, then judge.
What is the 50/30/20 rule in marketing?
The 50/30/20 rule allocates 50% to acquisition, 30% to retention and customer success, and 20% to brand and content. It assumes you already have customers worth keeping. At the earliest stage, retention is mostly a product problem, so this split is premature — but it becomes the right frame the moment churn starts eating your growth.
A rough signal: if more than a fifth of last month's customers have left, shift money toward retention before you spend another dollar on acquisition. Filling a leaking bucket is the most expensive way to grow.
What is a reasonable marketing budget for a startup?
Reasonable depends almost entirely on whether you have revenue. With no revenue, a working range is 5–15% of your total monthly spend, and often the bottom of that range is correct — you're buying learning, not growth. With revenue, many startups land between 8% and 20% of revenue, and the higher end is common when you're trying to establish a category.
Two things make the number swing more than stage does:
- Sales cycle length. Business-to-business deals with three-month cycles need budget that survives three months of no closed revenue. Consumer products with instant purchase need less patience.
- Whether the founder sells. If you're the one booking every demo, marketing's job is to fill your calendar — a much smaller budget than replacing a sales team.
One practical anchor: pick the number that lets you run two or three credible experiments per quarter without touching runway. If $2,000 a month can't fund a real test, either the test is too big or the budget is too small — but do not spend $2,000 on five things at once. That's not a budget, it's a scattergun.
A sample allocation you can copy and adjust
Numbers beat principles. Here's a shape I've used and recommended, scaled to a $3,000 monthly marketing budget at pre-product-market-fit stage. Adjust the channel names to your market; the proportions are the point.
| Line item | Monthly | Share | What it buys |
|---|---|---|---|
| Paid search / social test | $900 | 30% | Two small campaigns with distinct messaging |
| Content and landing pages | $750 | 25% | One comparison page, one case study, copy help |
| Outbound / community | $600 | 20% | Tools, event tickets, small sponsorships |
| Tools and analytics | $450 | 15% | CRM, tracking, email platform |
| Contingency | $300 | 10% | Whatever surprises you — it will |
Notice the contingency. I used to leave it out, reasoning that a tight budget should be fully allocated. That was wrong. Every quarter something unexpected needed money — a conference that turned out to matter, a competitor's launch, a tool that would have saved twelve hours. With no contingency, those opportunities came out of runway.
What counts as a marketing expense?
More than founders usually assume. Salaries for anyone doing marketing work, freelancers, ad spend, tools, events, content production, and the software that tracks it all — plus a share of the costs that only exist because of marketing, like the domain and the analytics stack. If you exclude salaries, you'll underestimate the true cost by a factor of two or three.
Tracking, and when to change the plan
A budget without a review cadence is a wish. I review marketing spend every two weeks against three numbers: cost per qualified conversation, cost per closed customer, and share of budget spent versus share of results delivered. If a channel takes 30% of the budget and produces 5% of conversations, it gets cut or rebuilt.
Two failures I've made personally and seen repeatedly:
- Judging too early. I killed a content channel after six weeks because traffic was flat. Six months later a competitor ranked for the same terms and I understood what I'd thrown away.
- Judging too late. The opposite mistake — I kept a paid channel alive for five months because each month looked slightly better than the last. Cumulative spend was $14,000 for eleven leads. The honest read was available in month two.
Set the review date and the kill criterion before you spend. "If cost per lead is above $60 at the end of the quarter, we stop." That sentence, written down in advance, is worth more than any allocation formula.
How often should you revisit the budget?
Every quarter for the allocation, every two weeks for the spend against results. Reallocating monthly is usually too slow; reallocating weekly is panic dressed as discipline.
The part most budgets get wrong
The real cost of an early marketing budget isn't the money. It's the attention. Every channel you fund demands someone to run it, read the results, and decide what's next — and at a five-person startup, that someone is usually a founder who should be talking to customers.
Which is why my advice, and I'll defend it stubbornly: fund fewer channels than feels comfortable. Two well-run experiments teach you more than six half-run ones, and they cost less than you think. The math isn't in the spreadsheet. It's in what you learn before the money runs out.