All articles
13 min read

Sprint Planning for Startups: How to Plan Sprints That Hold

A small magnetic planning board with five index cards in a row, a mechanical kitchen timer, a pencil and an amber desk lamp on a steel shelf against a plain wall.

A good sprint plan for a startup comes down to three things: one clear sprint goal, a backlog that is ready before the meeting, and a workload sized to the time your team really has. Everything else is detail. If the goal is vague, the work is unclear or the plan assumes five full days from every person, the sprint fails before the first line of code is written.

This guide shows how to plan a sprint step by step, how long the meeting should take, which best practices matter for a team of two to eight people, and where founders usually go wrong. It is written for founders and product owners who run, or are about to start running, a small software team.

TL;DR

  • Sprint planning answers three questions: why this sprint is valuable, what can be finished, and how the work gets done.
  • Plan one sprint goal in a single sentence, then pull work from the top of a prioritized backlog until the team’s real capacity is used.
  • Keep the meeting short: roughly four hours at most for a two-week sprint, less when the backlog was refined beforehand.
  • Leave a buffer for support, bugs and surprises, and let every developer take part in the estimates.
  • When the plan keeps slipping, the cause is usually unclear requirements or architecture, which is where senior technical guidance pays off.

What sprint planning is and what it should produce

Sprint planning is one of the five events of Scrum, the most widely used agile framework. It is the meeting that starts a sprint, a fixed period (usually one to four weeks) in which the team builds a usable piece of the product. Scrum Alliance explains that the team uses it to plan the current sprint by discussing why the work is valuable, what backlog items can be completed and how the work will get done.

Those three topics give you the structure of the meeting, and Scrum.org’s guide to the event assigns them roles:

  1. Why: the product owner (the person who decides what is most valuable to build, often the founder at an early stage) proposes how the sprint could add value. The whole team then shapes this into the sprint goal.
  2. What: the developers select backlog items they believe they can finish, looking at capacity based on past performance, the Definition of Done (the team’s shared checklist for what “finished” means), the current product increment and lessons from the last retrospective.
  3. How: the developers break the selected items into tasks and decide how they will build them.

The outputs are a sprint goal and a sprint backlog, which is the selected items plus the plan for delivering them. A useful test comes from The Scrum Book: if the team can explain how it will reach the sprint goal, it has probably discussed the work well enough that finishing on time becomes more likely.

Choose a sprint length first

You cannot plan a sprint until you know how long it is. Most startups choose one or two weeks, some three. The trade-off is simple. Short sprints give faster feedback and let you change direction quickly. Longer sprints give momentum and recovery time with less meeting overhead. The authors of Scrum and XP from the Trenches settled on three-week sprints after experimenting, and their advice is to pick a reasonable length, try it for a sprint or two, then stick with it for an extended period.

For an early-stage product where priorities shift often, two weeks is a sensible default. It is long enough to ship something meaningful and short enough that a wrong bet costs you at most ten working days. Whatever you choose, do not change it every sprint. A stable length is what makes your past results a usable guide for the next plan.

How long should sprint planning take?

The time box depends on the sprint length. The Scrum Book gives a rule of thumb: no more than four hours for a two-week sprint, and proportionally less for shorter ones. The Scrum framework’s own ceiling is eight hours for a one-month sprint. Swiss practice shows the range: the IFDS Scrum wiki describes an eight-hour meeting split into two four-hour parts, with the product owner present in the first and the team planning alone in the second. At the other end, the ScrumC handbook published by confirm.ch uses a 30-minute planning meeting for two-week sprints, because the backlog items arrive already prepared.

The lesson for a startup: the better prepared the backlog, the shorter the meeting. If planning regularly runs past the time box, the problem is almost never the meeting. It is that the work was not ready.

As for when to hold it, plan on the first day of the sprint. Planning kicks off the current sprint, not an upcoming one, so schedule it at the start of that sprint, not the week before.

Before the meeting: get the backlog ready

The product backlog is the ordered list of everything you might build. Before planning, the top of it should be in a state where the team can pick items up without long debate. The ScrumC handbook defines a Definition of Ready for backlog items, a checklist an item must pass before it is eligible for a sprint. An item is ready when:

  • it is written as a user story, a short description of what a user wants to do and why;
  • its business value is stated;
  • the team understands it;
  • acceptance criteria are defined, meaning the conditions that must be true for it to count as done;
  • dependencies are identified and no external blockers remain;
  • the team has estimated it;
  • it is small enough to finish within one sprint.

This checklist is the single most useful tool for a startup, because founders tend to arrive with ideas, not ready items. A sentence like “add payments” is not ready. “A returning customer can pay by card at checkout and receives an email receipt” with the two or three conditions that prove it works, is.

Hold a short refinement session a day or two before planning to get there. Scrum.org lists “lack of backlog refinement” and “ambiguous requirements” among the common antipatterns in planning events, and both are fixed before the meeting starts, not during it.

How to plan the sprint, step by step

Here is a sequence that fits a small startup team. It combines the three Scrum topics with the practical steps in DigitalOcean’s startup guide and the goal-first model Scrum.org describes.

1. Start with the goal, not the list

The product owner brings the most valuable outcome for the next sprint. The team discusses what success would look like and agrees on one sentence. “Customers can sign up and book a first appointment” is a goal. “Work on the booking module” is a topic. A real goal gives you a filter for the whole sprint: when an unexpected request arrives, you ask whether it helps reach the goal. If not, it waits.

2. Confirm capacity

Capacity is the working time the team really has. Check holidays, time off, other commitments and who has which skills. Scrum Alliance notes that developers consider past performance, time off, holidays and skill availability when choosing work. Do this before you look at the backlog, so the number does not bend to fit your ambitions.

3. Select work from the top of the backlog

Pull items in priority order until you reach capacity. Do not skip down the list to grab something that looks easy. Each item should have an owner, an estimate and acceptance criteria. If an item fails the Definition of Ready, set it aside and take the next one.

4. Break items into tasks

The developers split each story into tasks small enough to finish in about a day or two. This is the “how” part of planning, and it exposes hidden work: a migration, an API credential, a design decision. Discuss dependencies here, on other team members, on outside vendors or on third-party systems, because that is where sprints stall.

5. Confirm the commitment

Ask the team directly whether the plan is achievable. Scrum.org’s guidance on the event stresses that developers select the work, not the product owner. A plan dictated from above is a wish.

6. Make the plan visible

Put the goal and the sprint backlog on a board everyone can see, whether a physical wall or a shared tool. Then start work the same day.

A worked capacity example

Here is the arithmetic, using the sample in GitScrum’s sprint backlog guide. A team of five developers has a ten-day sprint, which is 50 person-days on paper. After subtracting paid time off, training, meetings and support work, 33.75 person-days remain. The guide converts that to about 67 story points, but because the team’s last three sprints delivered 48, 55 and 53 points, it targets 50 to 55 and keeps a 10 to 15 percent buffer.

The point is the method, not the figures. Start from the calendar, subtract everything that is not building, then check the result against what the team actually delivered recently. If you have no history yet, because the team is new, start modestly and adjust after two or three sprints. Story points, t-shirt sizes or dot voting are all acceptable ways to estimate, as long as the whole team uses the same one.

If story points feel abstract, one simple option is Curtis Schlak’s four-value method: two hours for cosmetic changes, two days for most well-scoped work, two weeks for anything ambiguous, and “too long” as a signal to slice the story into smaller pieces.

Best practices for small startup teams

Write a single sprint goal. Without one, a sprint becomes a list of tickets. With one, you get a clear yes or no at the end and a concrete thing to discuss in the retrospective.

Estimate as a team. Scrum.org warns against only the team lead providing estimates: strong estimates combine everyone’s viewpoint and knowledge. The junior developer often spots the awkward integration the lead has forgotten.

Leave room for the unexpected. Startups get production bugs, customer requests and investor questions in the middle of a sprint. A buffer of 10 to 15 percent, as in the example above, keeps one surprise from sinking the whole commitment. Some small-team guides suggest more, up to 20 percent, when the same people also handle support and sales.

Let the team pull work. Instead of the founder assigning tasks, developers choose items from the top of the list. Ownership rises and bottlenecks around one person fall.

Keep refinement separate from planning. Preparing items during the planning meeting is why meetings run long. Do it earlier, in a short session.

Define done. Undone work carried over at the end of a sprint is another antipattern on Scrum.org’s list. Agree on what finished means (tested, reviewed, deployed to a staging environment) and apply it to every item.

Talk about value, not velocity. Velocity, the amount of work completed per sprint, is a planning input. It is not a goal. A team that discusses only points has lost sight of what the sprint is for.

Common mistakes that break a sprint plan

  • Planning for full days. Nobody codes eight hours a day. Plan from realistic capacity.
  • An unavailable product owner. If the founder skips planning, the team guesses at priorities.
  • Taking items from the middle of the backlog. It erodes priority and confuses the goal.
  • Overloading the sprint to look ambitious. Missing a commitment every sprint teaches the team that plans are fiction.
  • Estimating during planning. Unestimated, vague items slow the meeting and produce poor numbers.
  • Changing scope mid-sprint. Urgent exceptions happen, but if they happen every sprint, the sprint length is wrong or the priorities are not settled.
  • Treating the plan as a contract with the client. The sprint backlog is the team’s working plan, and it can be adjusted as they learn, provided the sprint goal stays intact.

When the plan keeps slipping: the problem behind the process

Many startups run a flawless-looking planning meeting and still miss the sprint. The reasons are rarely about the ceremony. Requirements change after work has started. A technical choice made in sprint one makes sprint four painful. Nobody on the team has the experience to say, early, that a feature is three days of work or three weeks. A process cannot fix those problems, because they are about judgment.

That is where a senior technical voice makes the difference: someone who reads a backlog item and sees the dependency, the risky assumption or the cheaper way to reach the same outcome, before the sprint is committed. Founders without a technical co-founder often have nobody in that seat.

How Ampersand helps with this

Ampersand is a small Zurich-based studio that builds websites, web and mobile apps, and platforms for startups and teams. Every project is led end to end by senior engineers, from the first consultation to handover, so the person who helps you scope the work is the person who stays with it. That continuity is designed to avoid common agency problems such as rewritten requirements and mid-project staff changes.

For founders who run their own team and want an experienced technical view on planning, we offer advisory and part-time CTO support, as well as product coaching. Our Guidance plan is billed at 799 CHF per month. If you want us to build the product itself, we work from fixed-scope project fees, and our Run & monitor plan for ongoing support is 750 CHF per month. You can see all options on our pricing page, or start from ampersand.ch.

A simple sprint planning agenda you can copy

For a two-week sprint with a refined backlog, this fits comfortably into a morning:

Block Purpose Suggested time
Goal Product owner presents the most valuable outcome; team agrees on a one-sentence goal 15 to 30 minutes
Capacity Check time off, commitments and recent delivery 10 minutes
Selection Pull ready items from the top until capacity is reached 30 to 45 minutes
Task breakdown Split items into tasks, name dependencies 60 to 90 minutes
Commitment Confirm the plan and make the board visible 10 minutes

The times are a starting point and sit well inside the four-hour rule of thumb for a two-week sprint. A team new to working together may need the full time box for its first sprints. A mature team with a well-refined backlog may finish in under an hour. After each sprint, ask whether the meeting was too long, too short or missing a decision, and adjust.

FAQ

Is a sprint one or two weeks?

Either works. Scrum allows sprints of up to one month, and many startups choose one or two weeks for faster feedback. Pick a length, run it for a couple of sprints and keep it stable so your past results become a reliable guide for planning.

When should sprint planning be done?

On the first day of the sprint. Planning starts the current sprint, so it belongs at the beginning of it. Refine the backlog a day or two earlier, so that the planning meeting only needs to select and break down work.

How long should sprint planning take?

No more than about four hours for a two-week sprint, and proportionally less for a one-week sprint. With a well-prepared backlog it can take far less. Swiss teams using the ScrumC approach plan in around 30 minutes because their items are already ready.

Who needs to attend sprint planning?

The product owner, who brings priorities and the goal, and the developers, who select the work and plan how to build it. In a startup the founder is often the product owner. Anyone whose input changes the estimate, such as a designer, should join.

What if we do not finish everything in the sprint?

Review why in the retrospective. Common causes are overestimated capacity, unclear items and mid-sprint scope changes. Return unfinished work to the backlog, re-prioritize it and plan the next sprint on the capacity you actually delivered.

Sources

Talk to us

Have a project this touches on?

A free 10-minute call is the fastest way to find out whether we are the right studio for it.

Book a free 10-min call