Better your conversion rate optimization and experimentation strategies.
LTV = (Average revenue per account x Difference between revenue and cost of goods sold) / Customer churn rate.
Adjust this formula to account for other complexities such as segmentation, and cohorts, as your retention experimentation program matures.
For example, you can also include MRR fluctuations, non-linear churn, and enterprise customers. Ultimately, the formula will depend on your business model, and may be difficult to define in a single pass.
Choose the right retention metrics by understanding what best predicts the long-term success of your customers.
For example, for Facebook, this is users who add 7 friends in 10 days. For acquisition-level experimentation, example metrics might be: increase the number of visitors converting to email captures, increase the number of leads converting to customers, and so on.
These metrics will change as your data and retention program matures. For example, you might begin to rely on propensity modeling.
Use the customer lifecycle to loosely define your specific conversion points. For example, adoption to expansion.
For example, customers who bought product X, customers who have hired a partner, customers who have completed milestone Y but not milestone Z, customers who pay $XX/month in third-party fees, and so on.
Set a definition for your high-value customers, so you have a goal on how to turn more low-value customers into high-value.
For example, a high-value customer might be in the 80th percentile of MMR, or the customer on the most expensive plan.
Get a company-wide consensus on the definitions of key customer states, and ensure you track movement between those states accurately.
For example, high-value customers, low-value customers, and at-risk customers.
- Monitor and prevent experiments conflicting with one another, as this can skew results.
- Record which acquisition level experiment customers are assigned to at the funnel top.
- Rerun experiments to verify initial findings, and check for degradation.
- Use balanced metrics.
- Record and monitor incrementally.
Last edited by @hesh_fekry 2023-11-14T16:10:42Z