04 Mar 2015
Of course, behind this approach there was not a lot of data or analytics – nor did there need to be – but there was a key insight in understanding the power of suggestion at the point of sale. In a more complex environment, and for higher involvement products and services, this has evolved to 'cross-sell', 'up-sell' and 'next-best-product' conversations.
"The majority of companies seem to be using this technology in a very shallow way to just … sell more of their products or find new customers."
Patrick Maes, Chief Technology Officer, ANZ
Driven by improvements in data and analytical capabilities, these can be quite sophisticated conversations but are there hidden challenges in this field, traps as well as opportunities?
Analytics has been around for many years. In fact I started my career using analytics and artificial intelligence to underpin AXA's ambition to become “La banque de conseil" (the advisory bank) 25 years ago. While the algorithms we used in the eighties were pretty advanced, the main difference between now and then is we didn't have the computing power we do today and of course the data sets are far larger now due to the digitisation of our society.
We do read a lot about 'the Big Data promise'. The hypothesis made is pretty simple: we can use analytics on this data to provide deeper insights about our customers (such as aspirations, preferences, propensities, etc) enabling us to serve our customers better and create a positive effect on their well-being.
However, the majority of companies seem to be using this technology in a very shallow way to just find ways to sell more of their products (or services) to their customers or find new customers - this is what we in the industry call 'cross-sell', 'up-sell' and 'next-best-product'.
While these strategies can lead to a degree of commercial success, using analytics to be better at selling is not enough to give you an edge because everyone is getting into this sort of thing and it doesn't create a competitive advantage in a sustainable way. In fact the mass usage of “simple" analytics or what I call “McDonald's Analytics" will only bombard people with a massive amount of offers with very little sophistication. The result will be a new generation of spam, similar to the printed advertisements that caused an overflow of our post boxes in the 90s… 'digital junk mail".
Big Data and Analytics will only provide a better outcome if organisations move away from the pure transactional dimension to a genuinely customer-centric approach.
A lot of bank technology in recent years has been dedicated to products and services which were about 'empowering' the customer, especially through digital channels. But many of these were standalone offerings built from individual elements in isolation, creating an often inconsistent and confusing experience for the customer. For consistent user experience there is an urgent need to bring them together.
Data and analytics are crucial here. The model below outlines the five key components that I believe are required to offer a truly customer-centred approach.
The single-customer view is the key building block in any analytical capability. The most basic start in data is to consolidate all your current structured customer data and create that single customer view. Nothing fancy, just a bit of master data management (matching) and data clean-up - boring but very necessary. We are all aware of the 'annoying' experience of the next great offer 'empowered by analytics' when we already have the product or it is not relevant to us.
Once you have this in place, you can enter the world of analytics, not the cheap 'McDonald's' versions (often just correlation engines) that almost any technology and data company can offer you but advanced modelling techniques such as Baysian networks.
An example of this is a recent success ANZ enjoyed as part of a university collaboration program we are running. We won the Best Industry Paper award at the Australasian Data Mining Conference 2014 for our project on The Market Segmentation Of EFTPOS Retailers, where we applied clustering techniques on finding groupings of retailers who use the Electronic Funds Transfer at Point Of Sale (EFTPOS) facilities. Other approaches had focused on clustering the consumer but our approach allows a fresh way to look at data to gain business benefits beyond traditional approaches.
The key here is to find the data scientists who can work with these models, a skill not easy to find nowadays.
Imagine you have found a great analytical outcome, the next step is to explain this in the context of your business. For instance, a strong possible correlation between the success of Justin Bieber and the decline of the US economy post GFC is interesting but useless if you can't explain it.
This highlights the importance of heuristics (in computer science and mathematics, a shortcut to problem solving when traditional approaches would be too slow, including approximate solutions when there are no exact ones) and common sense which compliments the analytical approach to data and the introduction of self-learning technologies found in cognitive machines such as Watson (IBM).
Complementary to this artificial intelligence, the human brain can process complex data from visual stimuli and identify patterns far quicker and more accurately than computers are currently able to do. This means data visualisation, when done well (e.g. multi-dimensional), adds significant value in the speed and accuracy of decision making and sometimes provides more insight than pure analytical data crunching.
Last but not least, we humans like to play. Gamification technology allows us to explore scenarios in a “fail safe and fail fast" environment. Gamification can help you to take all this insight created from previous capabilities and bundle them all together in a compelling, engaging and fun interaction. That's exactly what Big Data should be about.
This brings me back to the real value of improvements in data and analytical capabilities.
We all know most banking products are commodities with very little room for differentiation. We can either be the trendy, social media obsessed, product bank that, after exhausting all (often defensive) marketing options, will probably end up differentiating on price. Or we can be the intelligent, customer-centric organisation that looks further than products and wants to play an important role in helping people and companies succeed.
When we complement our data sets with discovery and learning, we shift from products and services to value-adding advice, creating better outcomes for our clients and ultimately better results for our businesses - or in the words of Jack Byrnes (Robert De Niro) in Meet the Parents, we can be “in the circle of Trust".
Patrick Maes is Chief Technology Officer and General Manager GTSO at ANZ. You can find him on Twitter here (@DrsPatrickMaes) and on LinkedIn here.
The views and opinions expressed in this communication are those of the author and may not necessarily state or reflect those of ANZ.
04 Mar 2015
12 Mar 2015