Fellowship · Sprint lesson
Turn your startup into anoptimization problem
Choose one metric. Run four sprints. Review results and decide what to do next.
Turn your startup into an optimization problem
“Startup = Growth,” by Paul Graham
“We usually advise startups to pick a growth rate they think they can hit, and then just try to hit it every week. The key word here is “just.” If they decide to grow at 7% a week and they hit that number, they’re successful for that week. There’s nothing more they need to do. But if they don’t hit it, they’ve failed in the only thing that mattered… … “Focusing on hitting a growth rate reduces the otherwise bewilderingly multifarious problem of starting a startup to a single problem. You can use that target growth rate to make all your decisions for you; anything that gets you the growth you need is ipso facto right. Should you spend two days at a conference? Should you hire another programmer? Should you spend time courting some big customer? Should you add x feature? Whatever gets you your target growth rate.”
YCombinator
02
Part 1 · Concepts · Why
A clear target helps you choose where to focus
You have more possible tasks than time. A clear target helps you choose the work most likely to improve your business.
- Define the metric. One business outcome, suited to your stage.
- Establish the baseline. Its measured value today.
- Choose the November 20 target. The value you want to reach by the end of the fellowship.
Before choosing a task, ask: how will this help us reach our metric target?
Choose your metric that matters (MTM).
Your MTM measures the business outcome you want to improve during the fellowship.
Choose it once, in setup week. Use your MTM to prioritize work throughout the fellowship.
03
Part 1 · Concepts · What
Use the problem-solving loop to choose, act, and learn
Desired future state
Waypoints
Hustle and rigor
Accounting
Response
What is my current reality?
How do I start?
How do I start?
What’s critical to get right?
Where can we take risks?
What are the requests and promises?
Where can we take risks?
What are the requests and promises?
Did it get done?
What are the facts?
What are the facts?
Why did it happen?
What did we learn?
What did we learn?
What do we change?
A waypoint is an intermediate result
or a question to answer on the way
to your goal.
or a question to answer on the way
to your goal.
04
Part 1 · Concepts · What
A sprint is two weeks of focused work toward clear goals
01
Define
Set measurable or observable goals that advance drivers of your MTM.
02
Execute
Schedule time for the work. Use new evidence to adjust your steps during the sprint.
03
Review
Review steps and results on the scheduled date, even if work is unfinished.
The loop helps you decide what to do. The sprint sets the goal, ideas, steps, and review date.
Sprint goals focus on drivers or inputs to your MTM: leading indicators of progress toward the number.
Each goal needs a specific number or observable outcome, so you can answer: did it happen?
05
Part 1 · Concepts · Prioritization
Prioritize sprint ideas with ICE
Score each idea from 1–3 on three criteria.
| Impact | If successful, how much will this move the MTM? | 1 = Low · 2 = Medium · 3 = High |
| Confidence | How confident are we that it will produce that impact? | 1 = Low · 2 = Medium · 3 = High |
| Ease | How easy is it to deliver, given the time and resources required? | 1 = Hard · 2 = Moderate · 3 = Easy |
ICE score = Impact + Confidence + Ease
Total: 3–9. Higher scores get priority.
Base confidence on evidence: previous results, customer conversations, or a small test, not enthusiasm.
Compare ideas, discuss the assumptions behind the scores, and choose the highest-scoring idea that fits your sprint capacity.
06
Part 1 · Concepts · How, part one
Plan the goal, ideas, steps, and work to postpone
Why this sprint?We think [problem] is holding us back.Connect the problem to the MTM. Name what you are assuming.
GoalProgress on a driver or input to achieving your MTM.Set a specific number or observable outcome by the review date. Be able to answer: did it happen?
IdeasThe initiatives we will try to reach the goal.Use ICE to choose ideas that fit the sprint.
StepsBreak each idea into steps. Assign an owner, due date, and time estimate.Allow time between completing a step and measuring its result.
Evidence & reviewThe results or evidence to examine, and the review date.Agree on the evidence and date before starting work.
Work to postponeThe work you will postpone during this sprint.Name what you will postpone to make time for this goal.
Separate the goal, idea, and steps. Goal: contact 40 qualified prospects by sprint end. Idea: test targeted cold outreach. Steps: build a list, write and send messages, and track replies and bookings.
You will see this structure twice more: filled in for Belong Law, then blank for your own.
07
Part 1 · Concepts · Capacity planning
Plan work that fits your next two weeks
01 · Available time
Count the hours you have
Subtract delivery work, meetings, other commitments, and time off.
02 · Work estimate
Estimate each step
Break each idea into steps. Use a range when you are uncertain.
03 · Buffer
Leave room for surprises
Reserve time for unexpected work. Adjust the buffer as you learn.
04 · Scope
Make the work fit
If estimates exceed capacity, reduce scope or postpone steps.
Check each person’s capacity.
One founder’s spare hours cannot cover work only the other founder can do.
Effort is different from elapsed time.
Writing a quote may take 30 minutes. Waiting for a signature may take seven days.
08
Part 1 · Concepts · How, part two
Do the work, review the results, and choose the next step
Kick offAgree on the plan. Start steps that need time to produce results.Day 1
Midpoint check-inReport progress, obstacles, and help needed. Adjust steps using new evidence.Around day 5
Work continuesIf you add a new step, name which planned step you will postpone.Days 6 to 9
ReviewReview steps and results, including unfinished work.Day 10
Accounting · establish the facts
Response · decide what changes
01Did we do the work?Did we complete the planned steps? What was left unfinished?
02Did it work?Did the leading indicator improve, or did we answer the waypoint question? What changed in the MTM?
03What did we learn?Which assumption did the evidence support or challenge? What is still unanswered?
04What will we do next?Use what you learned to choose the next sprint goal, ideas, and steps.
09
Sprint 1 Illustration
Part 2 · Achint’s goal
Grow from 2 to a sustained 10 ICP meetings per week
Achint · Founder, Belong Law
MTM: completed meetings with petitioners who fit the ideal customer profile (ICP).
2 / weekCurrent baseline
→10 / weekSustained rate by November 20
Google Ads + cold outreachReach more prospects
→
Meetings bookedTrack upcoming demand
→
ICP meetings completedCount toward the weekly MTM
Assumed for this illustration: prospect volume is the constraint; prospect quality is adequate.
Count completed meetings by week. A single week at 10 is progress, not proof the rate can be sustained.
11
Part 2 · Define
Achint’s sprint plan: more volume from two channels
Why this sprint?We think too few ICP petitioners enter the funnel.
Goals1. Contact 300 distinct ICP prospects.
2. Launch 3 new AdWords tests.Both by October 2. Check each goal separately: did it happen?
2. Launch 3 new AdWords tests.Both by October 2. Check each goal separately: did it happen?
IdeasTest L-1 outreach to overseas founders planning U.S. expansion. Run 3 new AdWords tests.
StepsFind expansion announcements, identify ICP founders, send outreach, follow up, and launch the ad tests.Agree the increased ad budget and define the 3 tests on September 21.
Evidence & reviewOn October 2, compare replies and booking rates by message, then completed meetings and cost by source.
Work to postponeChanges to consultations or closing, and tests of additional acquisition channels.
Outreach hypothesis
Overseas founders planning U.S. expansion will book more meetings when outreach addresses transferring a founder or employee than when it offers general immigration help.
Illustrative test: compare an expansion-specific L-1 message with a general introduction. Offer a consultation on whether L-1 could fit.
12
Part 2 · Capacity planning
Budget time for outreach—and the meetings it generates
Illustrative estimates for the full two-week sprint. Confirm actual availability before committing.
| Planned work | Hours |
|---|---|
| Research and contact 300 distinct ICP prospects | 30 |
| Set up, launch, and manage 3 new AdWords tests | 10 |
| Follow up with prospects and schedule meetings | 10 |
| Hold petitioner meetings, including preparation and notes | 16 |
| Review results and plan the next sprint | 4 |
| Total planned work | 70 |
80 hours available
60 hours/week × 2 = 120 hours.
− 40 hours of existing commitments.
70 hours planned
+ 10 hours buffer
All 80 available hours are allocated.
Meeting hours reserve capacity; they do not forecast results. Ad spend is a separate budget.
13
Part 2 · Execute
Launch early, contact 300 prospects, and track meetings
Set the planConfirm goals, ad budget, 3 tests, ICP expansion signals, and two outreach messages.Sep 21
Launch and contactIncrease spend within budget. Launch 3 new AdWords tests. Contact the first 150 prospects.Sep 21–25
Continue and follow upContact the remaining 150. Follow up, hold meetings, and record their source.Sep 28–Oct 2
ReviewCheck both goals, then review bookings and completed meetings by source.Oct 2
Track steps and results separately.
Record contacts, replies, bookings, and completed meetings by outreach message. Track ad spend and results separately.
Compare the two outreach messages.
Use similar ICP prospects for both messages. Compare booking rates, not just counts. Track completed meetings separately from future bookings.
14
Part 2 · Review · Did it work?
Review meeting growth—and whether it can continue
October 2 review: planned questions, not actual results.
Accounting · establish the factsResponse · decide what changes
01Did we do the work?Goal 1: were 300 distinct ICP prospects contacted? Goal 2: were 3 new AdWords tests launched by October 2?
02Did it work?Which channels generated ICP bookings, at what cost? Did completed meetings rise above 2 per week?
03What did we learn?Did the expansion-specific L-1 message produce a higher booking rate than the general introduction? Is the evidence sufficient?
04What will we do next?Continue, adjust, or stop each test. Allocate the next sprint’s time and budget using the results.
A two-week sprint can show progress. Sustaining 10 meetings per week requires repeated weekly results, continuing prospect flow, and capacity to hold the meetings.
Compare ad spend and outreach effort separately. No-cost outreach still uses Achint’s time.
15
Part 3 · Moving forward
Run four two-week sprints between setup and final review
SetupSep 14 to 18
Sprint 1Sep 21 to Oct 2
Sprint 2Oct 5 to 16
Sprint 3Oct 19 to 30
Sprint 4Nov 2 to 13
Final reviewNov 16 to 20
Plan sprint 1 now. Use its results to choose the goal, ideas, and steps for sprint 2.
- Before sprint 1. Set your MTM target and baseline. Write the sprint 1 plan and schedule the check-in and review.
- At the final fellowship review. Assess MTM progress and choose the sprint that follows the fellowship.
16
Part 3 · Your turn
Write your first sprint plan
The numberOur MTM is , currently , with a November 20 target of .
Why this sprint?We think is holding us back.
GoalBy sprint end, our goals are .
IdeasThe initiatives we will try: .
StepsOur first step is , owned by , due .We have hours available. This work needs about hours.
Evidence & reviewOn we will review and decide .
Work to postponeTo make room, we will defer .
Make the goal measurable or observable: give each goal a specific number or observable outcome and deadline. At review, ask: did it happen?
Connect the evidence to your MTM: review the MTM and supporting evidence. If impact takes longer, set a follow-up date.
No measured starting value yet? Make measuring it your first step. Label any estimate.
17
Part 3 · Share and improve your sprint plans
Share your sprint plan and get feedback
- What problem is limiting progress? What evidence supports your explanation?
- Does each goal have a number or observable outcome? Which MTM driver or waypoint question does it address?
- Who owns each step? Do the estimated hours fit each person’s available time, with a buffer?
- What decision will you make using the review evidence?
- What work will you postpone to make time for this sprint?
Your sprint plan is ready when
it defines a measurable or observable goal tied to an MTM driver or waypoint question, names the ideas to try, assigns an owner and date to the steps, names the review evidence and decision, and lists work to postpone.
The facilitator picks one question per team, not all five.
ShareEach team reads its sprint plan.
FeedbackAsk the question most likely to improve each team’s plan.
Common issuesDiscuss issues shared by the four teams.
Follow-up stepsList assumptions still to check and assign follow-up steps.
18
Questions
Confirm who will do the first step and by when
Ask any questions you need answered before starting sprint 1.
Update your sprint plan. Confirm the first step, its owner, and its due date.
Sprint 1 starts September 21
Gateway X Fellowship 202619