One of the more difficult adjustments to make when transitioning from academia to industry is the change in vernacular required in communicating the machinations of models. In academia, the formatting and notations used generally indicate concepts that are taken for granted by the author and sometimes included only because, at a theoretical level, not including them could subject the author to ridicule.
But in industry, models must be more concretely defined before being introduced. Models in industry must also be practical and account for limitations of the analytics infrastructure and product. In addition, models should also be vetted, using real data, and made transparent and interactive with a software platform that all stakeholders (usually executives and the product teams) are familiar with. This precludes the use of statistical packages like R, Stata, and SPSS and requires output to take a form more instructive than command-line standard error measurements.
This transition pain is evident in the academic treatment of Lifetime Customer Value (LCV). The quantitative measurement of LCV has been explored at length primarily by two academics, Bruce Hardie and Peter Fader, who have seemingly made the process at least an ancillary portion of their research efforts, continuing the quantitative focus on LCV first established by Schmittlein, Morrison, and Colombo in the 1980s. An excellent primer on LCV is available in a paper contributed to by Hardie called Modeling Customer Lifetime Value. For more background on the Pareto/Non-Binomial Distribution method for estimating customer purchases pioneered by Schmittlein, Morrison, and Colombo, this series of blog posts provides a good overview.
Central to the academic definition of LCV is the notion of the time value of money as it relates to present value. The time value of money stipulates that a dollar received today is worth more than a dollar received tomorrow (or at some point in the future) because of the effects of compounding interest. If I receive a dollar today and invest it in an interest-bearing account, that deposit will be worth notionally more than one dollar tomorrow. Thus cash streams are subject to opportunity cost concerns: delaying the receipt of cash imposes an opportunity cost that must be accounted for by some mechanic.
In corporate finance, this mechanic is the Present Value calculation which reduces the value of a stream of payments in the future to a total value today given the opportunity cost (discount rate) of not receiving that money. The assumption that money always incurs an opportunity cost is grounded in the notion of a risk-free rate, or a rate of return on investment that is essentially risk free. In corporate finance, the risk-free rate is often defined as the interest rate on a 3-month US treasury bill, which is, for practical purposes, considered “risk free” given the low likelihood of default by the US government. At a high level, the risk-free rate exists as a permanent “Plan B”: an organization always has the option of taking their money and putting it in risk-free treasury bills as opposed to investing it in some other activity. This means that any project must exhibit a rate of return of more than the risk-free rate to be preferable to simply sitting on risk-free treasury bills.
In large, diverse organizations, such as software conglomerates, the discount rate is a real concern: projects are prioritized and financed based on their estimated Internal Rate of Return. But large, diverse conglomerates generally employ capital markets professionals — sometimes even in in-house groups – to manage their liquid assets. For many conglomerates, investing money into treasury bills (or other financial instruments) is a valid business activity; the preservation of the firm’s cash against inflationary decay, currency shocks, or any other of the myriad concerns facing businesses with appreciable amounts of liquid assets is extremely important.
For firms of this size, raising money from the capital markets is a fairly straightforward process, and the cost of raising that money is easy to quantify (through a model known as the Weighted Average cost of Capital). Thus, for large conglomerates and firms with sizable, diverse product portfolios, raising money and sitting on risk-free treasury bills are fairly commonplace activities that can be priced using historical precedent and transparent, liquid marketplaces. For such firms, a discount rate for a Lifetime Customer Value metric could be easily derived and put into practice, and the value of that discount rate could exert significant influence on resource allocation and project prioritization decisions.
For small, independent mobile gaming studios, this is not the case. For one, independent mobile gaming studios generally can only raise money by selling highly illiquid equity (and sometimes debt). Pricing such instruments is inexact at best and based totally on the demands of the buyer at worst; the price paid for such instruments does not represent a realistic cost of capital in future scenarios. Secondly, independent mobile gaming studios generally operate under a mandate, either from investors or the management team, to focus exclusively on the production of mobile games. Such a mandate precludes any participation in the capital markets using the firm’s liquid assets; the risk-free rate is not relevant to the decisions surrounding resource allocation, as the money dedicated to the development of mobile games can’t be allocated to any other purpose.
But perhaps the most poignant reason for not discounting the LCV metric in mobile gaming is that it introduces an unnecessary complication in the decision-making process that itself could cost the firm money in the form of inaction. Attaching the parameters and vernacular of academia to the LCV model when introduced in a practical setting doesn’t accomplish anything and more than likely obfuscates the framework used to value users. For all but the largest game development companies, dropping the present value mechanic from the LCV model when translating it from academic parlance into a practical, implementable solution provides more benefit than maintaining and explaining it.
Recent