LOIs, pilots, revenue: what investors actually rank

The "we have five LOIs" moment

A founder is three slides into a seed pitch. The traction slide goes up: "Five signed LOIs from mid-market logistics companies, combined pipeline of $1.4M." The founder pauses for the nod.

The investor asks one question. "Who is paying you right now?"

The founder explains that the LOIs convert to paid contracts after the pilot, the pilots start next quarter, procurement is slow in logistics, and the intent is real. All of which may be true. None of which answers the question. The room cools by about ten degrees, and the founder spends the rest of the meeting climbing back uphill.

The mistake was not the LOIs. LOIs are fine. The mistake was treating an LOI as the headline when the investor was scanning for paid usage, and then having no language ready for the gap between the two. The founder ranked their evidence by what felt impressive instead of by what the investor was grading.

Why founders mis-rank their own traction

Most founders build the traction slide by volume. Whatever they have the most of goes biggest: a long waitlist, a stack of LOIs, a logo wall of pilots. The instinct is that more proof equals more convincing.

Investors read traction the opposite way. They are not counting artifacts. They are pricing down risk. Every dollar they put in is exposed to a few specific questions: Will anyone use this? Will anyone pay? Will they keep paying? Will they pay more? A piece of traction is only worth what it does to one of those questions. An artifact that does not retire a risk is decoration, no matter how many of them you have.

This is why five LOIs can land softer than one paying customer. The LOIs gesture at demand but leave the two hardest risks fully open: nobody has paid, and nobody has stayed. One paying, renewing customer kills both at once. Volume of weak signal does not substitute for a single strong one, because they are answering different questions.

The founders who do this well stop asking "what is my most impressive number" and start asking "which risk is this investor most worried about, and what is the strongest thing I have that speaks to it."

The traction evidence hierarchy

Rank your evidence by how much risk it removes, not by how it sounds. From strongest to weakest:

Two rules for reading the table. First, an investor mentally collapses your traction to the highest rank you can defend, then checks whether the volume there is real. Leading with rank 7 when you also have rank 3 is leaving your best card face down. Second, the discount column is where deals get re-priced. Investors do not just credit the evidence. They subtract for the standard failure mode of each type. Know your discount and address it before they raise it.

When weaker proof still matters

The hierarchy is not a rule that only revenue counts. Weaker evidence is valuable when it is the right weak evidence for your stage and when it points at a risk that is otherwise invisible.

At pre-seed, you may have nothing above rank 6, and that is expected. Here a strong free pilot with a named owner, a written scope, and a go-live date beats a vague claim of paid revenue you can barely document. The investor is not pricing your ARR. They are pricing whether real buyers engage and whether you can run a sale.

Weak-but-specific also beats strong-but-vague. An LOI with named terms, a signer with budget authority, and a stated condition to convert ("go live by Q3, then $40k/yr") carries more weight than "we have revenue" with no contract you will show. Specificity is itself a signal. It says you understand what a real commitment looks like.

And some weak evidence retires a risk nothing else can. A waitlist that grew 4x in a month from a single Show HN post is weak on willingness-to-pay but strong on distribution. If your investor's real worry is "can this team reach its market," that artifact speaks directly to it. Rank tells you the default order. The specific risk in the room tells you what to promote.

How to present each one honestly

The fastest way to lose a traction slide is to dress an artifact up as a higher rank than it is. Investors have seen the costume. Present each at its true rank and name the gap to the next one yourself.

LOIs: say "signed LOI," never "contract." State the conversion condition and the date. "Three signed LOIs, each converts to a paid annual contract on pilot completion, pilots start in August." You just turned a soft artifact into a dated plan.

Pilots: separate paid from free, and name the success criteria the customer agreed to. "Two paid pilots at $10k, success defined as cutting their reconciliation time by half, decision date in 60 days." A pilot with no exit criteria reads as a customer stalling.

Revenue: show whether it is recurring, one-off, or discounted, and how many logos it spans. "$8k MRR across four customers, all annual, no discounts" is a different company than "$96k in bookings" that turns out to be one pre-paid year from one logo.

Usage: distinguish triers from repeaters. "400 signups, 60 weekly actives, 18 have returned every week for two months." The third number is the only one that means anything, so lead with it.

The pattern is the same every time. State the artifact at its real rank, attach the specifics that make it credible, and pre-empt the standard discount. Honesty here is not modesty. It is what makes the strong items believable, because you have shown you know the difference.

Where this breaks: you cannot find the proof when asked

The reason founders inflate traction in the room is usually not dishonesty. It is that the real evidence is scattered. The LOI is a PDF in email. The pilot scope is in a Notion doc. The renewal is a line in Stripe. The usage number lives in a dashboard nobody screenshotted. When an investor pushes on a claim, the founder cannot produce the artifact in ten seconds, so they round up and hope.

This is the workflow problem RoundOS is built for. Each traction claim in your round can carry the source behind it: the signed LOI attached to the LOI line, the pilot agreement attached to the pilot, the Stripe export or the cohort chart attached to the revenue and retention claims. When an investor asks "who is paying now," you are not reconstructing the answer from memory. You open the claim, the document is attached to it, and the number is defensible because the proof is one click away.

That changes the meeting. A claim with its source attached reads as fact. A claim without one reads as a hope, and investors discount hope.