Fundraising is an information problem.

A founder raising a seed round will talk to somewhere between eighty and two hundred investors. Most of those conversations were never going to happen. The firm had already deployed the fund. The partner who cared about the category left in March. The thesis moved two years ago and the website never caught up.

None of that is secret. It is published — in filings, in fund announcements, in job changes, in the rounds a firm actually led last quarter. It is simply spread across a hundred places and nobody has the hours to read all of it every week, least of all someone trying to run a company at the same time.

So the work gets done badly, by hand, under time pressure, by the person who can least afford to do it. A target list gets assembled out of a friend's spreadsheet, an accelerator's PDF and whatever the internet coughed up on a Sunday night. It is out of date before the first email goes out, and there is no way to tell which parts.

Signal has a half-life

An investor saying they are looking hard at seed-stage climate hardware is worth a great deal in the week they say it. A month later it is worth something. A quarter later that interest has either become somebody else's term sheet or quietly gone away, and from the outside you cannot tell which.

An investor's stated thesis is marketing. Their last four cheques are evidence.

Every fundraising tool we have used treats investor data as a reference work — a directory you consult, accurate as of some unspecified point in the past. We think it is closer to a news feed with a decay function on it. The question was never who invests in this? It is who is leaning towards this right now, and how do you know?

What we believe

• A shortlist beats a database. Two thousand investor records is not an advantage, it is the same problem in a different file format. The useful output is twenty firms and the reason each one is on the list.
• Recency is the signal. What a firm says it invests in tells you less than what it wrote a cheque for in the last six months. We weight behaviour over positioning, every time.
• Sourced or it didn't happen. Every signal carries a date and a link to where it came from. A tool that tells you something confident and unattributable is worse than no tool, because you will act on it.
• A score is an argument, not a verdict. When we put a number next to a firm you can open it and see what the number is made of. Anything you cannot interrogate does not deserve four months of your life.
• The founder's time is the scarce resource. Not the investor's. We optimise for fewer, better conversations — not for the appearance of momentum.
• Your pipeline belongs to you. We are not a marketplace. Investors do not pay us to be surfaced, nobody sees your list but the people you share it with, and we will never tell a firm who has been looking at them.

What we are not

We are not an intro broker, and we do not rank you by who you already know. Software that sorts founders by the strength of their existing relationships compounds exactly the advantage that makes venture funding so unevenly distributed in the first place: the best-connected founder wins again, and the algorithm takes the credit.

So the ranking is built on the investor's behaviour, never on your address book. Once a firm has earned its place on the evidence, we will gladly tell you whether there is a warm path to it — a path is useful. A ranking assembled out of paths is a caste system.

We are also not another CRM you have to feed for a fortnight before it gives you anything back. You should get something worth having on the first day, out of work we did before you arrived.

We are a research team and a piece of software doing the reading you do not have time to do, and handing you the part that changed.

How to hold us to it

Manifestos are cheap. These are the parts you can actually check:

• Every signal in the product has a date and a source you can open. If one does not, that is a bug, and we want to hear about it.
• When we are wrong about a firm there is a way to say so, and a person reads it. Corrections go back into the catalog for everyone.
• We publish what we watch and how often. No black box doing the difficult part in a single sentence.
• If we ever take money from an investor to influence what you see, this page will say so. It will not quietly stop being true.

Why the fish

Yellowtail is worth the most on the first day. Left on the counter it becomes the same fish at forty per cent off, and everybody walking past can read the sticker. A round works the same way: one that has been open for nine months is not the round you launched, whatever the deck still says. The market prices the wait, and it prices it against you.

We named the company after the thing we are actually selling, which is speed while you are still worth full price.

Why invite-only, for now

Because the quality of the answer depends on how well we understand the question, and right now that means working closely with a small number of companies rather than opening the doors and averaging everyone. That will change. Until it does, we would rather be useful to a few people than adequate to everyone.

If you are raising and this sounds like the tool you wanted to exist, ask for an invitation. Tell us what you are raising and we will tell you honestly whether we can help yet.