For the first handful of users, answering questions personally in a DM or a shared Slack channel is genuinely fine, and building a help center before you need one is wasted effort. The question is when that stops being true.
The signal is not revenue or headcount. It is whether you are answering the same question more than once.
Direct answer
A two-person startup needs a help center once you notice the same question coming up more than once from different users, even before you have significant revenue or any support headcount. Before that point, answering personally is faster than writing anything down. After it, a lightweight help center, even a handful of articles, starts saving founder time immediately, because every repeat question answered by an article is one that never reaches your inbox. It does not need to be large or polished to be worth having.
The actual signal to watch for
- A second user asks a question you already answered someone else, even in different words.
- You catch yourself copy-pasting a DM you wrote before.
- A user churns and mentions confusion about something that was never documented.
Any one of these is enough. Waiting for support volume to become a formal problem means you have already absorbed a lot of avoidable repeat questions.
It does not need to be big
A two-person team does not need a comprehensive knowledge base on day one. A handful of articles covering the questions you have actually been asked, kept current as the product changes, does more for a small team than a large, half-written one that goes stale within a month.
The part that actually matters at this stage
At two people, the product is likely changing fast, which means a help center written once will go stale quickly if nothing ties it to what you are shipping. The choice that matters most early is not how big to build it, it is whether it can stay current without becoming a second job for either of you.
Frequently asked questions
Is it too early for a help center pre-revenue?
Not if you already have external users asking repeat questions. Revenue is not the signal that matters; repeat questions from real users are.
How many articles should a two-person startup start with?
As few as the questions you have actually been asked. A small, accurate set beats a large, half-finished one that goes stale within weeks.
Who maintains it with only two people?
Whoever ships the change, ideally with help drafting so it does not become a separate writing task competing with building the product. See how founders can keep docs updated without hiring.
Conclusion
- Not needed on day one, but worth starting once the same question comes up twice.
- Small and accurate beats large and stale at this stage.
- Keeping it current matters more than how big it is.
- See getting started, or check pricing to start small.