Ah, thanks for the example. I'm used to thinking of costs as development (fixed costs, spread over an extended period before revenue is generated) and maintenance (proportional costs over the same timespan as revenue), but I never thought of customer acquisition as something occupying the "proportional cost, paid up-front" segment.

Not really relevant to tarsnap (I've never found an efficient way to spend money to acquire customers in the short term; instead I focus on longer-term branding) but I can absolutely see that it fits some companies.

Just checked out Tarsnap. Seems like a very good idea.

>Not really relevant to tarsnap (I've never found an efficient way to spend money to acquire customers in the short term; instead I focus on longer-term branding) but I can absolutely see that it fits some companies.

Isn't backup a kind of sticky service? I mean once I have purchased your service, as long as I am not facing any issues I am unlikely to change the provider. So LTV calculations should be relatively more predictable to your business.

For example, I use Dropbox for personal online storage and backup purpose. It has worked well so far and haven't considered switching in years.

Anyways, not trying to tell you how to run your business. Just curious. Yours is really an intersting product. Best of luck.

Yes, I have plenty of LTV per customer; but tarsnap's users skew heavily to the more technical and aren't exactly easy to targets for advertising.

how does one calculate the LTV/NPV of a customer?

The net present value (NPV) of a customer is just the value today of all of all future revenue minus the cost today of all future expenses associated with them. The key is that it factors in a "discount rate" to account for the fact that cash in the future is worth less than cash today. So future cash is discounted by some ℅, compounded annually.

If a customer will pay $100, and it costs $90 to acquire them, without taking time into consideration it would look like $10 in profit. But if you have to spend the $90 today, and you only get the $100 in five years, they might not actually be a profitable customer. Conversely, if that same $100 revenue customer cost $105 to acquire, but they paid today and you only spent $105 to acquire them in five years they might actually be profitable (this scenario is probably less likely, but maybe if you are acquiring them through some channel where the costs are deferred for some reason).

Excel and a lot of other spreadsheets have a built in NPV function to make the calculation easier. You can simply give it the series of expected costs and revenue associated with the customer, and a discount rate (how much to discount future cash by, compounded annually) and it will give you the NPV.

Churn rate gives you average customer lifetime. You know how much they pay over that period, you ought to know your customer acquisition costs & the cost to service your customer base already if you have any kind of business head on at all. From these numbers you can calculate the value of an additional customer to your business. How much of that you’re willing to spend on marketing costs is up to you and your risk tolerance :)