Commit Only to Dates Within Your Control
Roadmap the phase, not the feature: Discovery, then Solutioning, then Build, then a ship date.
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 4
- Confidence
- 95%
Calendly stopped promising delivery dates for work it had not yet discovered or designed. Instead the product development life cycle is split into four phases, and the team only commits to the phase directly in front of it. A quarterly roadmap item can legitimately be 'run Discovery on problem space X' — with the next phase earned, not assumed. An engineering delivery date is only given after Solutioning, when the solution is actually known.
Origin
Developed at Calendly under Annie Pearl over the year prior to this February 2023 conversation, as a correction to the standard practice of committing to feature dates six months out.
Core principles
- 01You can only estimate what you have already designed; a date for an undiscovered problem is fiction.
- 02Each phase is a commitable body of work with a knowable end, even when its outcome is unknown.
- 03The outcome of Discovery may legitimately be 'we are not proceeding' — that is a success, not a failed commitment.
- 04Precision increases as you move through the phases; only the final phase carries a delivery date.
How to run it
- 1
Discovery
Commit to a bounded research effort on a problem space with a known conclusion date. Do not commit to what will be built as a result — you do not yet know whether you will proceed at all.
Pro tip Put Discovery itself on the quarterly roadmap as the deliverable; if it goes well, the next phase takes the next quarter's slot.
Watch out Stakeholders will try to extract a build date here. Refuse — nothing after this phase is yet knowable.
- 2
Solutioning
If the problem space is worth pursuing, commit to exploring a few different solutions, running user testing, and landing on one. This is again a commitable phase with an end, not a delivery promise.
- 3
Build (and only now, estimate)
With both the problem and the solution known, engineering can do estimation planning and give an actual delivery date. This is the first point at which a ship date is honest.
Pro tip Announce the date only at this gate — that way every date you give is one you can hold.
Watch out Committing a date before this gate is where roadmap credibility is destroyed.
- 4
Launch, measure, and iterate
Ship, instrument, and iterate. Calendly uses in-product education (Pendo) to onboard users onto newly launched features.
In the wild
Calendly designed its whole product development life cycle around the four phases. A quarter's roadmap item may simply be a Discovery effort on a problem space; if it concludes positively, the following quarter carries the Solutioning phase, and only after that does an engineering delivery date appear.
→ The team got significantly better at making commitments about the work directly in front of it, instead of promising a project six months out with no discovery or design done.
Common mistakes
Giving a six-month-out delivery date
Pearl is explicit: committing to a project six months out, before enough Discovery, design, and ideation has happened, means there is no real basis for the estimate — the date is invented.
Treating Discovery as a formality on the way to a pre-decided build
The phase only works if 'we are not moving forward with investing in that area' is a genuinely permitted outcome. If the build is already assumed, you have re-created the fake date with extra steps.
Is it for you?
Best for
Product leaders whose roadmaps keep slipping because dates are committed at the idea stage, and who need a defensible way to say 'not yet' to date-hungry stakeholders.
Not ideal for
Hard-deadline contexts (regulatory, contractual, event-tied launches) where a fixed date is a genuine constraint and scope must flex instead.
From the transcript
“we've done over the last year is is really kind of moved to a model of talking about dates and promising and committing to dates…”
“making the commitments around the work that's right in front of us versus making a commitment around a project six months out when we haven't…”
“yeah so the first phase you know we sort of just call generally Discovery um the second phase we call kind of solutioning the third…”
From the episode
Behind the scenes of Calendly’s rapid growth
Annie Pearl (CPO)