Interview: Sheridan Hitchens, Auction.com on Customer Lifetime Value as the Cornerstone for Marketing Analytics

Sheridan Hitchens is currently VP of Data Products at Auction.com, the leading online real estate auctioneers in the United States.

He has also held positions at two leading online games companies, Kabam and Playfirst, where he has built out Analytics and Big Data groups from inception. Earlier on in his career, Sheridan worked at Procter and Gamble in the areas of Decision Support and Executive Information Systems. He also worked as a Managing Consultant at Towers Perrin.

Sheridan holds a Bachelor of Arts degree in mathematics from Cambridge University and an MBA from the Haas School of Business at the University of California, Berkeley, both with honors.

Here is my interview with him:

Anmol Rajpruohit: Q1. What are the key advantages of using Customer Lifetime Value (CLV) metric? For whom does this metric holds most promise?


Sheridan Hitchens: I think the key advantage is typically for the marketing department; if you understand lifetime value, particularly across demographics and acquisition channels, you can understand what marketing investments are ROI positive and where you should invest or cut back.


AR: Q2. What do you think about the current maturity level around the Customer Lifetime Value metric? How reliable are the current models?

SH: I find that the maturity levels differ across industry. In consumer focused businesses that are primarily online and do a lot of online acquisition, e.g. the online gaming industry, they can be pretty advanced. However if you’re in an offline industry, say steel production, they may not be very advanced, and quite frankly might not need to be.

The biggest factor in the reliability of online models, is our ability to predict future revenue flows. In a traditional industry such as utilities where customers pay monthly may be much better able to predict than an industry undergoing rapid change, like online media.

AR: Q3. In your talk, you showed the simplistic equation for calculating Customer Lifetime Value (CLV) based on future discounted cash flow. Is there an inherent assumption that no future investment would be required to sustain or improve CLV? Are there any Analytics tools that you would recommend for advanced CLV calculations based on wide range of inputs?

SH: The model looks at future flows of gross margin; if we believe we need to make an ongoing investments then we can either factor those in, or we can consider the new investment in its own right and understand if the change in cash flows ends up being positive from a net present value perspective.

As for tools, I’m sure there are some sophisticated ones out there, and no doubt I’m going to subject myself to a bunch of vendor calls, but most CLV models seem to be done in Excel, and you know that’s often fine.

AR: Q4. What are the biggest challenges in using the Customer Lifetime Value (CLV) metric to design business strategy? What kind of external data can be leveraged to improve the confidence in CLV metric values?

SH: I think the fundamental challenge is the future is hard to predict. Last time I checked, I wasn’t cruising around in my flying car. In strategy planning, it’s often useful to use more abstract analytical tools; I’m personally a big proponent of traditional scenario analysis. You might look at a variety of different scenarios around market growth, competitive pressures, and the regulatory environment that you may face. You can then get external data that helps understand and build these scenarios, and factor those into your models.

AR: Q5. In order to calculate Customer Lifetime Value (CLV), one needs to have the correct attribution model. In today's world of several consumer touch-points and multi-platform online experience, what are your recommendations to develop a reliable attribution model?

SH: Part of the problem of many of the efforts I see in this area is that they often start by asking question like “What is my Customer Lifetime Value?” or “How should I attribute my marketing spend?” These are somewhat abstract questions, and they don’t address the underlying decisions you’re trying to make. It would be better if you started with questions like “Should I invest more money in online marketing, and if so where should I invest it?” Those answers give me practical real world answers to decisions you need to make.

As for attribution models, they’re complex, and to understand how they may be subject to inaccuracies, you only need to consider yourself. If I buy a Gap Sweater, after I saw a billboard on my way to work, got a recommendation from a friend, and clicked on a keyword ad in google, I can’t tell you what percentage each of those three brand interactions contributed to my decision to purchase. Of course you can look at what effects different combinations have and there’s some very interesting work out there on this. One marketing tech company I came across, Abakus uses game theory to do it, which is absolutely fascinating. The bottom line is it’s a very complex problem and there’s no silver bullet.

Second part of the interview

Related:

• Alison Burnham, Scorebig on Optimal, Real-time Pricing through Analytics
• Pallas Horwitz, Blue Shell Games on Why Data Science is So Critical for Gaming Studios
• Hobson Lane, SHARP Labs on How Analytics can Show You “All the Light You Cannot See”