Growth Model Evolution (18-Month Rule)

"If you have a growth model that works for you, that's wonderful, good for you. Optimize it, grow it, scale it, create a team that will be nurturing it and that will be amplifying it, but you're going to need to evolve it, and that evolution needs to come through overlaying other growth models on top of it." - Elena Verna

What It Is

Growth Model Evolution is the principle that growth engines have natural lifecycles and companies must continuously layer new growth models to sustain momentum. Most growth loops produce meaningful results for only 5-7 years before spinning out, so teams must introduce new models every 18 months to ensure the pipeline of growth engines never runs dry.

This challenges the common practice of doubling down on what's working while neglecting the search for what's next.

How It Works

The S-Curve Reality

Every growth model follows an S-curve:

1. Early stage: Slow, experimental, lots of failure
2. Growth stage: Rapid scaling, high returns
3. Maturity stage: Diminishing returns, optimization yields less
4. Decline stage: Law of Shitty Clickthroughs kicks in

Most growth loops spin out within 5-7 years. Even exceptional ones (like Dropbox's sharing loop at 17+ years) are anomalies.

The 18-Month Introduction Cadence

"Every 18 months you need to introduce something new... every 5 years or so you for sure need new channel, new growth loops, new tactics, new engines."

The math works out:

• New growth loops take 6 months to 1.5 years to show visible results
• You need runway for many to fail
• If you wait until decline to search, you're already behind

The 20-25% Investment Rule

Elena recommends allocating 20-25% of growth team time annually to:

• Introduce new growth loops
• Test new channels
• Try new approaches

This time is:

• Not goaled on immediate metrics
• Expected to have high failure rates
• Measured differently than optimization work

Real Example: Mirrorverse at Miro

"Mirrorverse, which is a user-generated content library of all of the Miro boards that people create, it took us probably 18 months until we started putting metrics expectations on it. Before that, it was a thing that we were testing... And then it started taking off as both engagement engine as well as acquisition engine."

The lesson: New growth engines need time to mature before being held to metrics.

How to Apply It

1. Map your current growth model portfolio

   • What models are mature/declining?
   • What's in growth phase?
   • What's in early experimentation?

2. Assess vulnerability

   • How dependent are you on one growth model?
   • What happens if your main model declines by 50%?

3. Create protected experimentation time

   • Dedicate 20-25% of growth resources to new models
   • Don't goal this time on acquisition or monetization
   • Give experiments 6-18 months before metrics expectations

4. Think across model types

   • Product-led growth models
   • Marketing-led growth models
   • Sales-led growth models
   • Layer all three over time for diversification

5. Start before you need to

   • Don't wait for decline
   • The worst position: declining metrics + no pipeline of new engines

When to Use It

• During annual growth strategy planning
• When current growth is strong (counterintuitively, the best time to plant seeds)
• When noticing diminishing returns on primary channels
• When debating whether to "double down" vs diversify

The Alternative (Failure Mode)

Companies that don't evolve their growth models:

• Experience a "really big blip" of growth
• Potentially reach unicorn status
• Then start slowing down with no next act
• Leave gaps in the market for competitors to fill

Source

• Guest: Elena Verna
• Episode: "10 growth tactics that never work"
• Key Discussion: (00:50:58) - Evolution of growth models
• YouTube: Watch on YouTube

Related Frameworks

• Compounding Growth Loops - The types of loops to layer
• Distribution Platform Cycle - How platforms evolve
• Law of Shitty Clickthroughs - Why optimization yields diminishing returns