The Founder's Habit — Ten Sessions a Week
Watching ten real user sessions every week is one of the highest-value habits an early-stage founder can build. Learn which sessions to choose, what to look for, how to keep notes, and how to turn what you see into decisions.
Short answer: set aside forty minutes each week to watch ten sessions: five from new users who did not activate, three from users who did, and two from your most engaged customers. Write one line about each and choose one thing to change. At small scale this teaches you more than any dashboard, because every session is a real person meeting your product.
Table of Contents
- Why This Habit Matters Early
- Which Ten
- What to Look For
- The One-Line Log
- From Watching to Deciding
- Traps
- Key Takeaways
Why This Habit Matters Early
- Numbers are too small to trust. With forty signups a week, a conversion rate swinging from five to eight percent is one or two people.
- Users do not tell you. Most people who get stuck leave without a word.
- You are too close. You know where every button is. A new user does not, and you cannot see that from inside your own head.
- It keeps the team honest. A recording of someone failing to complete signup ends debates quickly.
- It is cheap. Forty minutes, no recruitment, no scheduling.
Interviews remain essential, because behavior does not explain motive. Watching sessions makes interviews sharper: you arrive knowing what to ask.
Which Ten
| Count | Who | Why |
|---|---|---|
| 5 | New users who signed up and did not reach activation | Where your funnel leaks |
| 3 | New users who activated in their first session | What success looks like |
| 2 | Your most active customers | What the product is really used for |
To filter this way you need two things in place:
deepsync("identify", user.id, { plan: "trial", signupDate: user.createdAt });
deepsync("track", "activated");Identifying users lets you find a specific person's sessions, and an activation event lets you separate the two groups (defining goals).
If you have no signups yet, watch ten visitors to your landing page instead: five who left and five who clicked your call to action.
What to Look For
In sessions that did not activate
- The first moment of hesitation
- Something clicked that does nothing
- A form field retried
- An empty screen with no guidance
- The last thing they did before leaving
In sessions that did activate
- The path they took. Was it the one you designed?
- How long it took to reach value
- What they skipped
In your best customers' sessions
- Which features they go straight to
- Features they never touch
- Workarounds: exporting, copying, keeping two tabs open
The gap between the first two groups is usually your most valuable finding. Often the users who succeed did one specific thing early that the others never found.
The One-Line Log
Keep a simple running document. One line per session.
| Date | Who | What happened | Note |
|---|---|---|---|
| 2 Oct | New, not activated | Created account, stared at empty dashboard, left after 40 seconds | Empty state gives no next step |
| 2 Oct | New, activated | Pasted snippet within two minutes, came from docs | Docs-first users activate |
| 2 Oct | Power user | Uses filters constantly, never opens reports tab | Reports tab unused? |
After a month you have forty observations. Patterns that were invisible in one week become obvious. This log is also the best onboarding document you can give a new hire.
From Watching to Deciding
- End each session with one decision. The smallest change that addresses the most common thing you saw.
- Ship it that week.
- Look for it next week. Is the friction gone in new sessions?
- Bring people in. Once a month, watch five sessions with the whole team.
- Follow up directly. If you can see who struggled, and your terms allow it, a short personal email offering help often turns a lost user into a conversation.
Fix bugs immediately. Do not put a broken button in a backlog.
Put it in the calendar
The habit dies the first busy week unless it has a fixed slot. Same time, same day, forty minutes.
Traps
- Reacting to one vivid session. Wait until you have seen something three times, unless it is a plain bug.
- Watching only failures. It is discouraging and it hides what works.
- Delegating it entirely. Summaries from someone else do not change your instincts the way watching does.
- Scaling it up. Once you have thousands of sessions, use filters and AI summaries to choose your ten. Do not try to watch more.
- Forgetting consent and privacy. Record with consent, keep inputs masked, and use what you see to improve the product.
Key Takeaways
- Ten sessions a week: five not activated, three activated, two power users.
- Compare the groups to find what successful users did differently.
- Keep a one-line log and make one change each week.
- Give the habit a fixed slot.
Conclusion
Most founders can describe their product in detail and their users' first five minutes not at all. Ten sessions a week closes that gap, and it costs less time than a single meeting.
Frequently Asked Questions
Related articles
Stay in the loop
Get the latest insights on product analytics and user behavior delivered to your inbox.



