7 in 10 plan on investing in analytics (and machine learning) this year: but why and where?

In today’s society, we are all leaving digital footprints in varying ways (for example on social media, by filling in forms, online or in apps). These footprints offer insights into our behaviour and our likes and dislikes. This information has value to everyone, if only you can work out what it all means.

What I intend on exploring in this blog is the concepts of AI (Artificial Intelligence), machine learning and advanced analytics. I thought given the hype around these at present it would be of use to not only outline what the challenges are, or could be, in adoption, but also what the opportunities look like post adoption.

A fear of the unknown and that a machine might replace a person unnerves many businesses – but it hasn’t deterred nearly every business from intentions to adopt (It’s now universally acknowledged by more than four out of five (82%) CEOs, just how important data, analytics and AI are to business). But at Experian we believe, and always have, that advanced analytics is the only way to extract the value out of data to enable you to make meaningful decisions for your customers and society in general. This, to us, doesn’t mean replacement – but enhancement.

Data in context

Making sense of complex data and turning it into useful insight that can be used to make decisions is not a new challenge. Last year we saw 82% businesses cite this as a problem through our research with Forrester, and this year the results are an exact replica – seeing 82% still state this as a challenge. What has seen change is a worsening customer view – as opposed to an ability to get one which last year was a pertinent challenge.

There are some new, very challenging aspects to the problem. The diversity and complexity of data today and a lack of available tools available means the ability for comprehensive complex problem solving and enrichment often isn’t there.

The fundamental challenge in using any data, particularly large and varied data sets, lies in understanding what has happened (how a person has behaved and what preferences have they shown, for example) and then understanding what will happen (predicting their future behaviour and preferences).

People are becoming more demanding. Part of that is driven by technological advances. Technology today makes things available to us that were not available before.  An obvious example being the adoption of smart devices, where we can order taxis or food simply through the touch of a phone. Technology brings significant convenience to everyday life. As a result, we then want more based on this transition, for example instant shopping, immediate decisions and swift identification. As we embrace these digital tools, we are leaving a trail of data in our wake.

Data generation

Different groups within society are leaving different amounts of data. Take Millennials, for example, – their attitude to data sharing is very different to the Baby Boomer generation. The Baby Boomer generation also referred to as the ‘ageing population’ are much more hesitant according to our research. Sharing digital data is still a relatively new concept for them, as opposed to the Millennial who have grown up with it. Important to note however, according to recent research we commissioned across 2000 consumers the gap of technology usage – including mobile banking – is closing. Three years ago, in the same research just 17% of those 55+ used mobile banking more than once a week, whereas today it is used on average once a day by them. Similar growth is seen in general app and internet engagement.

Cutting across that, we are all, to some degree, both embracing of and sceptical about what can be done with data. Many of us will be happy to consent to use some of our data for certain purposes, but six months later, when it becomes clearer what that data is being used for, or when we no longer have a value for it being used again and again, some people will object to this, even though they consented ‘willingly’ at the time. We need to keep those sorts of things in mind and being able to manage consent is therefore a critical priority.

Clearly, safety and security are an issue. Fraud concerns us all. There are several hurdles which all businesses holding personal data face, particularly around encouraging consumers to share. Transparency is therefore critical.

Challenges of expectations and fragmentation

People’s expectations are changing and people do not want to provide a lot of information to get access to products and services. Data and analytics can smooth the journey in the most frictionless way possible.

For example, data held on a person can be used to prepopulate fields – an attractive time-saver for most people today. Analytics is equally important as a tool to make decisions about whether the person is right to be approached – from a risk perspective but also an appetite to respond.

The solution needs people, technology and analytics 

You need to build a framework that incorporates the technology, and the people, which enables them to work together to conduct the analytics efficiently in a controlled and secure way. It is not about doing some clever mathematics, it is about solving the business problem and both are as essential as the other.

Today, advances in analytics provides the opportunity for more advanced processing and understanding of data. This doesn’t negate the need for people, it simply equips people with the tools that enable more sophistication to their job.

Employing people with the right skills is clearly very important, namely data scientists. Doing this however is a well reported challenge of which research states the skills gap is a significant concern for half of businesses. It is very important to recruit people with broad-based technical skills and experience of machine learning, signal processing and other areas of computer science so you can tackle some of the challenges posed by large scale datasets.

Replacing legacy technology with new technology which is more agile and able to link data and extract the insights that you need, will enable you to meet your strategic goals far more easily.

Solving problems through analytics

In the context of analytics, to get from raw data to something that can be used to make a decision is a well-defined and well-trodden route. You gather your data and then need to process it in a way that summarises it, aggregates it and turns it into something more useful. Instead of a lot of raw, uninterpretable information, you are now looking at how many, how much, how often. This categorisation of the data is the foundation for analytics further down the line.

Next you apply algorithms to solve a problem or answer a query. The range of algorithms now available and suitable for these purposes has increased significantly over the last few years.

The next stage is deployment. Here, for many, internal resourcing constraints and competing projects can be barriers to internal sign-off on deployment.

In addition to skills, right now, many organisations are still typically blighted by siloed-thinking with mixed customer messages and approaches – for instance marketing versus collections. But it’s possible to win a competitive advantage. Predictive analytics can help you consistently offer the most appropriate products and provide early identification of key lifestyle changes – from house moves to purchasing your dream car.

The last part of the process is to identify whether it is working at all. Can you monitor it? Is it doing what you expected it to do? If not, what should you do about it? Then you go back to the beginning, extract some more data and go around that cycle again, learning from the process. That process has been in place for many years but the technology now exists to go around that learning cycle much more quickly than in the past, in seconds for some problems. Or you might go through that cycle in days, months or even years in some cases. It depends on the problem that you are tackling but it can be automated through machine learning.

Analytics is bottom up, not top down

The hype around analytics, AI including machine learning is because of the benefit it can bring. But, instead of jumping on it – investing hard into it (we can see that 71% of businesses intend on investing in analytics over the next 5 years, with 78% increasing budget to allow for this development), you need to start from the bottom and work your way up. Analytics can help you do this. Refine and segment your data and more – but you need to understand the basics of what you need to do, what you want to do and then create a road map that is not only sustainable but considerate of where you want to be and how you will get there. Our view, which is evidence by a breadth of businesses who we have worked with, is that working with an experienced analytics partner will help you develop this opportunity and set you up for much more. Obviously, we would recommend you talk to us to do just this, so please do.

(Note, research contained within this blog is from our research commissioned through Forrester Consulting, November 2018, as well as consumer research via our panel partners C Space – Summer 2018)