Get with the Times: The Fair Value of “Big Data”

Consider the following eclectic assortment of vignettes:

• A few weeks ago, some colleagues from Mercer Capital attended the demo day for a Memphis-based summer startup accelerator. One of the presenting companies was developing a platform to match consumers who want to swap excess make-up supplies. Interestingly, the founder’s vision for creating value was based on more than the utility created for swappers. The second-side of the business would harvest, process, analyze, and package swap data to be sold to make-up buying clubs, manufacturers and marketers.
• “Data scientist,” a job title or description that was almost non-existent just a few years ago (subscription required) has now become a very important function within young startups across all segments, as well as traditional firms like retailers and banks. Data scientists with two years of experience can earn upwards of $200,000 a year, which is presumably reflective of the high demand for these positions.
• As a consumer, I have received “bespoke” (customized) coupons from Kroger that are based, presumably, on my purchase history with the grocery store. Some of the coupons are astonishingly relevant – they seem to know when I need to buy the next set of razor-blades – but most miss the mark, sometimes hilariously so. If I were a gambling man, however, I would not wager against these coupons becoming more and more useful to the point of being indispensable in the future, especially if Kroger decides to procure data on my interactions outside the four walls of their stores and incorporate them into their coupon-generating algorithms.

A recent WSJ article indicates that transaction-crossing platforms like traditional brick-and-mortar stores or online markets can and do collect data that consumer-product manufacturers are willing to purchase (subscription required) in order to fine-tune their own products and marketing. For instance, the article estimates that Kroger’s revenue from selling this type of data is on the order of $100 million annually. It would be logical to assume that the process of generating some, if not all, of the bespoke coupons mailed by Kroger is influenced by the fine-tuned marketing efforts on the part of the manufacturers.

Beyond traditional businesses, big data underpins the business models of tech companies – ranging from giants like Google and Facebook, and months-old startups – that collect information from their users on the one hand, and use that data to provide tailored advertising (or other) solutions on the other. At the moment, however, users of financial statements are unlikely to find assets that correspond to these datasets (and the resources invested in processing them) on the balance sheets of many companies, as the WSJ article notes:

Kroger does say that it follows generally accepted accounting principles, which prohibit companies from treating data as an asset or counting money spent collecting and analyzing the data as investments instead of costs.

Current accounting treatment stipulates that in-process research and development efforts are expensed as costs but capitalized as assets following business combinations (mergers and acquisitions). The WSJ article suggests that measuring the fair value of a “big data” asset may be difficult because “companies also would have to estimate the shelf-life of their data, figure out its future worth and track and report any changes in its value.” For fair value practitioners, however, these concerns may not seem insurmountable. Indeed, much like customer-related intangible assets, the following elements are likely key in reasonably considering the fair value of big data intangible assets:

• Revenue generated through the sales of the datasets can be a direct input in measuring fair value.
• In the absence of direct data sales, the expectation of enhanced product sales – by increasing customer retention, identifying better product fit, developing methods to maximize sales prices (through differential pricing, for example), or otherwise – can support the value of these assets.
• Historical or projected replacement cycles can be illustrative in determining the period over which these assets lose their ability to enhance product sales.
• As firms would need to assemble a number of other assets for the datasets to be valuable, a mechanism to consider contributory asset charges could also be necessary.

At the end of the day, financial statements need to reflect current economic realities to stay relevant. Tangible physical assets may have been the primary source of value creation in an earlier era, but intangible assets increasingly drive the business models of a newer generation of firms. While a degree of judgment is necessary in measuring their fair value, perhaps it is time to reconsider the need to recognize certain intangible assets on the balance sheets of companies even in the absence of (or prior to) business combinations.

Related Links

• 5 Things to Know About the Draft AICPA Guide on In-Process Research and Development Assets
• Valuation of Customer-Related Assets