A New Customer-focused Model for Stock Market Valuation Could Change the Game for Investors

Michelle Markelz
Current average rating    
Key Takeaways

​What? Traditional corporate valuation models do not adequately account for the influence of customer behavior.

So what? Researchers Peter Fader and Daniel McCarthy developed a model, based on customer data, that more accurately projects company valuation.

Now what? Investors can better know the risks posed by companies if they examine how well a brand retains its customers and how loyal they are.

Aug. 5, 2017

​Who could have forseen Blue Apron's disappointing IPO? Two marketing researchers with a market valuation model that could change the game for investors.

If you’ve listened to a podcast lately, you’ve heard of Blue Apron. The meal kit company appeals to would-be home cooks who are either strapped for time or don’t know where to start when it comes to shopping for dinner ingredients and trying new cuisines.

Blue Apron debuted on the New York Stock Exchange in late July but those paying attention saw it devalue its shares prior to its offering from an optimistic $17 per share to just $10. Though the decision was a rare move for companies approaching IPO, one person, at least, saw it coming.

Daniel McCarthy, an assistant professor of marketing at Emory University, flagged the company as overvalued and staring down a horizon of high churn and low ROI in an article he authored on LinkedIn about a month prior to the IPO. While analysts pointed to the company’s revenue increases and the increasing customer count, McCarthy noted that Blue Apron’s retention curve and customer acquisition cost trends were the real tea leaves, and they showed declining performance at an increasing marketing cost.


Graphs by Daniel McCarthy

McCarthy had good reason to examine these metrics; he’s been working with Wharton professor Peter Fader on a method to accurately value corporations based on publicly disclosed customer data. The two published the idea in the Journal of Marketing earlier this year and presented case studies of its application at the 2017 Summer AMA Conference.

Fader and McCarthy’s method involves running the available historical data through a statistical algorithm to project a firm’s future revenues and “back out” underlying unit economics. They contend that traditional methodologies used to project revenue are less accurate because they only focus on sales, ignoring or greatly misusing publicly disclosed customer data.

Rapid sales growth on its own does not guarantee financial success, he says, particularly if that growth is driven entirely by customer acquisition and is not supported by repeat business after those customers have been acquired.

“The gap in the literature to date is that the underlying models of customer acquisition and retention do not adequately reflect empirical realities associated with these behaviors (e.g., the effects of seasonality and customer tenure), and the valuation model used is not up to the standard of finance professionals,” they write in their article snapshot on ama.org.

Case in point: Blue Apron.

The professors demonstrated that their method, when applied to historical data for Dish Network and Sirius XM, projected stock market prices that were very close to the actual prices that shares sold for—on average, about 10% more accurate than analyst predictions. In one notable case, their modeling approach, when used by students at Wharton, more accurately projected Q1 2017 new customer acquisition for Dish Network than Wall Street.

While Dish Network and Sirius XM are subscription-based businesses—what Fader and McCarthy describe as contractual—the other two companies they applied their method to, Wayfair and Overstock.com, are “non-contractual”, or businesses for whom customer churn is not observed. For various reasons, loyal or retained customers may go long periods of time without making a purchase, while customers who have churned can be difficult or even impossible to identify. This complicates the task of modeling customer retention.

These types of non-subscription businesses are also more susceptible to fuzzy projections because they tend to define customer data more inconsistently than subscription businesses. Overstock and Wayfair both disclose how many “active” customers they have over time. Overstock defines an active customers as those who had made purchases in the last quarter, while Wayfair defines an active customer to be one who has made a purchase within the past year. Fader and McCarthy accounted for all of these challenges and complications in their customer-based valuation model.


 Peter Fader on Customer Centricity


With Overstock.com, their method implied a fair stock price that was within $2 of the actual market valuation. With Wayfair, in contrast, their method predicted a stock price far below the current market valuation ($23 per share, compared to the actual price of $74). Even after accounting for the fact that Wayfair’s future prospects are more uncertain, they nevertheless predicted that more than nine times out of ten, investors would lose money buying Wayfair stock at this price.

They are not alone in their assessment of Wayfair’s valuation. The company is the sixth-most-shorted stock in the internet retail sector, with some analysts calling it a pipe dream of a venture with little chance of reaching profitability.

There are limits to the applicability of these so-called customer-based corporate valuation models, McCarthy acknowledged, most notably businesses whose products or services are purchased less frequently by their nature (e.g., automotive manufacturers) and businesses that haven’t disclosed the right quality and quantity of customer data.


 Sign Up for Marketing News Weekly

Get the best marketing thought leadership delivered directly to your inbox!

Author Bio:

Michelle Markelz
Michelle is managing editor of the AMAs publications. She can be reached at mmarkelz@ama.org.
Add A Comment :

Become a Member
Access our innovative members-only resources and tools to further your marketing practice.