When it comes to analytics, data science is front and centre of how a revolution is happening.
The financial services, banking and insurance sectors all capture lots of data points and typically distil it to a single score denoting an individual’s risk. Risk analytics could be said to be about creating a number from people, but data science is about getting to people from numbers.
Traditional data analytics
Analytics is essentially using data points about individuals to better understand them. It can help us, as businesses, to have relevant, meaningful connections with individuals – just like human conversations. But we need to know something to start a conversation and in some cases, we don’t get the information right to even start talking. Why? Because form fills frustrate people and they may type their name as PPPPP rather than Paul. Or because they have got married, and the system hasn’t detected this and updated it. They therefore don’t want to talk to us, or we don’t know how to talk to them.
But it is vital to note that before you use the data available to you to do this, you consider its quality. This point might be self-evident, but as teams and databases merge and migrate and acquisitions within groups of companies occur, data quality can suffer. A solution is to combine your data with third party data (Experian data for example) to create a rich picture of your customers and validate what you believe.
The upshot is that businesses have made assumptions about their customers based on an inside-out view, seeing the customer the way the organisational culture tells them to. Fitting people into a box, into a segment – even if they only loosely belong in it. How can we change this so that we get to know each of our customers, without making assumptions based on this inside-out view? We flip it. We start with an outside-in view. We understand the external drivers of expectations, the drivers of each individual, their patterns, their needs and their wants.
As businesses and custodians of data we are beholden to use our customers’ data wisely and give a real value exchange should they grant us permission. From a customer’s perspective, sharing their data entitles them to expect relevant and engaging interactions. If you get this right, they will be more prepared to reward you with their loyalty and custom. But how do we get to a point where we can be engaging, be relevant and be valuable?
“It is not good enough to segment your customers by how much money they have got and how old they are. If you use that approach, you put the Queen and Mick Jagger together in the same segment.” Tom Spencer, Aviva
Applying analytics to customer problems.
Today there is huge amounts of data available. In the last decade alone for example, the bureau has seen more data; non-traditional and alternative data like rental data and utilities data. There are also more entries from across lenders giving a huge amount of insight that can be used to better understand people and better solve problems and enhance opportunities.
Earlier this year we saw bank account transactional data enter the market, and soon utilities data will be shared, and much more.
Open banking now allows us to add transactional data into the big data pot. Why would you want to add to the data volume and complexity? This type of analysis can truly start to signal an individual’s behaviour, their lifestyles, their needs and their motivations. And this is exciting. High level segments need to be broken down into micro-segments and used in specific use cases, where relevant.
Relevance is key
With any data, you need to ensure you are relevant. Being able to understand a person at an individual level has long been a problem of many businesses. Out of date data means marketing teams communicate with prospects and customers without the personalised content being relevant. Today, people expect that if you know so much about them, you deliver them value. You help them. You make it quick and easy for them to transact with you. Online shopping – specifically those who use recommendations and more, has spurred this desire along. Instant access to taxis without having cash, or single-sign on access means convenience and relevance is key.
For example, you send out an email campaign and there is a 10% open rate – great. But we should look at it another way. That 10% open rate is also a 90% irrelevance rate. You have been ignored by 90%, that’s nine out of ten, of the recipients. And by doing this you have potentially reduced the efficiency of your other marketing. Some of this 90% might remove marketing permissions; they may well unsubscribe.
We need to reach a position where for every individual customer we know which product we should talk to them about next, at what time and through which channel to maximise relevance. Ultimately, we should be applying data science to customer problems to help us improve failing customer journeys and be more relevant when we are interacting with customers.
Combining segmentation with personalisation
Customer segmentation is an age-old method of data analytics and provides some good data analysis to build from. However, combine this with advanced analytics and you can hyper-personalise conversations on a different level. You can understand not only a group based on their similarities and traits – but the person as an individual.
“Done correctly, segmentation combined with personalisation should be the keystone of your customer strategy and the foundation of a common customer language from your frontline staff to head office.”
How data science can help
That is where data science and technology come in. It can help organisations to bridge the gap between what they think and the customer reality – making it outside-in. Through advanced analytics we can get a better view of a customer’s individual behaviours and respond to them with knowledge and certainty.
Today there are vast opportunities from analytics, particularly data science. Machine learning for example can spot anomalies in data as opposed to solving problems it is being asked to. We have seen significant developments in data science in categorising of bank statements, to support open banking, as well as detecting fraud and enhancing marketing. Today, with the advent of technology, data – including non-traditional data – and computing power we are in a position where we can better understand a person, through better understanding the data: all of which is possible due to the advances in analytics we see today.
Machine learning and other advanced analytics techniques can crunch huge volumes of data, from disparate sources quicker, and more effectively than before. More effectively than a person can do. This, alongside computing power, means it is possible to access this higher level of sophistication today. We are not waiting, it is here.