LLenny's Podcast
← All frameworks
StrategyNoah Weiss (Slack, Foursquare, Google)

The Diversified Roadmap Portfolio

Treat a product roadmap like an investment portfolio, balanced across three axes of risk and intent.

Difficulty
Moderate
Time to result
~months to results
Steps
4
Confidence
83%

Weiss frames a feature team's roadmap as a diversified portfolio rather than a flat backlog. It should be balanced across three tensions: new capabilities vs. making what exists a little better every day; risky high-upside bets vs. known bets; and work meant to drive confident impact vs. work meant to learn about a new possibility space. Slack operationalizes the 'better every day' bucket through time-boxed Customer Love Sprints.

Origin

Articulated by Noah Weiss as Slack's approach to roadmap construction across feature teams.

Core principles

  • 01A healthy roadmap is deliberately diversified, not concentrated in one type of bet
  • 02New-capability work and incremental-improvement work must both be funded
  • 03Learning-oriented work has value even when it drives no immediate metric
  • 04Balancing known bets against risky bets prevents both stagnation and reckless swings

How to run it

  1. 1

    Split by new vs. incremental

    Allocate roadmap capacity between building genuinely new capabilities and making the thing you've already shipped a little bit better every day.

  2. 2

    Balance risky bets against known bets

    Include some bets you aren't sure will work but that carry large upside, alongside bets whose outcome you're already confident in.

  3. 3

    Balance impact work against learning work

    Separate work meant to drive impact you're already confident in from work meant to learn about a new possibility space, and fund both intentionally.

    Watch out Don't let confident-impact work crowd out learning work — the learning bets are what surface the next levers with years of room to run.

  4. 4

    Fill the 'better every day' bucket with Customer Love Sprints

    Because small polish work is hard to allocate across a quarter, run a dedicated two-week sprint on the lowest-effort, highest-impact changes to generate customer love.

    Pro tip Aim to ship all of them — this isn't throwaway hack-week code — and run it at least once a quarter for very user-facing teams.

In the wild

Customer Love Sprints at Slack

Slack teams run a two-week 'customer love sprint' — like a hackathon but working off a burn-down list of the lowest-effort, highest-impact changes that generate customer love in a feature area. Design, product, and engineering sprint together and demo a batch of shippable delight fixes at the end.

Small delightful improvements that customers love get shipped (not thrown away), and the change of pace re-energizes teams between multi-month feature efforts.

Common mistakes

Running a roadmap of pure incrementalism

Concentrating only on making existing metrics tick up a percent a quarter loses sight of what's beyond the horizon and starves the bigger, bolder bets that create new S-curves.

Refusing to fund learning bets

Because learning work may drive no metric this quarter, it gets cut first — but killing it means never discovering the new levers that pay off over years.

Is it for you?

Best for

Product leaders and PMs constructing a quarterly or annual roadmap for a maturing product with multiple feature teams.

Not ideal for

Very early startups still searching for initial product-market fit, where focus beats diversification.

From the transcript

the way that we think about it or I like us to think about our roadmap for any feature team at slack is that it's…

30:00

are there things that are meant to be risky that you aren't sure are going to work but might have a lot of upside there's…

30:30

almost like a hackathon but with that kind of burn down list of what we think is the lowest effort highest impact changes we can…

31:30

From the episode

The 10 traits of great PMs, how AI will impact your product, and Slack’s product development process

Noah Weiss (Slack, Foursquare, Google)