Most lead scoring criteria are a pile of points nobody can defend. Add ten for a job title, five for a company size, three for opening an email, and out pops a number that feels objective and means almost nothing. The problem isn’t scoring itself, it’s that the criteria got picked because they were easy to measure, not because they actually predict who buys. Good lead scoring criteria do the opposite. They start from your real customers and work backward to the handful of traits that genuinely separated the buyers from the tire-kickers, even when those traits are annoying to track.
For a solo founder, the point of scoring isn’t a fancy dashboard. It’s deciding who to spend your next hour on. So the criteria need to be few, honest, and tied to behavior you can actually see, not a wishlist of demographic boxes.
Lead scoring criteria that actually predict a sale
Split your criteria into two buckets, because they answer different questions. Fit criteria ask “is this the kind of customer I want,” and cover things like company type, size, and role. Intent criteria ask “are they showing signs of buying now,” and cover behavior: visiting the pricing page, replying to a message, asking about a specific feature. A lead scores high only when both are present. High fit with zero intent is a someday. High intent with low fit is a distraction. The ones worth your hour have both at once.
Of the two, weight intent more heavily than founders expect. A perfect-fit company doing nothing is colder than a slightly-off-fit company that just messaged you twice this week. Fit tells you who could buy. Intent tells you who’s leaning in right now, and right now is what you’re actually scoring for.
Build the criteria from customers you already have
Skip the generic templates. Look at the last ten people who bought and the last ten who ghosted, and ask what was different about them before they decided. Maybe every buyer had hit a specific team size. Maybe they all came from a particular channel, or all mentioned a particular frustration. Those real patterns are your criteria. Anything that doesn’t actually separate your buyers from your non-buyers is just noise dressed up as a number.
This is also why you should throw out criteria that never earn their keep. If “downloaded the ebook” scores points but ebook-downloaders never buy, that criterion is lying to you. Audit the scores against who actually converts and cut whatever doesn’t pull its weight.
Keep it simple enough to use
A scoring system with thirty criteria is a system you’ll never actually run by hand, and as a solo founder you’re running it by hand. Aim for five or six factors you can eyeball in seconds. Does this lead fit, are they showing intent, how strong is the signal, and roughly how urgent does it feel. If you can’t score a lead in the time it takes to read their reply, the system is too heavy for the team you have, which is one person.
Once you’ve got the criteria straight, the next question is how to turn them into a repeatable system. A look at the common lead scoring models covers a few simple frameworks you can run without a sales ops team or a spreadsheet that takes a PhD to maintain.
The best criterion is a signal you didn’t have to ask for
The sharpest scoring input isn’t anything a lead fills in on a form. It’s what they do in public before they ever talk to you: complaining about the exact problem, asking peers for a tool like yours, switching away from a competitor. Those unprompted signals are worth more than any self-reported field, because nobody games them. The trouble is they’re scattered across the internet, and watching for them by hand doesn’t scale past a handful of names.
That’s the part we built Unbound Compute to handle: catching the public signals that predict a buy and surfacing the people showing them, so your highest-scoring leads find their way to you instead of hiding in the noise. You decide who’s worth the call. We just make sure the strongest signals don’t slip past.