It is an interesting question because LTV is really tough to get “right”. What I’ve learned is that your goal should be to estimate LTV to the extent that it’s useful to you, and gives you actionable insights. If you can’t get to that, it’s probably not worth estimating.
The first approach as software companies calculate their LTV using the following formulas:
- ARPA = Average Revenue Per Account
- Gross Margin = the difference between Revenue and Cost of Goods Sold (COGS)
- Customer Churn Rate = Number of customers who churned in the period / Total number of customers at start of the period
The second approach considers account expansion. Account expansion refers to any increase in recurring revenue after the initial purchase, usually occurring from a plan upgrade.
This addition, proposed by David Skok accounts for some basic and consistent revenue expansion. Here, mm is the monthly growth in ARPA per account. We’ve also replaced ARPA in the initial section of the formula with ASP (Average Sale Price) – the average initial price (in MRR) that customers pay at the time of conversion.
From a company health perspective and what Andreessen Horowitz expects to see for a good SaaS business LTV/CAC>3 is very good.
Where, CAC- Customer Acquisition Cost, SME – Sum of all Sales & Marketing expenses, NCA – No of new Customers Added.
To increase LTV you should reduce Cust.ChurnRate and increase m (the monthly growth). I’d also invite you to check out the Simple and effective predictive analytics for your SaaS application and see how it can help you to increase LTV by reducing churn rate and enhancing the recurring revenue after the initial purchase mm.
As you quite rightly pointed out, things become complex when you have a number of payment plans, each with a different billing cycle.
Clearly, the LTV for customers on a 1 month plan is likely to be different than those on a 36 month plan. Those on a 36 month plan don’t even have the option to churn until 36 months after they buy!
Don’t estimate LTV across all of these different plans. They vary so much that the overall lifetime value isn’t going to be a number that’s useful to you. If you really wanted to do this you could multiply up all of the churn rate and ARPA values to the same sized period and produce a single LTV estimate. But it’s really not worth it.
What can be useful however, is measuring the LTV for each plan. With this, you can:
- Make decisions about how much you spend to acquire customers for each plan, based on the LTV
- Look at your pricing and assess whether it makes sense to offer as many pricing plans as you do.
I hope these can be helpful for you!
See case Study: Starbuck