Back to glossary

Now-Next-Later

A roadmapping format that organizes initiatives into three time horizons without committing to specific dates. Now items are in active development, Next items are planned soon, and Later items are being explored or researched.

Now-Next-Later roadmaps address a chronic problem with traditional timeline roadmaps: false precision. Committing to specific delivery dates months in advance creates pressure to ship features regardless of whether discovery validates them. By using relative time horizons instead, teams maintain strategic direction while preserving the flexibility to respond to new information.

This format is ideal for AI product teams because the uncertainty inherent in AI development makes fixed timelines unreliable. A model training effort might take two weeks or two months depending on data quality issues discovered along the way. Now-Next-Later lets teams communicate priorities honestly. Growth teams benefit because the format naturally accommodates experimentation: now might include a specific A/B test, next might include scaling the winning variant, and later might include exploring more sophisticated ML approaches. Stakeholders get visibility into direction without the team making promises that AI development timelines cannot reliably keep.

Related Terms