How much runway to spend on growth before Series A
The honest framing is not a percentage. It is how many experiments you can afford to lose before the round, and what the round needs you to have proven.
the short answer: the useful question is not what percentage of runway goes to growth. it is what your next round requires you to have proven, and what the cheapest experiment is that proves it. budget backwards from that, because a seed round is not funding growth, it is funding the evidence that growth is possible.
What the round actually needs
investors at series A are not buying your user count. they are buying a repeatable mechanism, and the difference matters enormously for how you should spend.
a chart going up that nobody can explain is worth far less than a smaller chart with a mechanism attached. “we spend a dollar here and reliably get this back within this period” is the sentence the round is for. two million users with no explanation of how they arrived is a harder pitch than two hundred thousand with a clear one.
so the budget question becomes: what is the cheapest thing that produces that sentence.
The arithmetic floor
there is a hard bottom, and it is not opinion.
about 5% of paid creatives become winners. the published production benchmark is roughly one new ad per $3,000 of monthly spend. so at $10,000 a month you produce three or four creatives a week and find a winner every month or two.
below roughly that, most consumer channels cannot produce a conclusion fast enough to be worth running. you can still grow, through community, content, and direct effort, but you cannot buy growth, and paying someone to manage a budget too small to learn from is buying an opinion.
A framing that works better than a percentage
instead of “what share of runway”, ask how many honest experiments can i afford to lose before the round.
one channel test done properly is about ten days and a defined budget with a written pass or fail line. if you have twelve months of runway and can afford to lose six of those tests, you have a real programme. if you can afford one, you are not running a programme, you are placing a bet, and you should place it on the thing you have the strongest prior about.
that framing also stops the most common failure, which is spreading a small budget across four channels simultaneously so that none of them gets enough spend to produce a verdict.
Sequence, not allocation
first, and cheapest: does anybody come back. if a cohort you did not pay for is not still active a month later, no amount of budget helps, and for subscription products the specific number is the share of first-month subscribers making a second payment. this costs nothing to measure.
second: the funnel step you have not looked at. almost always a cheap win and it costs a week rather than money.
third: one channel, tested properly. the full budget on one thing with a real gate, not a quarter of it on four.
fourth, and only after one channel works: scale, and add the second.
teams that invert this, starting with channel breadth, spend the most and learn the least.
What I would tell a seed-stage founder
do the first two before hiring anyone, including me. they cost time rather than money and they determine whether the rest is worth doing.
then, if you have $10,000 a month or more of genuine test budget and something people return to, the paid programme is worth starting and worth someone running properly.
if you have less than that, the honest answer is that your growth budget this quarter is your own hours in communities where your users are, and the money is better kept as runway. that is what i tell people on calls and it is why i turn work down. the four gates are in do not hire a growth marketer yet.
if you want the arithmetic run against your actual runway and round timing, let’s talk.
sources: adliftr analysis of 500,000+ meta ads and $1B+ spend · published 2026 meta creative production benchmarks · revenuecat AI app retention study for the renewal framing.