The 18-Month Growth Model Overlay
Spend ~20-25% of growth time annually seeding new loops so your S-curve never flattens with no successor.
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 4
- Confidence
- 90%
Every growth model and loop has an S-curve that eventually flattens (Andrew Chen's 'law of shitty click-throughs'). Verna's rule: continuously overlay new growth models (product-led, marketing-led, sales-led) so you're diversified and always have a second horizon. Concretely, allocate 20-25% of the growth team's annual time to seeding new loops, introduce something new every 18 months, and expect a genuinely new engine to take meaningful volume every ~5 years.
Origin
Elena Verna, referencing Andrew Chen's 'law of shitty click-throughs' and illustrated by Miroverse.
Core principles
- 01Over-optimizing the same lever yields minimal returns (law of shitty click-throughs)
- 02Most growth loops stop producing meaningful results within 5-7 years; a still-firing 17-year loop is an anomaly
- 03New loops take 6-18 months to produce visible revenue, so you must seed them before the current one slows
- 04Don't goal a new loop on growth metrics immediately or you'll cut it off at the knees before it can evolve
How to run it
- 1
Assume your winning loop will flatten
Whatever growth model works today, optimize and scale it, but assume it will slow and that you'll need a second horizon.
Pro tip Think in S-curves: each lever helps then slows; your job is to find the next curve before the current one flattens.
Watch out If you wait for the slowdown to hit before seeding new loops, recovering is nearly impossible because you need revenue now and new loops take months to fire.
- 2
Allocate 20-25% of annual growth time to new loops
Reserve a fifth to a quarter of the growth team's time annually (not every sprint or quarter) to introduce a new growth loop, channel, or engine, knowing much of it will fail.
Pro tip Overlay across the three models: product-led, marketing-led, and sales-led growth.
- 3
Introduce something new every 18 months
On an 18-month cadence, introduce a new loop/channel/tactic; on a ~5-year horizon, expect a genuinely new engine to be taking a big portion of your volume.
- 4
Delay metric expectations until the loop matures
Let a new loop run as a test before putting metric expectations on it, so it has room to evolve into an acquisition or engagement engine.
Watch out Goaling a nascent loop on monetization or acquisition too early cuts it off at the knees before it can become something.
In the wild
Miroverse, Miro's user-generated library of community-created boards, ran roughly 18 months as a test with no metric expectations. It was used by many people while the team figured out how it fit together.
→ It eventually took off as both an engagement engine and an acquisition engine, precisely because it was given room to evolve before being goaled.
Common mistakes
Over-focusing on the one big win
When a team hits a big result they keep squeezing the same lever past the point where there's juice left, instead of moving on to seed the next loop.
Is it for you?
Best for
Growth leaders whose current model is working and who need to avoid a future gap when it slows
Not ideal for
Pre-PMF startups without a single working growth loop yet
From the transcript
“I try to focus 20% 20 to 25% of growth team's time annually”
“every 18 months, you need to introduce something new in order for it to continue evolving”
“Most of growth loops spin out of their ability to produce meaningful results for you within the first five to six to seven years”
“it took us probably 18 months until we started putting metrics expectations on it”
From the episode
10 growth tactics that never work
Elena Verna (Amplitude, Miro, Dropbox, SurveyMonkey)