We are generating more data than ever before - and success and failure in business could be determined by which companies around the globe use that data in the smartest way, as Liat Clarke explains
By this time tomorrow, an extra 2.5 quintillion bytes of data will have been created. That"s a perfectly impossible volume to grasp. So if you need a mental picture, in one year that amounts to a stack of data-filled iPads that reaches the moon - and it"s a volume that"s expected to increase tenfold by 2020.
The build up to this volume of data has been extraordinarily brief. Ninety per cent of all today"s data has been generated in the last couple of years. This is partly down to the realisation of Moore"s Law, which dictates that computing power will double every two years as the cost of semiconductors plummets. This has led to the proliferation of mobile devices, along with sensors that are fitted into our cars, wearables and home appliances, which are all rapidly generating data of their own.
In most cases, that data is unstructured - taking the form of emails, photos, audio, social media content and much more - meaning it doesn"t fit neatly into rows and fields, like structured financial data. So it"s not surprising, then, that although we"ve all heard of big data, its value remains elusive to many. Indeed, at the time of a 2013 Gartner report, less than eight per cent of companies were deploying big data analytics.
Processing the data
But that doesn"t have to be the case. Just as the cost of electronics has plummeted, thus increasing their reach, so has the cost of processing and storing that data, thanks to cloud computing. For companies, that means realising the advantages of deploying customer relationship management (CRM) systems, fuelled by big data and complex algorithms, to help employees make novel connections, insights and decisions in real time.
As Chris Clark, Managing Director at consultancy firm Prosperity 24.7, explains: “It"s about analysing thousands of transactions and minuscule correlations, and tying that back to behaviours.”
For the JTC Group, that meant transforming its CRM from a "giant address book" containing thousands of client details, into a tool for driving global expansion.
“With big data analytics we could see how far business operations had moved and forecast how likely deals would progress,” says David Vieira, Group Head of Business Development and Marketing. “We tracked stats on a monthly basis to see the exact volume and rate of enquiries.” The improved CRM was stepped up just two years ago, and as a result has helped JTC go from having offices in five jurisdictions to 18.
As the group expanded into Africa and Latin America, analytics helped them target the right clients in the right way. In southern Africa, for instance, it showed that clients were accessing the company website on mobile devices - e-marketing campaigns were tailored for the platform and a mobile responsive site launched.
Making the connection
Social media is also increasingly becoming a part of building client profiles. JTC integrates LinkedIn into its CRM, which then flags up when a client has changed jobs. “We"ll get in touch to congratulate them and set up another meeting,” says Vieira. Other firms might just lose contact, receiving bounce-back emails from the client"s old address.
JTC also uses behaviour tracking to see which products clients are clicking on through their site. Although this could be perceived as intrusive, Vieira notes it"s simply a matter of “taking real world systems online” - if you met someone at a conference, say, you"d make a mental note of which brochure they picked up.
Client managers just need to adopt an indirect approach, recommends Clark. “Phone up and mention a particular client service to strike a chord, don"t go straight in and suggest the specific product they had looked at.”
Telecommunications group Sure has been integrating research from customer focus groups and public comments on Facebook and Twitter into its CRM. “It"s about finding out how this data relates to the analysis we"ve already done,” says Gavin Price, Head of Commercial Integration. “We can then start tagging the data to automatically segment customers [according to preferences].”
If companies manage to avoid unnerving clients by uncannily pre-empting their needs, the advantages will be huge. Between e-marketing, cold calling and snail mail junk, the public is, unsurprisingly, a little jaded. A genuinely helpful call or a bespoke e-marketing campaign means clients will be more likely to pick up the phone or divert emails from junk.
However, all this relies on having clean and relevant data to begin with. “Companies have massive amounts of data collected over a number of years, and information from 10 years ago might be patchy,” says David Carney, a Senior Manager in risk assurance at PwC. Many organisations don"t even have a handle on what data they hold, he says, and start building CRM systems without deciding what questions need to be answered.
If data"s not clean and reliable, and a cross-departmental strategy isn"t decided upon, big data won"t be the solution - it will be the start of a complicated mess. “It"s about stepping back to ask yourself, "is it all relevant?"” says Carney.
It"s why all the analytics are done in-house at Sure. “We had pockets of skills, in terms of Excel resources, and we could build in dependent variables that made it meaningful to us,” says Gavin Price
Data protection
Keeping employees in the loop is also key to protecting client data - you can"t provide the proper privacy and security if you don"t know who is handling it and how. Every employee must be aware of data protection regulations, says Price, who notes that most breaches are attributable to human error. “The tech side of things will do as it"s told - it"s about educating people.”
According to Europe"s Data Protection Directive and the Article 29 Working Party - made up of regulators from all member states - customer consent is needed when “tracking and profiling for purposes of direct marketing, behavioural advertising, data brokering, location-based advertising or tracking-based digital market research”. But as Karen McCullagh, Law Lecturer at the University of East Anglia, points out, company terms and conditions are rarely straightforward, and so could prove problematic.
The EU Directive also states that data processing and collection should be proportionate to the original use. “However, big data analytics often involves repurposing of data,” says McCullagh.
Vieira says JTC sticks to a policy of transparency to make sure it"s in line with the directive, while Price argues that all of Sure"s data is anonymised. However, anonymised data can get personal - and quickly.
“Organisations tend to claim that data is not personal because it"s held in an anonymised or statistical format,” says McCullagh. ”However, given the volume processed it can lead to surprisingly accurate profiling.”
Jennifer Golbeck, Director of the Human-Computer Interaction Lab at the University of Maryland, recently revealed how her team predicted how smart a person was based simply on whether or not they "liked" a curly fries Facebook page. This was down to a network effect - smart people tend to know other smart people, and the page founder was smart. However, other elements as seemingly trivial as "liking" curly fries were also shown to help accurately predict political preference, gender, religion and age.
When everyone from Barclays to the NHS says it wants to sell anonymised customer data, the dangers to the public appear to be growing. Carney notes: “If you haven"t logged in, but a company already has information on you and how to track habits, how many clicks does it take to figure out it"s you?”
Change is afoot, however. Pending changes to the EU Directive will force companies to be more responsible when it comes to profiling - both in terms of excessive collection and getting informed consent. As a consequence, they will have to provide customers with access to their profile, as well as information on how it was built.
Companies need to be prepared for this. As the public becomes more aware of what analysts are doing, they will want to know their data is being safeguarded. Ultimately, these adherences will be worth it because, used correctly, big data can help any company grow. A report published in the Harvard Business Review in 2012 showed those in the top third of their industry were six per cent more profitable than competitors when they made data-driven decisions.
At the end of the day, though, to have a successful experience with big data, companies really just need to continue doing what they should already be doing - listening to clients, keeping interdepartmental communications open and asking the right questions. The analytics will just help the answers come faster. And with the world"s total data output expected to reach 40 trillion gigabytes by 2020, that help couldn"t come soon enough.
The robot uprising
As more systems become automated and we rely increasingly on analytics to make our decisions for us, employee numbers are likely to be culled. The signs are clear across multiple industries.
Thanks to an algorithm that draws in data from multiple sources - including the US Geological Survey - the Los Angeles Times published a news article about an earthquake earlier this year, minutes after it happened.
Text-mining could make paralegals redundant, and image processing software could do the same for lab technicians. But in most of these cases, it"s likely to be a mistake.
A great example of the inherent problem that comes with letting a robot do a human"s job came when high frequency trading algorithms caused the stock market to temporarily crash last year. Why? Associated Press"s Twitter account was hacked and a tweet posted stating that the White House had been bombed and the US president injured.
And "misread" data can go a very long way to seriously damaging customer relations. When an angry father marched into a Target store in the US furious because the retail branch had been automatically sending his teenage daughter vouchers for cots, CRM was to blame. His daughter had been buying unscented lotion and supplements - classic signs, according to Target"s "pregnancy predictor" analytics, that she was approaching her second trimester. It turned out the system was right, but that didn"t matter much to the furious father and his mortified daughter.
Despite abiding by US privacy laws, Target - and the statistician that implemented it eight years prior, adding $23bn in growth to the company - had managed to totally overstep what is acceptable in customer relations, proving that behind every good analytics system, there needs to be an even better human team. A cautionary tale well worth bearing in mind as manipulating big data becomes common practice.