Make the bet explicit.
Write your hypothesis, set your odds, and define what success looks like. Name the stakes and what could go wrong—before the outcome rewrites your thinking.
A hypothesis, not a hunch.
A decision journal for founders & their teams
Place better bets.
Keep the receipts.
Every feature, growth channel, and change of direction is a bet. Keep the thinking and the evidence together.
State your hypothesis and the odds upfront. Let your agents keep the evidence current. Review what happened—and carry the lesson into your next decision.
Free during beta. No credit card required.
A web app for you. A CLI for you and your agents.Bet 041 / ProductRunning
A shorter path to the first win.
Our hypothesis
Odds: 70%
Replacing the blank dashboard with a guided setup will lift trial activation to 40%.
Baseline
24%
This week
36%
Target
40%
Week 1Week 6
Latest update
Activation is climbing. Support requests are flat. Keep running for two more weeks before the call.
From “we think” to “we learned”
One bet. Its own metrics. The whole story.
01 /The expensive part is forgetting
You shipped it.
But what did you learn?
The plan is in a document. The numbers are in a dashboard. The reason you stopped is somewhere in Slack.
Three months later, the same idea comes back. Different name. Same assumptions. Another month of work.
Betlog gives each bet a home—so your next decision starts with evidence, not a better story about the last one.
02 /A small habit. A lasting record.
Write it down.
Keep score. Come back wiser.
Keep the evidence close.
Give every bet its own metrics and review cadence. You or your agents log the numbers, updates, tasks, and milestones. The original target stays in view.
A scoreboard, not a status meeting.
Leave a useful receipt.
Close with a postmortem that separates decision quality from the outcome. Keep the learning searchable, track your calibration, and put that knowledge to work on the next bet.
A lesson, not a lost document.
Fits the tools you already useFathom for metric syncLinear for linked tasks
03 / Your agents keep score. You make the call.
Less updating.
More deciding.
Give your agent a schedule and a workspace. It can log metrics, post updates, and maintain tasks and milestones—so your review starts with the evidence, not a hunt for the numbers.
Before it proposes another experiment, let it search what you’ve already learned. The CLI works for humans and agents: read and write the same record, from the tools you already use.
Your agentsYour scheduleYour workspace
betlog / terminalHumans + agents
# What are we running?
$ betlog bet:list --stage=running --json# Keep the scoreboard current
$ betlog entry:log 41 activation=36# Have we tried this before?
$ betlog learning:search pricing --jsonGet your workspace token and installation instructions in Settings → API after signing up.
Ready for Friday
“Update the numbers for our running bets. Flag what changed. Bring the relevant lessons from past experiments.” Set the routine with your own agent. Arrive at the review ready to continue, change course, or stop.
04 /Built for the calls founders actually make
Different bets.
The same discipline.
Product
Will a better first run turn more trials into customers?
Keep the reason you built it next to the evidence that it worked.
Activation · trial-to-paid · support load
Growth
Is this channel working—or are we just still spending?
Give every channel a target and a reason to keep going.
Qualified leads · CAC · conversion
Strategy
Should we go upmarket or double down on self-serve?
Revisit the assumptions behind your biggest commitments.
Deal size · sales cycle · retention
05 /Inspired by Thinking in Bets
A good decision
can still lose.
A lucky outcome doesn’t make the reasoning sound. A disappointing result doesn’t make the original bet foolish. Learning starts when you can tell the difference.
Betlog’s postmortems separate the quality of your thinking from the result. Your calibration record compares the odds you stated with resolved outcomes, so you can see where your confidence is earned—and where it needs work.
Read Annie Duke’s Thinking in Bets06 /One plan. Your whole team.
Keep the receipts.
Skip the seat math.
Unlimited bets and teammates. Predictions, postmortems, and calibration. CLI and API included. One flat price for your whole workspace.
See pricing & what’s includedYOUR WHOLE WORKSPACE
$390/year
Two months on us when you pay yearly.
or $39 month-to-month
Your next bet is already taking shape
Give it a record
you can learn from.
Start your first bet
Free during beta. Built by a founder, used to run real bets.