At least once a week, someone sends us a pitch deck and a prototype looking for investment.
The decks are created with AI and follow the typical format: problem, solution, market opportunity. Then they get noticeably weaker when talking about sales, marketing, or customers.
The product itself looks respectable, with nice functionality and basic design. Though that Michael Chen seems to use a lot of brand new software.
It really is amazing how far Claude Code and OpenAI images can take you.
But a working prototype and a nicely designed pitch deck aren't enough to get the attention of seed investors.
These days, you can build a product in a weekend. But you can't build a business that way.
If you're going to get the attention of stakeholders or investors, you still need to put in some work.
The easy parts don't tell us anything
Five years ago, a good deck meant you'd hired a designer or spent your weekends learning to do it yourself. A working product meant you could code, or you'd convinced someone who could to join you. Both were expensive, so both told us something.
They take a weekend now. Which means when we open your deck and click through your demo, we're not learning much from any of this:
- Your pitch deck
- The product landing page
- Product demo
- Brand
- Market research or data
- Testimonials
- Survey results
You should absolutely have all of this because showing up without it tells us something too.
Just don't expect these to carry the meeting and guarantee our interest. Almost everyone has this stuff, so it's not going to move us. If the deck and the demo are all you've got, we'll pass, and so will most of the funds you're about to email.
Five things that still count as evidence
Everything on that list is something you can produce by yourself, in a room, in about a week. What's left over is the part that requires another person to do something they didn't have to do.
That's the whole test. Did somebody outside your company take a real risk on you?
This isn't a Stackhouse list. When we're comparing notes with other seed funds or sitting on a co-investment call, some version of these five is what everybody is circling.
Somebody paid you
Not a pilot, not a discount that's really a giveaway, and not "they're using it and we'll figure out billing later." Before we invest money, we want to know if real customers are paying money.
Money is the only feedback that costs the other person something. Everything else is free.
"Compliments are the fool's gold of customer learning," writes Rob Fitzpatrick in The Mom Test. People tell you they love it because they like you, or because the meeting is almost over. Fitzpatrick's fix is to stop collecting opinions and start collecting commitments, and a commitment always costs one of three things: time, reputation, or money.
Watch what happens to the same company described two ways.
- "We have twelve organizations in the pilot program."
- "Four of the twelve are paying $180 a month. The other eight passed when we started charging."
The second version contains eight rejections and it's still the stronger claim. Somebody made a decision. The first version tells us twelve people accepted a free thing, which is not a decision.
If you're in the middle of a pilot right now, you don't have to end it. Put a price at the end of it and find out who's still there.
They paid again
The first payment says you can sell. The second says the product works. Those are two different claims and founders collapse them into one constantly.
Uri Levine is blunt about this in Fall in Love with the Problem, Not the Solution. "Retention is the only measure of product market fit," he writes. Not signups or week-one usage. Whether they're still there.
For monthly software, month four is usually where you find out. That's about when the person who championed you has seen the charge enough times to wonder about it.
If you sell annual contracts, your first real renewal might be twelve months away. Don't wait on it. Find the proxy:
- They added seats
- They rolled it out to a second location, campus, or department
- They asked about the higher tier without being prompted
- Somebody who wasn't in the original meeting is now using it every week
One renewal is worth more than three new logos, and almost nobody in a seed pitch can say a sentence like this one: "Nine of the eleven customers who signed in January were still paying in June."
Somebody told somebody else without being asked
You can buy attention. You can't manufacture a recommendation.
A referral you requested is a favor. Your customer likes you and wants to help, which is generous and isn't evidence. A referral that showed up on its own means somebody put their own credibility behind you in a room you weren't in.
So know where every customer came from. Specifically.
- An operations director mentioned you in a peer group thread
- A bookkeeper passed your name to two other clients
- Somebody found you through search and can tell you the exact phrase they typed
That last one counts. It means the problem is urgent enough that people go looking for a solution on their own. Levine makes the point that a real problem doesn't need a demo. If somebody is out searching, they've already decided they have the problem, which is the expensive half of any sale.
One unprompted referral beats fifty names you collected at a conference.
Somebody moved money to make room for you
New categories don't have budget lines. Somebody has to create one.
So a payment is only the start of the question. The second question is what they gave up in order to make it, and most founders never ask.
- Did they cancel something else to afford you?
- Did they pull from a discretionary fund, or did they get a line added?
- Did somebody else's request get denied so yours could be approved?
- Did they put you in next year's numbers and defend it in a meeting?
If nobody can answer, you may be living in discretionary spend. That's the first money to disappear in a bad quarter, and it explains a lot of companies with real revenue that come apart quickly.
A customer who fought for a budget line is not going to churn quietly. They'd have to explain it to the person they fought.
You can explain how each one happened
Founders usually underrate this one.
Five customers you can't account for is luck. Five you can walk us through, step by step, is a repeatable motion.
For every deal, write down five things:
- Where they came from
- What they pushed back on before they said yes
- How long it took from first conversation to payment
- Who signed, and who else had to approve it
- What it cost you in hours or dollars to get there
Levine describes a roadmap as a list of experiments you run until you find the one that works. Your sales process is the same thing, and you can't tell which experiment worked if you never wrote down what you tried.
This is also the part nobody prepares for. We'll ask how you got customer number three, and the answer is usually a shrug and a story about a conference. So will everybody else you pitch. The founders who can answer it precisely are the ones we keep talking to.
Four out of five puts you ahead of most of what comes across our desk.
The number might be lower than you think
The Stackhouse Fund looks for about $5,000 in monthly recurring revenue and five paying customers before we write a check. We invest in faith tech, so those customers are usually churches. The bar itself has nothing to do with who you sell to.
Founders hear that number and assume we're being polite. We're not. It's low on purpose.
At five customers we're not asking whether you've scaled. We're asking whether the thing works at all. Five buyers who found you, evaluated you, and paid you means there's a path. Fifty who came through your own network means you have a good network, which we already knew, because you got a meeting with us.
A bigger number from a warmer source is weaker evidence.
Other funds draw the line somewhere else. Some want to see twice the revenue, some don't use a revenue threshold at all, and a few will go earlier than we will on a founder they've been tracking for years. What we've never heard from any of them is a different list of what counts as proof. The threshold is a house rule. The five things aren't.
We weight founder quality at 35% of our scoring, more than any other single factor, and the evidence is a big part of how we read the founder. We're not impressed by $5,000 a month. We're interested in what it took to get it.
Slow buyers make better proof
If you sell to churches, schools, clinics, city departments, or nonprofits, you've complained about the sales cycle. Six to eighteen months. Most decisions clear a committee. Budgets get set once a year and defended by someone whose name is on them.
That's a real cost and we're not going to pretend otherwise. It works in your favor in exactly one place, and this is it.
A single customer in a market like that has already produced three separate things. Someone with authority said yes. That yes survived a group. And a budget line got created for a category that didn't have one before.
None of that got shorter because your demo is clean. It's the one part of this business that hasn't gotten cheaper, which is exactly why it still counts.
The reverse costs founders more time than anything else we see. A buyer who loves it in March and can't purchase until the next fiscal year hasn't told you much. Enthusiasm without a purchase order is the most common false positive in these markets.
If you're not there yet
You're probably closer than you think, and the fix usually isn't more building.
Charge before you feel ready. Most founders wait because the product isn't finished. You might be thinking, "We'll start charging once it's actually worth paying for." The price is how you find out whether it's worth paying for. A free pilot gets you a nice conversation and no information.
Sell five strangers before you sell fifty friends. Your network will say yes for reasons that have nothing to do with your product. That's generous of them. It's also useless to you.
Get to a renewal before you get to a pitch meeting. One renewal changes the conversation more than doubling your customer count.
Write down how each one happened. Where they came from, what they pushed back on, how long it took, who signed. Do it while you still remember it. That record is what we'll ask about, and almost nobody has it.
Where to start
Before you touch the deck again, write out every fact about your company you could still say out loud if the person listening assumed you were exaggerating by half.
Then cross off anything you produced by yourself. Cross off anything that took less than a week.
What's left is your evidence. If it's a short list, you know what the next ninety days are for.