It’s been said that in a digital century data is the new gold. On the surface data and big data analytics promise much. The combination of:
- Vast data generation – Eric Schmidt, Chairman of Google noted that over five exabytes of data is created every 48 hours (that is more data than was generated in all of human history up to 2003)
- Exponential increases in processing power as Moore’s Law continues unabated; and
- dictum “what gets measured, gets managed”
has led many companies to invest in data analytics initiatives.
There is no question that data, correctly leveraged can deliver huge benefits. Over 87% of company value today is now in intangible assets and data, alongside brand, software code and confidential information, are critical to everyday business. Your customer list? That’s data. Your inventory management system? That’s data. In fact many of a company’s most basic functions from invoicing to advertising would grind to a halt without data.
Data analytics promises efficient supply chains, better product design and individualised marketing to name but a few opportunities. For example it has been estimated that each Fortune 1000 company could on average generate an additional $65 million annual net income through a 10% increase in data accessibility to its workforce. Whole business models have sprung up around selling data – check out creditkarma.com for example.
In other cases the value of the company *is* the data. We recently completed a sell side mandate for a 30 year old financial services company. The owners had originally been advised the company would sell for no more than 4 times EBTIDA. After analysing the business we realised that the company’s most valuable asset was not even on the balance sheet – it was 30 years of data it had collected. So instead of targeting people who wanted to buy the operating business we targeted people who wanted the data – a completely different set of buyers with much bigger cheque books. The result: the business sold for 32 times EBITDA.
This is not an isolated incident: in 2013 when Caesar’s Palace went into bankruptcy the data (again not on the balance sheet) was conservatively valued at US$1 billion, making it the single most valuable asset in the entire company. When online retailer Kogan’s bought the failed Dick Smith’s customer database and brand there was commentary to the effect that Kogan made 15 times their money in under three months.
It’s unsurprising then that many c-suites and boards have begun to ask “we’ve collected all this data – what are we doing with it? What *could* we do with it? Can we sell it? Are we sitting on a gold mine?”
Before you reach for the shovel to stake your gold rush claim there are some significant issues to be aware of. These divide into three groups:
- technical (what can and can’t be done from a technical perspective)
- reputational (the brand impact of using data in ways that customers may not expect); and
- legal (what can and cannot be done legally)
The starting point is the inconvenient truth that less than 0.5% of data is actually ever analysed and used. As Todd Park, ex CTO for President Obama points out “data by itself is useless. Data is only useful if you apply it”. A key driver here is that frequently data is collected, but not in a useful form. Sometimes the data is in deep silos that make integration impossible or prohibitively expensive, other times even small but persistent inaccuracies can render data sets effectively useless – for example 75% of businesses believe even something as simple as customer contact information is incorrect.
On other occasions the problem is that expected insights just aren’t there. Data Scientist, Maksim Tsvetovat states “there has to be a discernible signal in the noise that you can detect and sometimes there just isn’t one.” Sometimes the problem is not the absence of a signal but the absence of people who can interpret it – CapGemini found that 37% of companies have trouble finding skilled data analysists to make use of their data.
Perhaps the most significant technical challenge is Recency Bias, a cognitive bias that gives greater weight to more recent events. Given that 90% of the world’s data was created in the last few years this is a serious problem because new data tends to unquestioningly replace older data even when they are equivalent or older data is superior. This is an extremely difficult phenomenon to overcome because it is essentially hardwired into the human brain and affects everything from sports casting “30 greatest plays of all time” to hedge fund performance.
Businesses face significant reputational damage if customers (consumer or business) discover that their information is being used in ways they never anticipated or approved. In the rush to exploit the new data gold the potential benefits need to be carefully weighed against risks to the broader customer relationship.
Straying dangerously close is an entire new $24 billion industry: selling data. Many telco’s facing diminishing subscriber growth are looking to exploit the tremendously rich stream of data they receive as their customers surf, text and call throughout their daily lives. For example SAP’s Consumer Insight 365 business receives up to 300 cellphone events from as many 25 million mobile subscribers *per day* which it then on “sells” to various companies. SAP won’t disclose the carriers providing this data. Similar initiatives are being driven by other major data accumulators such as banks, utilities and insurance companies.
There are very significant reputational risks here: not the least of which are whether customers have actually given their consent to these kind of practices. With the advent of social medical brand velocity (the speed at which brands are built and destroyed) has increased dramatically in the last decade. It takes only small but highly publicised mistakes to do significant brand damage: witness when Target’s big data practices sent coupons for pregnancy related products to a teenage girl who had not yet told her parents about the impending arrival.
Next Up: Part 2: The Legal Risks Around Big Data