Despite a multitude of algorithms crunching mountains of information, there is a vital missing link in the data analytics rush. Finding it is key to relevance, perhaps even survival.
What is the hardest thing for a bank to acquire? Not data, certainly. As an institution that its customers trust, a bank has more data than it knows what to do with.
Data is compiled, sorted, analysed and put to use in myriad ways today. For starters, banks serve up static information to the customer. It can aggregate your deposits and showcase your portfolio value. It might even recommend a loan for you, based on your balances. But all of these, even the last instance, are simply rule-based responses, not necessarily intelligent ones.
Not quite there yet
Consider the rash of new offerings that a bank pushes your way. These are typically based on some pattern. For instance, those customers in the age group of 30-40 are skewed towards more buying, more transactions. Those older tend to save more. These are all non-feedback-oriented prompts. As long as there is a discernible pattern, a bank is fairly confident of assessing what its customers need, and pushes products accordingly. But why is it that most offers don’t really assess your need? Why do they tend to be irrelevant and lacking in context?
All banks want to get to a stage where any information they display to a customer – static or dynamic – is qualified with something contextual. A stage where a customer is prompted not post-facto, but right at the outset of a transaction, based on data which the bank already has. For instance, when a bank displays the balance of an account, prompt the user if it is low and recommend that he pulls funds from somewhere. Or give the customer a heads-up to ensure there is sufficient balance, aware that there’s a payment coming. The more contextual the prompts, the more discreetly they can be served to the customer, the the more attractive the bank is for new customers, and the stickier the customer’s interaction with the bank.
But wait a minute. Is that what the customer really wants? A stickier experience?
Does a bank need to do more?
Not really, one could argue. A customer logs on to the bank’s portal not to spend time on it, but to get things done efficiently and very quickly.
After all, if the business goals of a bank are met, why bother with obsessive number crunching or contextual offers or ‘superficials’? This isn’t an isolated opinion. Take any cooperative bank. They have not quite harnessed technology. They sport a portal that only reflects their rudimentary process flows. Many of the bigger, universal banks aren’t doing much better. They approach portal or channel as an aesthetic exercise, which are online reflections of their fragmented product portfolio*.
If only things could remain the same, if only change would give us a heads-up and happen over time.
As we explored in a previous blog, a fundamental shift in the benchmarks of customer experience – blame FinTechs, or the larger ecosystem of addictive enterprises like Amazon or Netflix – has forced banks into quickly putting together their digital aspirations. And the fuel for this process is clearly data.
And to process this data in a way that can actually help banks instead of creating more packets of data, we need to look at intelligence not as some nebulous fixture, but infused into every step of every user journey.
Waking up to intelligence
Consider this sentence again – A bank has more data than it knows what to do with. This is a telling indicator of where they stand in the data analytics rush.
What banks want to do is to provide the right information to the right person at the right time. Even in read widgets, and write widgets, there should be intelligence to see if it is straightforward transaction, or a path to authentication, or should a second or third factor be invoked. In other words, making it contextual at every block.
There are algorithms, yes, but it isn’t just the algorithms themselves, but the inputs that should feed the algorithm to arrive at a specific, actionable insight.
A common misconception is that intelligence comes from outside. In truth, intelligence is a bottom-up approach. Collecting data, tagging it smartly, and making it work for you; this is where intelligence comes in.
The first question a bank needs to ask itself is, ‘Am I using intelligence effectively’? The answer, across all institutions, is a negative. Once this is recognized, the possibilities are endless.
When banks begin to recognize intelligence as an over-all concept, and not something compartmentalised, they can begin to truly explore the applications of intelligence.
Let us look at just two use cases we all recognize and can realise, from existing tools. AML, for instance. The traditional use of algorithms in AML is post-facto. To detect an AML pattern, it needs to be seen through to the end. However, AML is now moving into fraud detection. There are mechanisms where you can score at transaction, and provide security at the customer level. You still allow the transaction to continue, but you allow it with a fraud control mechanism baked in Intelligence.
A second use case is in pushing new offers to very specific customer segments. Currently, banks use very broad patterns to identify prospective customers for a particular product. What if we used more advanced statistical models – like Adjusted R-Squared – instead? Without getting too deep into the tech, a large number of variables decreases the ability of a model to predict accurately. Adjusted R-Squared, on the other hand, accounts for these variables better and provides the most accurate degree of relationship within a particular segment / population. The end result is that the information you serve becomes more contextual to your customer.
The whole purpose of looking at intelligence and its application seriously is the answer to the question we asked right at the beginning – What is the most difficult thing for a bank?
The answer is – to understand your customer’s intent.
Irrespective of the data you have at hand, the patterns of customer behaviour and transactional information you record, the most elusive bit of intelligence, and also the most valuable, is the customer’s intent. Why did he make a series of fixed deposits? What are his goals, his aspirations? What anxieties prompt him to make a particular series of decisions?
This is a greenfield area, one which we are actively working on. We are building what we call the Purple Fabric to help coalesce data and intelligence from disparate systems and deploy it in a manner that is unprecedented, and evolving all the time.
The era of analytics without conclusion is coming to an end. It is time we stopped crunching numbers and patterns for the sake of it and instead put them to work.
This will be the key to banks’ survival in the increasingly competitive, tech-driven world.
In subsequent blogs, we will explore cutting edge applications of intelligence, and also grapple with questions of regulation, privacy and security.
*There are, of course, exceptions to this on both ends of the spectrum. For instance, Co-op bank either don’t have portals, or they approach it in a 'brandless' way, without applying aesthetics and design. We also need to consider what ‘portal’ itself means to banks. For instance, some mid to large retail banks have fragmented product portfolios and portal is the place where all of them aggregate, consolidate.
Harish has a career record spanning about 13 years as a hands-on technology expert, presales, product manager and solution architect for strategic projects and product initiatives. He has been working with Intellect for the past 11 years.