In our last blog on customer analytics, we highlighted three things you need to do customer interaction analytics well. As a refresher, these were:
- A readily accessible and complete source of customer interaction data
- A range of complementary data like demographics, BIN lists, ATM descriptions, etc.
- A tool to explore, segment, and present customer interaction data
In today’s blog, we are diving deeper into the first requirement.
Let’s start with a working definition of customer interaction data. Customer interaction data is the sum of all the “digital footprints” a customer leaves behind as they complete a banking transaction.
Examples of these digital footprints include: when they logged in and from what device, reading about a loan product on their tablet, checking their balance prior to withdrawing cash from an ATM, etc.
Now, in reading the prior paragraph, you’re probably already thinking about all the different applications, systems, channel teams and departments who may have captured these digital footprints (and how hard it will be to get them all back out and organized into a “ready to analyze” state).
This is why it’s so important to make customer interaction data readily accessible. You need to get a “data lake” together to facilitate your analysis. You have to find a way to either:
- Get customer data out of all those places in a simple and timely fashion and then pool it for analysis
- Capture customer data and pool it before, or as it is being written, into all those locations
The first way is the classic data warehouse strategy – extract, transform, and load data from source systems into a single analytic database. It’s like backing up to an ocean with a tanker truck, filling it with data and then trucking it to your lake…over and over again. In our opinion, this strategy is not working for most banks, and especially for channel managers. The data is not readily accessible, requires IT service requests every time you need a different data slice, and the information is typically out-of-date before you receive it.
A second way is a new approach, made possible by streaming transaction data platforms such as INETCO Insight. Data can be captured in real-time of the network as customers are interacting with various channel applications and systems. Nobody needs to change the underlying systems and the platform brings all the data together immediately for self-service access (we’ll talk more about this in a future blog). It’s like redirecting a river into your own lake bed…
We also used the word “complete” in our definition above. This is because everyone keeps a record of a customer’s interaction – that’s what a core banking system is for after all – but it’s too sparse for the kind of analysis we want to do. What ends up in your core banking system is more of an accounting record – funds moved from account A to account B – not a customer record. A customer record tells you all the various steps the customer took to get to that transaction (and ideally, why they did it).
A complete source of customer interaction data should “sessionize” all the digital footprints it captures so you can see this individual customer journey. A simple example is showing all the transactions a customer completes (and in what sequence and over what timeframe) from the moment they insert their card into the ATM, to the point the ATM returns their card.
So, by using a streaming transaction data platform to capture customer interaction data as it happens, we can develop a complete customer record. In our next blog, we’ll look at how to enrich this core data stream with complementary data sources to make your analysis and decision-making efforts even easier.
Part 3 and Part 4 of this series is now available!
To request a demonstration of the INETCO Insight or INETCO Analytics products for fraud analytics, please contact us.