TractionJune 18, 2026·8 min read

What Actually Counts as Traction at Seed

Traction is not a chart that goes up. It is evidence the chart will keep going up after you stop pushing it. Most of what founders call traction is just the sound of their own effort.

The Roast My Startup firm

Investor-style feedback, written down

Traction is the most overused word in a pitch and the least examined. Founders say “we have traction” and point at a chart that goes up and to the right. But a chart going up is not traction. It is a chart. Traction is the narrower, harder claim that the line will keep climbing after you stop pushing it, and almost everything founders put on the traction slide is evidence of the opposite: proof of how hard they are currently pushing.

The distinction that matters is between effort-driven growth and pull-driven growth. You can move nearly any metric for a while with brute force: founder-led sales, a paid budget, a launch, a personal network. None of it tells an investor what happens when the founder stops being the engine. Pull is when the metric grows for reasons that are not you. That is the only kind of growth worth calling traction.

The two kinds of growth, and only one is traction

Effort-driven growth is real work and worth doing, but it is linear and it stops when you stop. Pull-driven growth has a flywheel: users bring users, a channel pays for itself and scales, a product becomes part of someone's routine. An investor's whole job is to tell these apart, because they are buying the second derivative, not the first. A smaller number that grows on its own is worth more than a bigger number you are carrying on your back.

Signal strength by metric type

Retention curve that flattens100

Paid growth at a payback that works82

Organic word of mouth76

Revenue from a repeatable channel64

Total signups, no retention shown22

Press and launch-day spikes12

How much a given metric actually tells an investor about durable demand. Notice that the loudest numbers, signups and press, sit at the bottom, and the quiet one, retention, sits at the top.

The traction ladder

It helps to rank what you actually have rather than lumping it under one word. The higher a signal sits on this ladder, the harder it is to fake with effort and the more an investor will pay for it.

From vanity to durable

Durable, hard to fake

Retention curve that flattens

people stay without you pushing

Repeatable paid growth at a payback that works

you put in $1, you get more than $1 back

Organic word of mouth

users arriving because other users sent them

Revenue you can attribute to a channel

growth with a known, repeatable source

Raw signups or downloads

a number that says nothing about staying

Press, launch spikes, waitlists

attention you paid for in effort

Vanity, easy to manufacture

Lead with the top of the ladder if you have it. If everything you can show sits near the bottom, you do not have a presentation problem, you have a traction problem, and it is better to know that before an investor tells you.

Why retention is the only honest traction

Retention is the one number you cannot fake with effort. You can buy signups, manufacture a launch spike, and will a few logos into a deck through sheer hustle. You cannot make people stay. If a cohort keeps using and keeps paying month after month, something about the product is genuinely working, and that is the closest thing to proof a seed-stage company can offer. It is also why a forecast without a cohort table is so suspect, a point we made in the numbers that kill a seed round.

Growth tells you that you can fill a bucket. Retention tells you the bucket has a bottom.

Traction that proves the wrong thing

A metric can be completely real and still be the wrong proof, which is a subtler trap than vanity. Demand for a free tool is not demand for a paid one. A pile of pilots is not a pile of contracts. Enthusiasm from your existing network is not evidence that strangers will care. When the traction you show proves demand for something easier or cheaper than the thing you are actually selling, an investor notices, and it counts against you rather than for you. This is one of the quieter reasons investors pass.

Durable signals vs. things that look like traction

A cohort that is still active and paying months after signup
Customers arriving from other customers, with no spend behind it
A paid channel that returns more than it costs, repeatably
A big signup number with no retention curve attached
A launch-day spike presented as a trend
Strong usage of a free tier with no evidence anyone will pay

The top items survive your absence. The bottom ones evaporate the moment you stop pushing, or prove demand for a product you are not selling.

How to present traction honestly

Lead with retention and cohorts. It is the hardest number to fake, so showing it first signals confidence and earns trust for everything after.
Separate organic from paid everywhere. Blending them hides whether the growth is pull or purchase.
Show the metric that would embarrass you if you hid it. The thing you are tempted to leave out is the thing an investor most wants to see.
Name what each number proves, and what it does not. A founder who says “this shows demand for the free tier, not yet for paid” is more credible than one who pretends the distinction does not exist.

Honest traction is more persuasive than impressive traction, because investors have seen every way a chart can lie and they are scanning for the founder who is not trying. Before you put your numbers in front of one, it is worth pressure-testing the deck they sit in. If you want a firm of analysts to tell you which of your metrics is real traction and which is just effort in disguise, that is what Roast My Startup does.

tractionmetricsretentionfundraising

Find the holes before an investor does

Roast My Startup is a firm of AI analysts that tears apart your deck, model, forecast, and data room, then tells you exactly what an investor would use to pass. Brutal first, constructive second.

Roast my startup See how it works

Keep reading

ForecastForecast Only What You Have Proven (Investors Delete the Rest)When you raise a seed or Series A, investors mentally erase every unproven line in your forecast and judge you on what is left. Build the model on proven revenue, because speculative revenue reads as risk, not ambition.Read OutreachHow to Reach Investors, and Why the Warm Intro Is OverratedThe warm intro raises your odds, but it is a weight on a scale, not a gate. A specific cold email from a founder who has done the work beats a lukewarm referral. Here is how to write one that gets answered.Read Round MechanicsHow Much to Raise, and What It Really CostsThe right raise is not the biggest one you can get. It is the smallest one that buys you to a milestone worth a markup. A look at dilution, runway, and how to size a round without giving away the company.Read ProcessRun Your Raise Like a Process, Not a LotteryA fundraise is a sales process with you as the product. Founders who run it sequentially and hope lose to founders who run it in parallel, on a clock, with a real pipeline. Here is how to run it.Read InvestorsChoosing Investors: The Cap Table Is ForeverYou will spend years with your lead investor and you cannot fire them. A guide to choosing investors for the relationship after the wire, not the size of the check or the logo on the term sheet.Read ThesisNon-Consensus and Right: The Only Way to Win a Funding RoundGreat startups are built on a contrarian truth that turns out to be correct. Here is why being non-consensus and right is the only quadrant that pays, and how to tell a real secret from a bad idea.Read ThesisThe Case for Being Roasted Before You PitchPraise is the most expensive thing a founder can buy. The market only tells you the truth after it has already decided to pass. The fix is to manufacture an honest adversary on purpose.Read SurvivalDefault Alive: The Only Forecast Number That Survives a DownturnDefault alive is the single question that decides whether your startup controls its own fate: would you reach profitability on your current growth and burn before the money runs out, if you never raised again?Read Data RoomWhat a Data Room Actually SignalsA startup data room is not storage, it is a credibility test you take before anyone opens a file. Here is what its structure quietly tells investors, and a minimal room that signals you are ready.Read TimingWhy Now: The Question That Separates a Company From a FeatureThe 'why now' question decides whether you have a company or a feature. Here is how investors read startup timing, and how to prove the window is open instead of asserting it.Read ForecastThe Numbers That Kill a Seed Round: Reading a Forecast Like a VCFounders model hope, investors model decay. These five numbers in a startup financial forecast quietly decide whether the rest of the model gets read.Read Business ModelThe Moat Question: What Stops Anyone Else From Doing ThisEvery investor eventually asks what stops a competitor from copying you. Here is how to think about a startup moat, why defensibility compounds, and the difference between a real moat and a head start.Read Pitch DeckHow to Pressure-Test Your Pitch Deck Before You RaiseYour pitch deck is not a document, it is an argument. Here is a repeatable way to red-team your deck and find the holes before an investor uses them to pass.Read TeamFounder-Market Fit: Why ‘Why You’ Is the Hardest Question to FakeInvestors ask ‘why you’ to test founder-market fit. Here is what they are actually scoring, why resume logos barely count, and how to show the earned insight that makes you the right team for this market.Read FundraisingWhy Investors Pass, and Why It Is Almost Never the IdeaA pass is a verdict on your thinking, not your market. Here are the real reasons VCs say no, and why most of them are smaller and more fixable than founders expect.Read