And how every single one of us produces data worth a couple of hundred bucks
Andreas Bayerl & Ines Sophie Scholtes
There is common agreement on the fact that data is the oil of the 21st century (The Economist 2017; New York Times 2018; WIRED 2019) and there is no doubt that “data has great value to institutions” (Li et al. 2017, p. 79; see also Wixom and Ross 2017, p. 10; OECD 2013, p. 16; Spiekermann et al. 2015, p. 161). Data can be used to enhance marketing capabilities, lower advertising costs, improve customer relationship management (CRM), allocate budget more efficiently, predict trends, increase profits through price discrimination (Acquisti 2010, pp. 13f), or build competitive advantages and entry barriers by locking in customers (Spiekermann et al. 2015, p. 161). To be precise, not data by itself has this power, but rather information extracted from data which then leads to knowledge and ultimately wisdom (Chaffey and White 2010) – a hierarchy as depicted by Rowley (2007) in the “DIKW pyramid”. But can the economic value of data actually be quantified?
When researching about this question, one quickly realizes that the enquiry about the value of something is not new: Some argue it can even be traced back to the old Greeks and Aristotle (Johnson 1939). Without a doubt, Adam Smith, David Ricardo and Karl Marx put a lot of thought into it in their respective works. A common theme in their elaborations is a paradox now termed the Diamond-Water-Paradox. The Diamond-Water-Paradox first gained attention as part of Adam Smith’s Wealth of Nations and highlights how the value from using a product, which is high for water and low for diamonds, may greatly differ from the price received or paid in return for a product, which is low for water and high for diamonds (White 2002, p. 662).
“Wealth of Nations”
“Principles of Political
Economy and Taxation”
Heckman et al. (2015, pp. 3f) point out that existing approaches, also newer ones for the evaluation of intangible goods with cost-based and market-based methods, are inappropriate for data. Nowadays, managers are generally aware that data is a valuable resource. However, many find it hard to grasp what constitutes its value (Green 2012), because after all, data differs in its value (Faelten 2018) contingent on especially the context (Acquisti, Taylor and Wagman 2016, p.464) and on “potential avenues for which [it] is used” (Heckman et al. 2015, p.7).
In the following we want to give a short overview of some reports and scientific work elaborating on the economic value of data. We structured the literature reviewed in this blog article in two broad literature streams.
Research Stream 1: Firm Value Creation Through Data
McCarthy et al. (2017) demonstrate how publicly available data about customers can be used to evaluate subscription-based businesses in an – according to their opinion – better way than existing Accounting and Marketing literature. While providing valuable first evidence, they did not put a €-label on data or even on individual data points. This is something the following papers managed to do.
Meire, Ballings and van der Poel (2017, p. 27) assess the predictive value of data sourced from commercial data sellers, and websites in terms of its ability to forecast the conversion of a prospect to an actual customer in a B2B setting. They find that Facebook data is most valuable. The table below shows the added monetary value of using data to convert prospects into actual customers. The baseline represents conversion without any data usage and commercial refers to bought data from specialized providers. The numbers refer to Coca Cola Inc. and were calculated based on one million customers. They reflect the improved response rate of prospects when using data to identify them as such. Thus, the monetary added value of each dataset is several billions for the example of Coca Cola Refreshments USA (Meire et al. 2017, p. 32, 34).
Another example is the work done by Wu et al. (2015, p. 739). They assess the economic value of online reviews and find that one review on a Chinese rating platform for restaurants has a value of 7 CNY (~ USD 1*) to each private user and 8.6 CNY (~ USD 1.1) per private user to each restaurant under review (exchange rate as of January 2020). These findings exemplify the value of data stemming from user-generated content, but are also an example of how both consumers and firms can benefit mutually.
In conclusion, measuring the value of data can be extremely complex because “data value is determined by many, rather than one, attribute[s]” and “by the complex interaction of multiple factors” (Yu and Zhang 2017, p. 2).
Research Stream 2: Valuing Data by Assessing Customers’ WTA for giving away data
Another stream of research (e.g. Staiano et al. 2014; Carrascal et al. 2011) takes a completely different approach by investigating the relationship between personal data trading and privacy, trying to put a price tag on the loss of the latter. In this context, customers become an integral part of the supply side, because their evaluation of the trade-off between costs and benefits associated with disclosing their data determines whether a transaction will occur (Acquisti 2010, p. 7), and their willingness-to-accept (WTA) provides another data point to approximate a price range for personal data, as it represents its floor. Several authors have tried to quantify this trade-off between benefits and costs of sharing personal information in the form of a price that consumers are willing to trade their data for. In a study from Carrascal et al. (2013), participants were occasionally asked about their minimum monetary value for which they would sell different categories of personal data while browsing on the Internet. With a reversed price auction, they arrived at a median bid value of 7€ for the browsing history and 25€ for information related to subjects’ offline identity, such as age, gender or financial status. In a similar setup with mobile phone users, Staiano et al. (2014, p. 583) also use reverse f price auctions and find that location is the data category that most participants opt out for, i.e. are not willing to accept any price.
The above-mentioned valuations seem fairly economic. However, there is “evidence that consumers cannot act rationally […] when facing privacy trade-offs” (Acquisti 2010, p. 38). One example of this is the so-called privacy paradox (e.g. Morando, Iemma and Raiteri 2014, p. 3; Acquisti 2010, p. 37). It describes how people value their privacy but at the same time do not want to pay for it and are willing to disclose sensitive information for even small incentives, i.e. “people do not ‘live up to their self-reported privacy preferences’” (Morando, Iemma and Raiteri 2014, p. 3). Moreover, people’s evaluations also depend on the locus of collection and how the data will be used, illustrated by two important findings: People do not like passive data collection (Spiekermann et al. 2015, p. 165) or their data being monetized (Carrascal et al. 2011, p. 197). All in all, “the price at which people are willing to sell their own personal data depends on their perception how the data will be used and on how much they trust the entity receiving the data” (OECD 2013, p. 32).
Federico Zannier, an alumni of New York University, had similar questions to the ones tackled in this blog post. He wanted to know what his personal data is worth to the world. Therefore, he created a Kickstarter project where he offered a recording of his online activities (HTML pages he visited, the position of the mouse pointer, a screenshot of what he was looking at, every 30sec a webcam image of him looking at his computer, his GPS location and a log of the apps he was using).
His plan was to sell his personal digital tracks for 2 US-$ per day over one month. While he guessed he would earn around 500 US-$, he exceeded his target more than fivefold, accumulating 2,733 US-$! This inspired us for our own mixed-method calculation about what the the data we produce and access during a day is worth:
|7 am||You wake up and check your social media profiles. If someone got access to your username and password, your account would retail for around 2.50 US-$ (krebsonsecurity.com). In total, according to howmuch.net, you as a Facebook user are worth 158 US-$ to Mark Zuckerberg (howmuch.net)|
|7:20 am||While you prepare breakfast you listen to your “Wake-Up-Playlist” on Spotify. If someone could have gotten access to your Spotify password, it would sell for around 2.75 US-$ at the black market (Keepersecurity.com). If your preferred music streaming service was iTunes, a hacker could even make up to 8 US-$ (krebsonsecurity.com) from selling it.|
|8:30 am||On your way to work, you stop by at your electronic store to buy some gadgets. The cashier asks you for your Zip Code. An information that Cataline Marketing Corporation would pay up to 40 US-$ if you also reveal your preferred supermarket (Jaising et al. 2008, p. 858).|
|10 am||Naturally, during your work day, you use Google a couple of times. In total, you are worth 182 US-$ to Google, as you sometimes click on one of their ads provided to you (howmuch.net)|
| ||In your lunch break you make a bank transfer. Information on financial transactions has a median value of $15.5 (Carrascal et al. 2013).|
|3 pm||You finish your working day off in a coffee shop with public WiFi, which is a paradise for every hacker. Assuming a hacker could gain access to a password that would walk him into your bank account, he could sell it for around 100 – 1000 US-$ (Bloor 2013).|
|6 pm||At home you receive an email, that asks you to rate your experience at the coffee shop. Such an online review can have a value of around 1 US-$ (Wu et al. 2015). And since you are in the mood for some additional online reviews, you also write one about your experience of last week-end`s hotel stay and you rate your employer. Two reviews which arguably are worth even more than the coffee shop review you started with, because the decision about where to vacation and where to work is a more involved decision.|
|7 pm||At home, on the couch you decide to do some online shopping on Amazon. To them you are worth a staggering 733 US-$. For eBay you would be worth a bit less: 474 US-$. After finishing your online shopping, you decide to check out the career process of some of your high school mates on LinkedIn. Your value to LinkedIn is about 69 US-$ (howmuch.net).|
|10 pm||Before it’s bed time you watch your favorite Netflix series. On the black market your Netflix password would be worth around 3 US-$ (Bloor 2013).|
While this is of course only case-study evidence, the question about what the economic value of data is cannot be finally answered. The properties of data as a product make pricing decisions extremely complicated. Data owners’ costs are correlated with data content, data bits are correlated among each other, both data owners and buyers are heterogeneous by nature and these gaps are amplified by contextual influences. All of these complexity drivers have been found to be influential for pricing, and current models have incorporated some of them, but not all.
For companies right now it is probably the best plan-of-action to try to understand how data can help them enhance marketing capabilities, lower advertising costs, improve CRM, allocate budget more efficiently, predict trends, increase profits through price discrimination, or build competitive advantages and entry barriers by locking in customers. Assessing a company’s performance with respect to the aforementioned points in online field experiments with two different scenarios can lead to results: One test setting with using data and one without using data might be a solution. The surplus from using data can then be labeled with a quantifiable amount of Euros and, depending on the particular context the company was wondering about the economic value of their data, a quantifiable assessment of the economic value of a company’s data is possible.
The future will show if new value theories just like the old ones from Smith, Ricardo and Marx will help to settle the question about the economic value of data. We are excited!