Backward Growth Model
Set a growth target, calculate its customer demands, then choose channels
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 99%
The model starts with the company's current scale and a specific future revenue or customer target. The team converts that destination into the number of retained customers required, incorporating pricing, churn, conversion, and timing assumptions. It then works backward to assign plausible contributions to acquisition channels and determine the activity, investment, or audience each channel needs. The result is not expected to predict reality perfectly; Nagpal noted that Teachable often missed its modeled numbers. Its value is forcing founders to understand what one million, two million, two-times growth, or three-times growth actually demands. Rebuilding the model monthly creates a disciplined conversation between ambition, channel performance, and resource allocation instead of relying on unsupported numbers.
Origin
Nagpal said Teachable built a monthly model, selected the growth it wanted, and worked backward to determine the required channels. Extracted from Marketing Against The Grain.
Core principles
- 01Growth targets must translate into customer math
- 02Channels should serve quantified goals
- 03A useful model can miss while still improving decisions
- 04Two-times and three-times growth require materially different plans
- 05Founders should understand the numbers themselves
How to run it
- 1
Establish the Baseline
Record current recurring revenue, customer count, pricing, retention, and recent growth. Use consistent definitions so the model can be compared month to month.
Pro tip Separate new, retained, expanded, and churned revenue.
Watch out A vague or inflated baseline corrupts every downstream calculation.
- 2
Set the Destination
Choose a specific customer or revenue objective for a defined month or year. Make the target ambitious but explicit enough to calculate.
Pro tip Model more than one scenario when uncertainty is high.
Watch out Do not substitute a slogan such as 'grow fast' for a numeric destination.
- 3
Convert Growth Into Customer Math
Calculate how many customers, conversions, leads, or audience members the target requires after accounting for churn and pricing. Show every major assumption.
Pro tip Use ranges for uncertain conversion rates.
Watch out Ignoring churn will understate the acquisition requirement.
- 4
Work Backward to Channels
Assign realistic contributions to existing and experimental channels. Determine the volume and resources each channel would need to supply its share.
Pro tip Use actual historical channel performance wherever available.
Watch out Do not make the spreadsheet balance by assigning impossible output to an unproven channel.
- 5
Operate and Reforecast
Compare actual performance with the model each month, diagnose the gaps, and update assumptions or channel investments. Preserve prior forecasts so learning remains visible.
Pro tip Focus reviews on why assumptions changed, not on defending the original forecast.
Watch out Missing the model is not permission to stop modeling.
In the wild
Teachable selected a desired monthly growth rate, calculated what different revenue milestones meant, and worked backward to identify the channels needed. The company frequently missed the exact target, but leadership retained a concrete understanding of the difference between alternative growth paths.
→ The model improved channel planning and founder numeracy despite imperfect forecasts.
Common mistakes
Pulling Numbers Out of the Air
A target without customer and channel math is an aspiration rather than an operating plan.
Treating the Forecast as Certainty
The model organizes assumptions; it does not eliminate market uncertainty or guarantee the result.
Ignoring the Acquisition Gap
If modeled channels cannot produce enough customers, the team must change the target, economics, or channel strategy rather than hide the gap.
Is it for you?
Best for
It is best for startup founders and go-to-market teams deciding how much growth they need and which channels could realistically produce it.
Not ideal for
It is not ideal as a precise forecast in a market with no usable assumptions, although it can still expose what must be learned.
From the transcript
“Every single month we had a model.”
“We tried to figure out how much we wanted to grow every month and then work backwards to find our channels.”
“And we usually didn't hit those numbers, but we always had a very good idea of what it meant to be a million in a…”
From the episode
Why This Startup Founder Is Spending $0 On Paid Marketing with Ankur Nagpal (#104)
Ankur Nagpal