Incubate-Iterate-Integrate for Innovating Inside a Mature Product
Build the new experience beside your core product, perfect it, then fold it back in — so you don't break existing customers
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 3
- Confidence
- 90%
To innovate inside a mature product with hundreds of thousands of customers, don't build directly into the core where you'll break millions of experiences. Instead incubate the new UX on the side (still on the platform), iterate with a small set of users until it's right, then integrate it back into the main product. This lets you escape the constraints of your existing UI/workflows when your audience needs something the platform can't give them.
Origin
Crusson credits an article by Noah Weiss (he says 'Ben way no sorry Noah way') on innovating in a successful mature product, using Instagram, Twitter, and Foursquare as examples (e.g. Instagram's side 'popular tab' becoming the Explore tab and eventually the main feed).
Core principles
- 01Don't break existing customers to experiment — carve a space away from the core while staying on the platform
- 02Escape your platform's UX constraints only when your new audience genuinely needs it
- 03Perfect the experience with a few users before you integrate it back
- 04Integration is the last step, not the starting assumption
How to run it
- 1
Incubate on the side
Build the new experience in the product but extracted from the core components, so experimentation can't break the experience millions of existing users depend on.
Pro tip Justify detaching from the platform UX only when it can't serve your audience — Crusson's PMs needed something snappy, visual, and free of Jira's strict workflows.
Watch out If the platform UX is already good enough for your audience, detaching is hard to justify — don't do it reflexively.
- 2
Iterate until it's right
Rebuild the UX that works for the new audience and iterate with a small number of customers over months until the experience genuinely fits their needs.
- 3
Integrate back into the core
Once the side experience is proven, integrate it back into the main product where portfolio fit makes sense.
Watch out Portfolio fit may reroute the feature elsewhere — Confluence Whiteboards started in Point A and was rolled into Confluence instead of shipping standalone.
In the wild
Product managers told Crusson they didn't want to work in Jira — too strict, too many workflows, owned by IT. So the team built JPD on the same platform but detached from Jira's core UX, creating a snappy, visual space PMs could present to stakeholders, then began integrating it back.
→ JPD delivered an experience its audience actually wanted; at time of recording the team was in the integration phase.
Common mistakes
Forcing the new product to live inside the core from day one
Feeling limited by your platform's UI leads you to ship a compromised experience your target users reject. Figure out first what the thing can be as its own experience, then integrate.
Is it for you?
Best for
A team building a new experience for a new audience on top of an existing large-scale product
Not ideal for
Cases where the platform's existing UX already serves the new audience well — added separation is unjustified overhead
From the transcript
“talked about three stages of incubate iterate integrate”
“we're going to build it in jir but we're going to kind of extract ourselves from the a lot of the core components”
“this concept of in incubate by working on something on the side iterate until you've got it right and then integrate it back into the…”
“most of the managers I talk to were like I don't want to do this in J it's too strict”
From the episode
Hard-won lessons building 0 to 1 inside Atlassian
Tanguy Crusson (Head of Jira Product Discovery)