Customer growth model
How conversion, LTV, ROAS and churn affect sales
Growth models become misleading when every percentage is treated as the same kind of improvement. These four levers affect different parts of the system and should be tested separately before their combined effect is considered.
Start with a visible customer journey
A funnel is easiest to manage when each stage begins with a real number of people and each movement is controlled by one conversion rate. The example below follows the baseline used in the Growth Simulator.
Awareness
Starting audience
Consideration
1% of awareness
First conversion
20% of consideration
Retention
30% of first conversions
Loyalty
30% of retained customers
Illustrative model: customers can be fractional because the purpose is to compare scenarios consistently, not forecast individual people.
1. Conversion changes volume
A conversion rate controls movement between two stages. Improving an early stage has a compounding effect because every later stage starts with a larger base.
Moving awareness-to-consideration from 1% to 1.2% creates 120 considered customers instead of 100. With every later rate held constant, first conversions rise from 20 to 24 and retained customers rise from 6 to 7.2—a 20% improvement throughout the remaining funnel.
2. LTV changes the value of each customer
Customer lifetime value does not create more converted customers by itself. It estimates how much value an average customer produces over the relationship. Raising LTV by 10% should therefore increase the modeled value of the existing customer base by 10%, assuming customer volume and all other inputs remain unchanged.
In practice, LTV may improve through stronger repeat purchase, frequency, average order value, margin or customer longevity. Those drivers should be checked before using one headline percentage.
3. ROAS changes advertising efficiency
Return on advertising spend compares revenue attributed to advertising with the advertising cost. It is a useful efficiency measure, but it is not the same as profit and it does not prove that advertising caused every attributed sale.
A scenario model can use ROAS to represent how efficiently budget creates revenue or effective reach. The assumption must stay visible: improving ROAS while holding spend constant creates more modeled value; lowering spend while holding revenue constant creates the same ratio but a different growth outcome.
4. Churn changes what remains
Churn measures customer loss over a defined period. A fall from 10% to 8% is a two-percentage-point reduction, but it is a 20% relative reduction in churn. Mixing those two descriptions can overstate or understate the result.
Churn should be applied to the relevant active or retained base—not to the awareness audience. It mainly changes the customers and value that remain available for repeat purchase, loyalty and future revenue.
How to test a growth scenario cleanly
- Set the baseline.Use current reach, stage conversions, LTV, ROAS and churn from the same time period.
- Change one lever.This reveals which part of the result that input actually controls.
- Read the downstream effect.Compare customers, retained customers, loyal customers and sales value—not just the changed rate.
- Combine only realistic improvements.Avoid assuming every lever improves at once without the investment or operational change required.
- Replace assumptions with evidence.Update the model as campaign, sales, CRM and retention data becomes available.
A simulator supports a decision—it does not predict it
The most useful output is not the largest number. It is a clear view of which change produces enough commercial value to justify the work, investment and time required. That is why the model should be used for prioritization and scenario planning before it is treated as a forecast.