Early startup evidence
What counts as startup traction?
Signups show that you earned attention. Traction begins when a defined group repeatedly reaches a valuable outcome and makes future growth more predictable.
A usable definition
Traction reduces uncertainty about the next customer.
A launch spike says your message travelled. A signup says someone was curious enough to begin. Neither tells you whether the product became useful. Stronger traction makes the next unit less mysterious: another user is more likely to activate, return, pay, or arrive through an existing user.
This is why there is no honest universal number for “good traction.” A tax product may matter once a year; a team chat product should matter daily. A marketplace must create a successful match on both sides. The interval and core action have to match the job.
The startup traction evidence ladder
Move from attention to evidence that compounds.
- 01
Attention
A visit, impression, waitlist entry, or signup. It proves the promise earned a look, not that the product delivered.
- 02
Activation
The user reaches the first valuable outcome. Measure the outcome, not account creation or onboarding completion.
- 03
Return
The same cohort performs the valuable action again at a natural interval, without being individually chased.
- 04
Dependence
The product enters a real workflow. Users import data, invite teammates, build history, or feel a meaningful loss when it disappears.
- 05
Payment
A real buyer accepts the price and buying friction. This tests value capture, but not retention or healthy unit economics by itself.
- 06
Referral
Users bring relevant users because the product worked, not only because an incentive paid them to share.
Choose evidence that fits the product
The metric changes with the business model.
Consumer or social
Watch: core-action activation, cohort return, contribution, and organic invites.
Do not confuse with: downloads, raw time spent, or notification-driven opens.
B2B software
Watch: repeated workflow completion, account retention, paid conversion, and expansion inside the team.
Do not confuse with: booked demos, unpaid pilots, or founder-operated usage.
Marketplace
Watch: match rate, time to match, repeat transactions, and contribution after incentives.
Do not confuse with: listings, registrations, or one subsidized side of the market.
Transactional product
Watch: conversion, repeat purchase at the natural interval, refunds, and contribution after acquisition cost.
Do not confuse with: traffic, gross sales, or discount-dependent orders.
Worked example
What do 300 free signups actually tell you?
The numbers300 people sign up, 120 reach the core action, 25 return once, two pay, and most return visits happen after a founder message.
The honest readingThe positioning can acquire attention and activation exists. Independent repeat use is still weak. Two payments are promising cases, not yet a repeatable sales motion.
The next testChoose the 25 returners, identify the job they share, remove personal reminders, and set one return threshold for the next cohort.
The decisionIf that narrow cohort returns independently, improve distribution to more people like them. If it does not, more traffic will mostly enlarge the leak.
A minimum evidence record
Write five fields before calling a metric traction.
- Cohort
- Which exact users started in the same period and source?
- Value action
- What did they complete that represents the product’s actual job?
- Interval
- When would a real need naturally bring them back?
- Threshold
- What result would change the next product or distribution decision?
- Intervention
- Would the behavior happen without reminders, discounts, or founder labor?
Read the combination
One weak metric does not always mean the same problem.
High acquisition, weak returnImprove the product, audience fit, or first value moment before buying more traffic.
Low acquisition, strong returnThe product may work for a narrow group. Clarify positioning and find more people like that cohort.
Strong use, weak paymentTest the buyer, urgency, packaging, and price. Usage alone does not prove a business.
Payment, weak useLook for procurement effects, service work, or a sale that did not become an adopted workflow.
Use benchmarks carefully
Compare like products, then compare your own cohorts.
Amplitude’s Product Benchmark Report covers more than 2,600 companies and reports materially different acquisition, activation, engagement, and retention ranges by industry and percentile. Sequoia’s product-health work makes the more durable point: a retention curve that stabilizes above zero shows a core group continues to find value. First establish the right action and interval, then compare cohorts using the same definition.
Frequently asked questions
Startup traction without vanity metrics
What counts as traction for a startup?
Repeated evidence that a defined customer reaches a valuable outcome and returns, pays, depends on the product, or refers others. The evidence must fit the product’s natural cycle.
Are signups considered startup traction?
Signups prove the promise or distribution earned attention. They become stronger traction when a meaningful share activates and later returns, pays, or refers others without repeated founder intervention.
How many users does a startup need to have traction?
There is no universal count. Ten recurring users in a narrow, valuable workflow can teach more than thousands of one-time visitors. Define the cohort, behavior, and interval.
What traction metrics matter before revenue?
Track the core action, time to first value, cohort retention, repeat frequency, and independent return. Commitments such as pilots, deposits, data migration, and team invitations can add stronger evidence.
Can a startup have traction without revenue?
Yes, but keep the claim narrow. Repeated valuable behavior can establish product traction. It does not yet establish pricing, unit economics, or a durable business.
Sources and scope
Evidence behind this guide
- Amplitude Product Benchmark Report: product data from more than 2,600 companies
- Sequoia Capital Data Science: measuring product health and retention curves
- Y Combinator Startup School: repetitive use and flattening retention curves
The evidence ladder and minimum evidence record are practical synthesis tools, not universal benchmark or investment-readiness tests. The founder snapshot is self-selected and descriptive.
Continue the decision
Use the evidence instead of staring at the total.
See where your instincts lead
Customer love or business durability?
Choose between real founder tradeoffs, then see how your answers compare.
Answer the founder questions