Why 2017 will be a Big Year for Big Data
Hold on – surely big data’s been a thing for ages? Well, if your company is in the business of making informed decisions rather than implementing the HIPPO* you’ll have been poring over data sets since before Wintel roamed the earth.
At Fospha, we reckon a number of trends that our team has been working on for a while now will become technology’s equivalent of popular culture in the months ahead. Here, Head of Solutions Rob Williams and Data Scientist Sepi Pouryahya give you the low down on the Seven Key Technology Trends that we think will come to the surface in 2017.
1. The rise of the machines
“Machine Learning and Artificial Intelligence have been buzzwords in tech for a while,” says Rob. “But we’re now beginning to see so many stories in the news – your fridge stocking itself, your car finding its own way home – that these concepts are going to break out beyond techies and marketers and start appearing further up boardroom agendas: they’re going to be asking, what are we doing with ML and AI?”
And the answer to that is…ML and AI are not some silver bullet.
Sepi explains: “Machines are clever at repetitive tasks but what they don’t have is common sense. You have to feed them meaningful data to drive clever algorithms.
“So, what businesses need to be looking at first is the value proposition of big customer databases, because if you want to leverage ML and AI you need that granular level of detail. It is the raw material that informs personalisation.”
2.This time it’s personal…
Personalisation? Well, it might sound obvious, but…
“Yes, it’s about learning what people’s preferences are and acting accordingly,” says Sepi. “But you also have to train personalisation processes. That involves examining the negatives, too: you need to learn the things that people dislike as much as the things they do like because both are critical factors in decisions. So there has to be a lot of trial, error and experimentation if you’re going to deliver the right kind of experience.”
3. A big sea of data
So you’ve got loads of records which show who your customers are and which device they accessed your platform through? Congratulations – you’ve got a contact list…
“People use the expression ‘Big Data’ like a generic phrase but big in this context can be measured in different ways,” says Rob. “Some might see it as large numbers of rows of search data. But what’s in that data? If all you know is where your customers live and what device they use you don’t know much about them at all.”
“In the age of ML and AI, meaningful Big Data shows features about people, not the numbers of them,” Sepi adds. “It’s better to know a lot about 10,000 customers than it is a little about a million. If you want to build real personalisation into the experience, then you need to be able pull in data from different sources – an address from one system, browsing data from another – because that is the kind of granular data that ML thrives on. It’s what defines personalisation.”
4. Sherlock Holmes and the true cost of acquisition
“This is also going to be the year when people start re-thinking the cost of customer acquisition,” says Rob.
“The trend is to focus on channel cost, with lots of clients talking about how much they get from comparison websites, for example. But they’re not looking at all the manpower and other operational costs going into the marketing, all of which has to be absorbed by the customer.
“So, 2017 is also going to be the year when we start using big data to measure the true cost of acquisition, putting emphasis on things like lifetime value rather than simple ‘cost divided by customer’ equations. It’s a bit of a detective story, really.
“When you look at customers who’ve been with you for a few years are they really premium customers? Or is it years since they delivered any real value, despite all the marketing you’re firing at them? It’s about investing in processes which really help you to understand your customers beyond the sums they spend. You need the granular detail to do that.”
5. Fantastic journey
Buying stuff online is usually a pretty smooth journey. So how come wandering from clicks into bricks is like travelling back in time?
“The concept of the seamless journey from online to offline for the consumer is another important trend,” explains Sepi. “When you’re shopping online you’re routinely generating data about you and your individual preferences. That data needs to follow you into the bricks and mortar store, so that we get away from a situation where the offline experience is impersonal and disappointing to one where it’s personalised and fantastic.
“This is about the ability to stitch different sets of data from different sources together. The technology certainly exists to deploy it – wouldn’t it be smart if your phone could tell you where stuff is, if in-store displays showed your related items and reviews relevant to what you’re looking at? That would be a neat trick and data can drive that.”
6. Access all areas
“People are also beginning to realise that they need direct access to data rather than just trusting that some system will sort everything behind the scenes for them,” says Rob.
“If you want the system to have more information, to do ML, to deliver meaningful results then people outside your technical team have to have access to it. And the more access the better – hiding the data away isn’t good enough.”
“Improving access by improving visualisation is important, but people shouldn’t be afraid of looking at the raw data,” adds Sepi. “You don’t want to be in a situation where you’re looking for answers to an important question but you have to ask someone in IT to write the code for a query. You want answers quickly.”
7. Spark life
And talking of everyday access to data, you’ve heard of Spark, right? Come on, get with the program…
“Apache Spark is the new kid on the block for us,” explains Rob. “It’s a data processing platform that has built up a real head of steam and I think everyone’s going to be talking about it in 2017. A lot of companies are using it successfully because it’s great for number-crunching huge data sets.”
Sepi concludes: “This is why 2017 is going to be a big year for big data, and we can see its applications trickling down to SME level. Everyone’s hiring data scientists now – it’s getting to the point where you won’t want to open a food stall in Hammersmith market without one!”
*Surely you know what the HIPPO is? You don’t even need to be Sir David Attenborough to realise that it is, of course, the Highest Paid Person’s Opinion!