LLenny's Podcast
← All frameworks
InnovationMegan Cook (head of product, Jira)

Stage-Gated Product Incubation

Grow new products through five funded stages — wonder, explore, make, impact, scale — validating at each gate.

Difficulty
Advanced
Time to result
~months to results
Steps
5
Confidence
92%

A way for a large company to incubate new products with a startup mindset. Ideas pass through five stages, each adding a little more investment and headcount only after passing a gate check. The goal at every stage is to iterate fast toward proof — either that it can be a business, or that it can't — while protecting the small team from heavyweight company processes.

Origin

Atlassian's incubation model as described by Megan Cook; produced Jira Product Discovery, Atlas, and Compass.

Core principles

  • 01Stage investment against validation — add people only after proof
  • 02Small teams move fast; premature scaling slows discovery
  • 03A gate is a real kill/continue decision, not a rubber stamp
  • 04Hold incubated products to a different standard than mature ones
  • 05Innovation can come from anyone — let anyone pitch

How to run it

  1. 1

    Wonder — one person with an idea

    Let anyone in the company pitch a new-product idea; the first stage may be just the person who had it.

    Pro tip An internal innovation program plus hackathons ('ShipIt') keeps a pipeline of ideas flowing.

  2. 2

    Explore — a tiny team validates

    Add a couple more people (about three) to build a prototype, find interested customers, and put together what the roadmap looks like.

    Pro tip The prototype can be simple and cobbled together — its job is to learn, not to ship.

  3. 3

    Make — a full (small) team builds it

    Give it a full team of around a dozen people to actually build the product and try to win more and more customers.

    Pro tip Even a 'full team' stays around 12 people — resist throwing four or five teams at it.

    Watch out Getting excited and over-staffing early makes things run slower, not faster.

  4. 4

    Impact — prove it can sustain itself

    Reach the point where the product is self-sufficient in the revenue it generates.

  5. 5

    Scale — launch it broadly

    Invest to take the validated product to full scale and put it on the website.

    Pro tip Let the incubated team solve problems differently — their innovations can be pulled back into the core products.

    Watch out Holding an early-stage product to the resilience/scale standards of a mature product will slow it down far too much.

In the wild

Jira Product Discovery

PM Tony Christian saw demand for a tool to shape roadmaps before ideas get committed to Jira. Rather than force it through processes optimized for big products, Atlassian ran it through the stage gates, partnering heavily with customers already signalling interest in existing products.

It found product-market fit and became a shipped product; some of its spreadsheet-style ideas are now being pulled back into Jira Software.

The whiteboard product that got killed

A whiteboard product was being incubated as its own product, but at a gate the team realized it was really a useful document type, not a standalone product.

They killed it quickly as a standalone and shipped it as a feature inside Confluence.

Common mistakes

Over-resourcing a seed idea

Throwing many teams at an exciting early idea creates coordination drag and slows the very validation you're trying to accelerate.

Running incubated products through full company process

Company-wide planning, business reviews, and quality bars optimized for big products crush a seed startup that isn't ready for them.

Is it for you?

Best for

Large, multi-product companies wanting to seed new bets with a startup mindset without smothering them in process

Not ideal for

Early-stage startups that are themselves still hunting for their first product-market fit

From the transcript

we changed that to create really small groups with stage gates that we call wonder, explore, make, impact, and then getting to scale

51:30

at each at each one of those stages there's that check on whether or not we continue to go ahead

54:00

starting really small with the idea and solution to be proved and don't don't add too many people

56:00

for something small like this where you're proving product market fit, you don't need to think about that initially. So, you you've got to hold…

57:00

From the episode

Lessons from Atlassian: Launching new products, getting buy-in, and staying ahead of the competition

Megan Cook (head of product, Jira)