LTV-CPI=Disaster

I’m on day four of Casual Connect, close to returning home.  It’s a gaming conference that’s been a great way to meet up with colleagues and partners in the part of the gaming market that Disruptor Beam inhabits, but it’s been marred by a piece of “received wisdom” that has become so commonplace and accepted—while being so wrong—that it threatens to sink a lot of game companies.

Lifetime Value Squad
(Photo credit: Flickr: http://www.flickr.com/photos/vogelium/3911033827/ )

The mantra is the advice that one can simply take Lifetime Value of a customer (LTV) and subtract the cost-per-install (CPI, the amount paid to a third-party to acquire the customer—usually some third-party advertising network), and that when LTV > CPI, you’ve got a perfect business model.

Here’s the reasons why this is such a lethal and self-destructive way to look at your business:

1) Most people are looking at LTV as “total revenue” that will come from a customer.  Of course, total revenue is precisely the worst way to define LTV.  You’ve got to subtract all of the costs of delivering service to your customer as well.

2) Let’s say you were smart enough to accurately predict costs, and represent LTV as more of a “net cashflow” figure.  That’s good, but you’re still missing a critical aspect: when thinking about LTV it’s really about the present value of all your future net cashflows from that customer.  You get present value by discounting all those future cashflows.  In other words, if it takes you a year to hit your LTV for a customer, you’ve got to account for the time value of money; waiting for the money to come in is a very real cost to your business, and should also be accounted for in your calculations.

3) OK, let’s say you’ve really crunched the numbers and figured out LTV that properly discounts all possible costs, including the cost of waiting to be paid your LTV.  There’s still one absolutely essential aspect that most people are missing.  In many cases, you’ll be paying a CPI for a customer who might have eventually become a customer anyway.  In this case, you’re effectively paying CPI simply to accelerate the cashflow from certain customers.  The better the game is, and the better your organic channels for customer acquisition, the more you’re paying to make things happen sooner.  A good thought experiment to ask yourself is: how many of the customers am I paying a CPI for that I may have eventually gotten to become customers anyway?  And how much am I willing to pay to get them sooner? For example, would I pay $1 CPI today to get a customer who would have signed up organically (at $0 CPI) tomorrow? If the answer is no, and 24 hours is worth being patient for, then just how much time is right for you to wait? If you're willing to accept that a certain amount of your CPI is going to be wasted (by buying eventual customers sooner than you'd have wanted), then that might be OK, but are you really factoring that in as a cost--which it most certainly is?

Now, all this isn’t to say that advertising and other paid customer acquisition vehicles don’t make sense.  There are plenty of cases where you might be able to pay to acquire a customer who you might not have ever gotten at all, or significantly sooner than if you’d have waited for organic channels to work.  The point is that the CPI you ought to be willing to pay is a lot less than many people are leading you to believe.  I hear a lot of people willing to pay half—or more—of their forecasted LTV towards CPI (and in some exceptions, I’ve met people willing to pay CPI higher than their LTV based on hopes of leveraging “network effects” and increasing “enterprise value” to make up for today’s expenses—a recipe that’s as sure as the dot-bomb of circa 2000 to destroy far more wealth than it can ever create).

What’s the right CPI to pay?  I don’t know, but my guess is that unless you have a team of analysts and a very high confidence level of your LTV, then it ought to be less than half--for most companies, a lot less.  If you think you need to pay more, ask yourself who is telling you that you should go higher—my guess is that you’re hearing it from ad networks, themselves largely an arbitrage business with perpetually-diminishing margins, desperate for any revenues they can extract in their race to consolidation.  Caveat emptor... Or worse, you're hearing it from other game companies who are repeating this formula as if it's a magical incantation, but without really understanding the important nuances.

Paid customer acquisition, used intelligently and conservatively, can help grow a business (what large company of the past hasn’t invested in sales forces and/or marketing programs?)… When over-simplified, is likely to spell disaster for otherwise promising businesses. LTV-CPI should be the last thing to look at after you're sure you've done everything else you can to attract organic, loyal fans to your product.