Outcome-Based Roadmap: How to Kill the Gantt Chart (With Template)
Feature-by-date roadmaps grade the wrong thing. Here is how to build an outcome-based roadmap with Now/Next/Later, keep the few dates that matter, and copy a working template.
Share
An outcome-based roadmap plans product work around the customer behaviour you want to change, not the features you will ship by a date. Each item names an outcome (for example, "raise week-1 activation from 28% to 40%"), the evidence behind it and the bets you will test. The most common format is Now/Next/Later, which swaps the calendar for three levels of certainty.
Key takeaways
Most roadmaps are still feature lists. In ProductPlan's 2023 survey of more than 1,500 product managers, 54% of roadmaps were designed around outputs and only 44% communicated outcomes (ProductPlan, 2023).
Feature lists ship features nobody uses. Pendo's analysis of 615 product subscriptions found 80% of features were rarely or never used, while 12% of features drove 80% of daily usage (Pendo, 2019).
The Now/Next/Later format started near the end of 2012, when Janna Bastow and Simon Cast of ProdPad replaced a timeline with three columns (ProdPad, 2019).
Dates don't vanish. Keep them for the handful of real commitments, which Marty Cagan says should be the exception, not the rule (SVPG, 2020).
The best outcome sits at the product level: a behaviour change one team can move. Not revenue, and not clicks on a single feature.
What is an outcome-based roadmap?
A roadmap tells people where the product is going and why. The difference between roadmap types is the unit you plan in.
An output is the thing you make: a feature, a page, an integration. An outcome is what changes because you made it. Josh Seiden's widely used definition: an outcome is "a change in human behavior that drives business results" (Intercom, 2019). Revenue, retention and margin are the business results further downstream.
So "ship the onboarding wizard" is an output. "More new teams finish setup in their first week" is an outcome. The wizard is one guess at how to get there. If a setup checklist works better, you swap the bet and the roadmap item stays the same.
Teresa Torres and Hope Gurion split outcomes into three levels (Product Talk, 2022):
Level
Example
Who can move it
Good roadmap item?
Business outcome
Grow net revenue retention
Several teams together
Too broad for one team; use it as the "why"
Product outcome
Teams invite a second user in week 1
One product team
Yes, this is the sweet spot
Traction metric
Clicks on the new invite button
One team, easily
Too narrow; it presumes the solution
Torres's test for a traction metric is simple: can a customer be happy without ever using that feature? If yes, the metric is too narrow to plan around.
Feature roadmap vs outcome roadmap, side by side
Output (feature) roadmap
Outcome roadmap
Unit of planning
Features and projects
Customer behaviour changes
Top axis
Calendar months or quarters
Certainty: Now, Next, Later
Success means
Shipped on the date
The metric moved
When evidence changes
Rewrite the roadmap, renegotiate dates
Swap the bet, keep the outcome
Question it answers
"What are we building, and when?"
"What are we trying to change, and why?"
Why do Gantt-chart roadmaps fail product teams?
A Gantt chart is a fine delivery tool. It is a poor roadmap, for three reasons.
It measures the thing that slips least. McKinsey and the University of Oxford studied more than 5,400 IT projects with budgets above $15 million. On average they ran 45% over budget and 7% over time, and delivered 56% less value than predicted (McKinsey, 2012). These were large IT programmes, not product teams, but the shape is telling: the schedule miss was small next to the value miss. A timeline roadmap tracks dates, the 7%, and stays silent on value, the 56%.
Long plans get worse with length. The same study found every additional year of scheduled duration raised cost overruns by 15%. A 12-month Gantt chart locks in guesses made when you knew least.
It grades shipping, not results. If 80% of features in the average product go largely unused (Pendo, 2019), a roadmap that counts features delivered is counting mostly waste. Janna Bastow makes the same point from the other side: a timeline forces a date onto every item, which creates pressure and padding (ProdPad, 2023).
Where did the Now/Next/Later roadmap come from?
Janna Bastow, co-founder of ProdPad and Mind the Product, developed the format with co-founder Simon Cast near the end of 2012, in a café in Wandsworth, south London (ProdPad, 2019). A customer had asked to move a large number of roadmap items at once. That request showed the team that people were not delivering to timeline roadmaps anyway. They decided a roadmap should be a strategic communication tool, not a release plan.
The first version had three columns with plain names: Current, Near Term and Future. They later became Now, Next and Later:
Now: the outcomes you are actively working on. Highest certainty, bets are scoped, the team is busy here.
Next: problems you have chosen and are still shaping through discovery. Medium certainty.
Later: directions you believe in but have not sequenced. Lowest certainty, fuzzier scope.
The columns describe confidence, not calendar slots. That is the whole trick: no one can read "Later" as "October 18".
How do you build an outcome-based roadmap? A 6-step method
This is the sequence we recommend for a team of 3 to 30 people. Budget about half a day for the first pass.
Start from the strategy, not the backlog. Write down the one or two bets your strategy makes and the metric that says customers get value. If you have neither, do that first with a one-page strategy canvas and a North Star metric.
Translate each business outcome into 2-3 product outcomes. "Grow retention" becomes "teams invite a second user in week 1" and "teams create a second project by day 30". These are behaviours your team can influence.
Write each outcome as metric, baseline, target and check-in date. "Week-1 activation: 28% → 40%, reviewed after 6 weeks." No baseline means you cannot tell whether you moved it, so measuring the baseline is often the first task.
Attach the evidence and the riskiest assumption. List the interviews, tickets or usage data that make you believe this outcome matters, and the one assumption that would kill it if false.
List bets under each outcome, as hypotheses. "We believe a setup checklist will lift week-1 activation because 9 of 14 churned trials stalled at setup." Bets can change weekly; the outcome should hold for a quarter or more.
Set a review rhythm. Review the Now column monthly and the whole roadmap quarterly. Items move between columns when evidence arrives, not when a date passes. Turning evidence into a ranked plan is covered in our guide to going from customer feedback to a product roadmap.
What does a Now/Next/Later roadmap look like? Template and example
Copy this table into Notion, Confluence, a doc or a spreadsheet. One row per outcome; keep the Now column to 1-3 rows per team.
Horizon
Outcome (metric: baseline → target)
Why we believe it (evidence)
Bets we will test
Riskiest assumption
Confidence
Next review
Now
High
Now
High
Next
Medium
Next
Medium
Later
Low
Worked example: a 6-person B2B SaaS team (illustrative)
Say a 6-person team sells a reporting tool to finance teams. Their strategy bet is "win by being the fastest way to a first board-ready report". Their filled-in roadmap:
Horizon
Outcome
Evidence
Bets
Confidence
Now
Week-1 activation: 28% → 40% by end of quarter
9 of 14 churned-trial interviews stalled at data import
CSV import wizard; pre-built sample report
High
Now
Time to first report: 3 days → under 1 day (median)
Usage logs show a 3-day median gap between signup and first export
Template gallery; guided first report
High
Next
Second user invited within 14 days: 18% → 30%
Last year, accounts with 2+ users renewed at 85% vs 61% for single-user accounts
Share-a-report link; comment notifications
Medium
Later
Finance teams run month-end close inside the product
4 sales calls asked for it; untested
Not yet chosen
Low
Notice what is missing: dates on the Next and Later rows, and any feature without an outcome above it. If the CSV wizard fails to move activation after 6 weeks, the team swaps in a different bet and the Now row stays. If you want a ready-made file, our free Now/Next/Later roadmap template downloads as a CSV with example items and success metrics, no account needed.
How do you handle deadlines on a Now/Next/Later roadmap?
Some dates are real: a trade show, a partner contract, a tax-season deadline, another team waiting on your API. Cagan calls these high-integrity commitments and gives three rules for them (SVPG, 2020):
Do discovery before you commit. Judge whether the solution is valuable, usable, feasible and viable, often with a feasibility prototype, then name the date.
Track them separately from key results. A commitment is binary: you deliver it or you don't, and you raise problems early.
Keep them rare. If most objectives turn into dated deliverables, Cagan warns they become "little more than a reformatted roadmap".
In practice, add a "Committed: [date]" label to the one or two items that carry a real external date, usually in the Now column. ProdPad's own guidance makes the same exception for legal and contractual dates (ProdPad, 2023). Everything else carries a review date, not a delivery date.
How do you get stakeholders to accept a roadmap without dates?
Stakeholders ask for dates because dates are the only proof of progress a feature roadmap offers. Give them better proof:
Report movement, not percent complete. "Activation went from 28% to 33%; the import wizard explains most of it; the sample report did nothing, so we dropped it."
Name what you are not doing. A "not now" list prevents the same request coming back every quarter.
Pair it with OKRs, don't duplicate them. The roadmap's outcomes and the quarter's key results should be the same numbers. Our post on how product teams use OKRs shows the link.
Common mistakes and when this doesn't work
Features renamed as outcomes. "Launch onboarding v2" in the Now column is still an output. Test each row: could two different features achieve it? If not, it is a feature.
Revenue as the outcome. Revenue lags by months and depends on sales, pricing and marketing. Put it in the "why", and plan against the product behaviour that leads it.
Traction metrics. "Increase use of the export button" presumes the solution. Use Torres's happy-customer test from earlier.
Later as a graveyard. If an item has sat in Later for three or more review cycles with no new evidence, delete it or move it to a parking list.
No baseline. A target without a starting number cannot be judged. Measure first, even if that takes two weeks.
When a Gantt chart is the right tool. Outcome roadmaps assume you can change course. They fit poorly when scope is fixed by contract (agencies, enterprise deals with roadmaps locked 12-18 months out), by regulation, or by hardware lead times. Even then, keep the Gantt chart as the delivery plan for one committed project and keep the roadmap about outcomes. The two can coexist; they answer different questions.
Where Specky fits
The hardest part of planning by outcomes is the evidence column: proving why an outcome matters and whether a bet moved it. Specky reads customer calls, feedback, Slack, Jira and analytics into a Product Graph, so each outcome can link to the signals behind it, and it checks after launch whether the bet worked. Strategy Canvas, OKRs and the Decision Log sit in the same place, so the "why" travels with the roadmap instead of living in someone's head. Faster learning loops matter more than faster shipping, as we argue in learning velocity as a competitive advantage, and scaling teams need the same loop across many squads, which we cover in the product operating model for scaling teams.
Frequently asked questions
What is the difference between an outcome-based roadmap and a feature roadmap?
A feature roadmap lists what you will build and when; an outcome roadmap lists the customer behaviour changes you want and why. In a feature roadmap, success means shipping on the date. In an outcome roadmap, success means the metric moved. Features still appear, but as bets under an outcome that can be swapped when evidence says they are not working.
Who invented the Now/Next/Later roadmap?
Janna Bastow and Simon Cast, the co-founders of ProdPad, developed it near the end of 2012. The first version used three columns called Current, Near Term and Future, which later became Now, Next and Later. They built it after seeing that teams were not delivering to timeline roadmaps, and decided a roadmap should communicate strategy rather than act as a release plan.
Can an outcome-based roadmap have dates?
Yes, but only where a date is real. Keep firm dates for the few items with external constraints, such as a contract, an event or a dependent team, and do enough discovery before committing. Marty Cagan calls these high-integrity commitments and says they should be the exception. All other items carry a review date for checking the metric, not a delivery date.
How many outcomes should be in the Now column?
One to three per team. Each Now outcome needs active bets, a baseline and someone checking the metric every few weeks, and a small team cannot do that well for more than three at once. If your Now column holds eight items, most of them are really Next items without anyone working on them yet.
How does a Now/Next/Later roadmap relate to OKRs?
They should share the same numbers but do different jobs. OKRs set the target for a period, usually a quarter. The roadmap shows sequence and certainty across several periods: what you are working on now, what comes next and what is a direction. A good test is that each Now outcome maps to a key result.
Is Now/Next/Later suitable for a startup?
Yes, and it often fits early teams better than timelines because priorities change monthly. A solo founder can run it as one page with three columns and one outcome in Now. Start from your strongest customer evidence, pick the single behaviour that best predicts retention, and review the board every two weeks.
GDPR-aligned. EU workspace data is stored in Frankfurt, Germany. GDPR and EU data Built in Vienna, Austria. A product of influence solutions fm GmbH (FN 675374y). Imprint 14-day trial. Card required, and nothing is charged until day 14. 30-day refund after subscribing. See pricing