Product roadmap template

A product roadmap is a shared view of what you intend to build, in what order and why. A product roadmap template gives you the page to fill in. For the full definition and how roadmaps differ from backlogs and release plans, read What is a product roadmap? and the glossary entry. This page is the part you paste.

TL;DR

• Group the work by bet, the outcome you are after, not by team or by date.
• Give each item what becomes true, the evidence and its stage.
• Use now, next and later, or an outcome and a number. Skip dates past the next few weeks.
• Keep it short enough that a customer, a teammate or an agent can read it in a minute.

When to use it: you have more than one thing to build and someone asks "why that first?" When not: you are building one thing this month. A to-do list is enough, and a roadmap would be a second place to keep it.

A product roadmap template

Copy it, fill the brackets, delete what you do not need. One bet per heading, a few items under each.

# Roadmap: <product or team>, <month>

## Bet 1: <the outcome you are after, in a line>
Why this bet: <one sentence>

- **<What becomes true, in a customer's words>**
  - Evidence: <the requests in their words, how many accounts, revenue, or "an idea, no customer yet">
  - Stage: <open | ready | built | shipped | shared>
- **<What becomes true>**
  - Evidence: <...>
  - Stage: <...>

## Bet 2: <outcome>
Why this bet: <one sentence>

- **<What becomes true>**
  - Evidence: <...>
  - Stage: <...>

## Bet 3: <outcome>
- **<What becomes true>**
  - Evidence: <...>
  - Stage: <...>

## Not now
- <what you decided not to do> because <a few words>

## Order inside each bet
Broken first, then what customers asked for, then ideas. Within each, more accounts and more revenue first.

## Last checked
<date>. Next look: <date>.

What goes in each part?

Bet. An outcome you are after, in a line: "Customers who start get to a first result in a day." Two to four bets is plenty. If you cannot name the outcome, the items under it are probably a wishlist. In Annsa you set your bets and their order.

Item. What becomes true, in words a customer would use: "Import a spreadsheet and see your first list", not "CSV importer v2". Nine words or fewer helps.

Evidence. The requests in the customer's words, how many accounts, what they pay. "One loud user" and "five quiet accounts" are different lines, so write which. An idea with no customer behind it is fine. Say so.

Stage. Where it stands: open, ready (the spec is written), built, shipped, shared (you told the people who asked). A stage tells a reader what to expect without a status meeting.

Order. Inside a bet, put what is broken first, then what customers asked for, then ideas. Within each, more accounts and more revenue come first. Order is the real content of a roadmap: if it cannot explain why one item is above another, it is a list.

Not now. Write down what you decided not to do, with a few words of why. It saves the next debate.

Three examples, filled in

All three are examples, made up to show the level of detail. None is a real product.

An example: a startup roadmap

One founder, a few bets, no dates.

Bet 1: New customers get a first result in a day

• Import a spreadsheet and see your first list. Evidence: 6 of 9 trial accounts stopped at the empty screen. Stage: built.
• Invite a teammate in one click. Evidence: 2 accounts asked, $98 a month together. Stage: ready.

Bet 2: Customers we have stay

• Pause a subscription instead of canceling. Evidence: 3 accounts in 5 weeks, 2 more canceled and said "seasonal". Stage: open.
• Invoices open on the first click. Evidence: found by an agent, marked broken (P1). Stage: shipped.

Bet 3: People hear about us

• A page for the one question 40 people asked. Evidence: an idea from search data, no customer yet. Stage: open.

Not now: a mobile app. Nobody has asked, and the web version works on a phone.

An example: a now/next/later roadmap

Columns by confidence, not by calendar.

Now (being built)

• Pause a subscription instead of canceling. Evidence: 3 accounts. Stage: ready.

Next (understood, waiting its turn)

• Invite a teammate in one click. Evidence: 2 accounts, $98 a month.
• Export a report as a PDF. Evidence: 4 accounts, 1 of them large.

Later (an idea, not yet checked with customers)

• Connect to a calendar. Evidence: a teammate's idea, none yet.

Items move left when the evidence grows or the work is done, not when a date arrives.

An example: an outcome-based roadmap

Each bet has a number you will check.

Outcome: more trial accounts reach a first result. Today 3 of 9. Aim: 6 of 9 by the end of next month.

• Import a spreadsheet and see your first list. Evidence: 6 of 9 stopped at the empty screen. Stage: built.
• A checklist of three steps on day one. Evidence: 4 trial interviews. Stage: ready.

Outcome: fewer cancellations for seasonal reasons. Today 2 in 5 weeks. Aim: 0.

• Pause a subscription instead of canceling. Evidence: 3 accounts. Stage: open.

Check the number after each ship. If it did not move, the item was a guess. Note that and look again.

When is a roadmap overkill?

When you are building one thing, or the work is a week long. Then a list is enough. A roadmap is for the moment you have more to build than you can, and someone, including you, needs to know why this comes first. If you are weeks from that moment, write a PRD for the one thing and stop there.

How does Annsa help?

Annsa is the source of direction for your agents. The roadmap is a view of it.

• The roadmap is your priorities grouped by bet, kept current as status changes (open, ready, built, shipped, shared). You do not copy it anywhere or update it by hand.
• Evidence comes in on its own. Requests arrive from Slack, Intercom, call transcripts, surveys, CSV and Google Sheets imports, GitHub, the API or by hand. Each keeps the person, their exact words, the account and its revenue, and matching requests are grouped.
• The order is explained. Inside your bets, in your order: what a person marked broken first, then what customers asked for, then ideas. Then severity, how many accounts, revenue and urgency. You choose; Annsa never decides for you.
• Agents use it. Claude Code, Cursor, Codex and Grok connect over MCP, claim a priority, read the spec and record evidence against each Done When line.

See the roadmap and how a priority is ranked.

Sources: Technology roadmap, Wikipedia, for where roadmaps come from.