For years I’ve been trying to get people and businesses to think better with data – to come up with new ways of solving problems. We finally seem to be moving in that direction.
Prompted by the release of the Productivity Commission’s report into Data Availability and Use, which essentially found that we need to invest in and support opening up access to data, Australian business and governments are finally beginning to talk about the impact of data on our economy.
It’s good timing, too. A recent report from The Economist made the claim that as oil powered the 20th century, so too will data propel the 21st – and the numbers certainly look that way.
Google, Amazon, Facebook, Microsoft and Apple make up the five most valuable listed firms in the world. And as a separate report from The Economist also points out, much of the reason that’s true is because of data those companies control.
Facebook’s information on your habits sold to advertisers, Uber’s data on traffic supply and demand – these are the real golden nuggets. And if these companies’ expenditure is anything to go by, they think so too: Amazon, Google parent Alphabet and Microsoft spent $23 billion in 2016 in capital expenditure in leases to help house all this information.
That’s not to mention the $26 billion Microsoft spent acquiring professional social networking site LinkedIn last year, which according to Salesforce Chairman and CEO Marc Benioff claimed was a directed move to control LinkedIn’s vast swathes of data.
It certainly looks like an economy of data is starting to rise. But there’s just one problem:
We don’t have a unified, comprehensive way to value that data. We’re just making educated guesses.
All the things we can do with data – better target customers for leads, provide messages to people need it, and even find cures for disease or provide better hospital care – can’t prosper in that new, data-driven economy without proper valuation.
And because we can’t value data, we’re faced with three key problems:
- Businesses sitting on top of massive databases are undervaluing themselves
- Revenue models based on the use of data aren’t being implemented (this includes not-for-profit organisations)
- We’re missing out on the extra demand that attributing value to data could create
So why hasn’t this happened yet? Why don’t have we standard ways of valuing data?
After spending more than 10 years working in and alongside big business, I think there are two main reasons.
The first is that most businesses, although they increasingly view data as an asset, see it mainly as a back-office issue. Something to be maintained, like a tool.
The second issue is that businesses simply don’t know how to value data – no general accounting principles exist to put data on the books.
That second point is especially tricky. Data is unique, its value differs for whatever entity access it or maintains it, and as a result creating a valuation figure is hard. Even more unusual, is the realisation that the value of data doesn’t decline with use, and is usually variable.
For instance: information held by a retailer on a customer may be worth a fixed amount in relation to what that retailer can extract from the information. To another business in a different field, that same data could be worth exponentially more– especially when combined with other data sets.
This is one economic basis for exchanging data on a marketplace, like the one the team and I at Data Republic are building. Unlike other assets like iron or coal that have ‘instrinsic’ or transparent value, there isn’t a single set of unique, accounting standards or methods by which data can be valued from an economic perspective and traded. This obfuscates its value.
The good news is that this isn’t a new problem, and nor is it impossible – otherwise I wouldn’t be giving Data Republic everything I’ve got to succeed!
The first step in the journey towards bringing data onto the balance sheet is for business to evaluate and treat data as an asset, much like their brand.
So, what attributes do you need to consider when attempting to value data?
Business value: This, the most important method, creates tension between the buyer and the seller – with prices usually starting higher than they would others. However, this isn’t usually a transparent method.
Incisiveness: Data that can provide insight is clearly valuable, although this comes with caveats. As a marketer, data that identifies the 1% of the market I need to convert into sales clearly saves me time and money. But the timeliness of this data is crucial, and it needs to be constantly updated.
Uniqueness: In an age of data sharing, unique data will become more valuable. But keep in mind, data held by one entity can be identified by collecting other data from multiple sources. True value comes from the ready application of this ‘unique data’.
Company size: As businesses grow, they generally pay less per item of data – but this may not be sustainable. Granularity: Are we learning about people based on a “yes” or “no” question, or is that data more detailed?
Cost context: The cost to collect data, maintain and protect it should be factored into the overall value. This cost will drop as machine learning techniques mean businesses can use data in cheaper ways.
Surety: How sure are you that this data is reliable or covers the market you want?
Refresh frequency: How often is the data updated?
Contact permission: Can the owners of the data be contacted or not, and how much more is it worth to contact them based on the method they’ve allowed?
Collection cost: How much does it cost to collect that data? (Keep in mind that data can be a revenue stream that can offset cost over time – however, without a marketplace, the cost of actually selling data can often negate revenue or the cost of collection.)
Risk: Risk is the final, and perhaps most important, metric to consider. What risk does a vendor have in providing data, and what risk lies in protecting it from intrusion?
There’s something else to think about: what risk lies in actually executing the data?
After all, data is only valuable once it’s used. (This is why personal data needs to be appropriately separated and maintained in silos to reduce the risk of intrusion – another topic for another time!)
The context of any data’s use should be the primary model for how data is valued, not the implied or assumed value. This also leads to another question: what is the permitted use of any data set, and how is it regulated? Knowing the borders of that permitted use, and the risk of stepping outside it, is a crucial factor in determining monetary value.
Finally, what’s the “reasonable expectation” users have of the data?
You might expect a business to use customer data for advertising on social media, but it’s not reasonable to expect a phone call will be made to that same customer. That unique combination of expectations influences value, and establishing these boundaries is crucial before attempting any sort of valuation process.
So where does that leave us?
It’s clear that figuring out how to value information is a complex beast, and one that requires controlled experimentation from organisations to explore both what the market can bear, the usefulness of this data when applied and how risk can be managed without missing opportunity.
The first step in this process however is for organisations to shift their thinking.
Data is not back-office information on a page. It’s an asset whose value needs to be tested in the market.
Economic activity is derived from value. If we don’t value data appropriately then we don’t secure it properly or trade it in any meaningful way. This hinders the ability for data to flow through our economy and fulfill its productivity potential.
As the Productivity Commission rightly put it; “The potential value of data is tremendous; as is the scope for Australia to forgo much of this value under the misconception that denial of access minimises risks.”
It’s time we got started.