If your growth only looks good monthly, it doesn't look good: what growth rate seed investors expect
If your growth only looks good monthly, it doesn't look good.
On May 15 I spent a call with a founder in London who'd just pivoted. He'd spent a year and a half building a multi-agent AI video editor, then learned from users what a lot of AI founders are learning: experienced editors were faster doing it by hand, they wanted fine control, and the AI couldn't make sense of vague creative requests. So he turned the company into a service, using his internal tools to make videos for clients. Smart pivot. He wanted to know how to raise.
I gave him the same order I give every founder: market size, then traction, then defensibility. Market size carries most of the weight, I'd say 60 to 70 percent of what an investor is weighing, and video is a big market with a lot of competition. Defensibility is the question every AI founder gets: why can't the big platforms build this next quarter? But the part he needed to hear, and the part most founders skip, was the number in the middle. I told him I want to see "three to four, maybe five percent a week growth minimum."
A week. Not a month.
Where three to five percent a week comes from
The weekly lens isn't my invention. Paul Graham's essay "Startup = Growth" says a good growth rate during Y Combinator is 5 to 7 percent a week, that 10 percent a week is exceptional, and that if you can only manage 1 percent, it's a sign you haven't yet figured out what you're doing.
My floor is a little lower, three to five percent, because I'm talking about the minimum to start a credible seed conversation, not the bar for a top accelerator demo day. But it's the same idea. Growth gets measured per week at the early stage, and the bar is high.
Compounding is why. Graham's own table: 1 percent a week is about 1.7x in a year. 5 percent a week is about 12.6x. Those two companies might both say "we're growing" in a pitch. One of them is a business that might matter. The other is a business that exists.
Why investors think in weeks, not months
Monthly numbers hide things. A company that adds roughly the same number of customers every month can show a chart that goes up and to the right forever. Graham calls that out directly: if you're adding a constant number of new customers each month, your growth rate is falling, and you're in trouble.
Weekly numbers expose that fast. Four or five data points a month instead of one. Momentum shows up, or it doesn't. Seed investors are betting on a curve that bends, and the only way to see whether it's bending is to look closely and often.
There's a practical reason too. Weekly growth forces weekly focus. A founder who reports weekly knows exactly what they did this week to move the number. A founder who reports monthly can lose three weeks to things that don't matter and still produce a respectable slide.
What counts as growth after a pivot
The London founder had a specific problem. His old metrics were about a product he'd abandoned. His new business was services. Which numbers count?
The answer: the clock resets. Investors won't credit you for the users of a product you no longer sell. What counts now is revenue from the new model, measured weekly from the day the new model started. Graham makes the same point: revenue is the best metric, and active users are the next best for companies not yet charging.
That's hard to hear after eighteen months of building. But it's also freeing. A services business with paying clients can show weekly revenue growth quickly. The old product's usage numbers were never going to impress anyone anyway.
How to show weekly growth honestly in a deck
He didn't have an updated deck. The old one told the story of the product he'd walked away from. This is how I'd rebuild the traction slide:
- One chart. Weekly revenue since the pivot, bars or a line, nothing else.
- One number in big type: the average weekly growth rate over the last eight to twelve weeks.
- A short note on what's driving it: new clients, repeat clients, larger projects.
- No cherry-picked weeks. Show the dips. Investors trust a jagged real line more than a smooth suspicious one.
If the average is below three percent a week, don't hide it with a monthly view. Fix the growth, then come back to the slide.
When to go back out and raise
We agreed on the obvious next step on that call: focus on revenue and traction, and reconnect when the metrics improve.
That's not a brush-off. Raising too early with a weak curve is worse than waiting. Every investor who sees you now remembers the flat line. Come back eight weeks later with three to five percent a week compounding, and you're a different company to the same investor.
In the meantime, there are useful things to do besides pitch. Build relationships with investors before you need them, so that when the numbers are ready, you're calling people who already know you. We invited him to the Sunny Ray Show and to pitch on Spark Tank once he had a deck and traction to show. Neither is a raise. Both are ways to meet the people you'll want in your round.
Then hit the number, week after week, and go.
When your weekly numbers are moving and you're ready to raise, get in front of investors.
Frequently asked questions
What is a good weekly growth rate for a startup?
Paul Graham's essay puts a good rate during Y Combinator at 5 to 7 percent a week, with 10 percent exceptional. My minimum before a seed conversation is three to five percent a week.
What traction do I need for a seed round?
Weekly growth in revenue, or active users if you're not charging yet, sustained over a couple of months. A large market and a reason the big players can't copy you come with it.
How do I present growth after a pivot?
Start the clock from the day the new model began and show weekly revenue since then, dips included. Investors won't credit metrics from a product you've abandoned.