Big data and your credit union: Black magic or salvation?

The stampede is on. Suddenly credit unions across the nation have become convinced that big data is the path to prosperity. The good news: that’s half right.

The bad news of course is that it also is half wrong. But there are ways to get this entirely right.

First, however, understand that big data really is all its proponents say it is. The plainest examples are how Netflix and Amazon know – really know – what to recommend to you if you are a frequent customer.

Most credit unions are far from that standard.  

“It’s poorly understood among credit unions,” said Paul Ablack, CEO of OnApproach, a data analytics company focused on credit unions.

In a survey of hundreds of financial institutions, MX, a data focused fintech serving some 1800 credit unions and banks, asked them if they believed they were making efficient use of their data. 92% said no, according to MX CMO Don MacDonald.  He added that many credit unions, when asked why they found themselves in that predicament, responded that they did not know how to get started.

Keith Brannan, CMO at Kasasa, a financial services company that serves credit unions and community banks, said “The biggest issue – and the reason most credit unions don’t do big data  – is a lack of data. Most of the tools require mountains of data,” said Brannan.

Ablack agreed – in his mind, maybe one credit union has enough internal data to run meaningful big data projects. The others don’t have the bulk.

This is not good. Big banks are awash in data and they are becoming experts at using it. “This is becoming a story of the haves and have-nots,” said Brannan, where the haves truly understand their consumers and prospects. And the others are left to guess.

Guesses often miss the mark and that’s a prime reason why so many credit union marketing initiatives fail to hit the bullseye.

But now there are emerging paths that get credit unions into big data in ways that work.

For instance: there’s a growing number of big data-based fraud prediction services that, drawing on knowledge of fraudsters, sell their service on a subscription basis to credit unions.

Case in point: PointPredictive which has amassed 57 million auto loan credit applications, drawn from both lenders and car dealers. According to Frank McKenna, chief strategist, PointPredictive uses its data to hunt for fraud in auto loan applications and, he said, they find a lot. Every application submitted into the system is scored and it’s the choice of an individual lender to fund or decline a loan – but with that data, their decision is fact-based.

Know that there are plenty of other fraud prediction services – it’s becoming a hot area for big data services aimed at small and mid-sized financial institutions. Look around and compare. But understand that subscribing to a service helps level the playing field for credit unions as they combat increasingly sophisticated criminals.

Can’t a credit union go all in with big data and begin applying it to a range of activities, from account opening to knowing what offers to extend to which members, when?

Doing that requires a lot of data and, said Ablack, for starters look way beyond payments data and core system data. He itemized some of the data sources he tells his credit union clients to gather up when they start a big data campaign:

* Payments
* Every transaction a credit union does
* Loans
* Checks
* Debit card activity
* Credit scoring info
* Call log info
* Online banking info (with attribution – where were they when they logged in?)
* Who is the car insurer? When is the policy up for renewal?
* Building permits
* Vehicle info

That sounds like a lot of data? Get this: you need more. That’s been the hold up for meaningful big data projects at most credit unions but there now are initiatives intended to solve the problem of not enough data.

How? Ablack’s perspective is that for credit unions to really get the benefit of big data they need to pool their data, in what experts call a shared data lake. Get 100 credit unions putting data in a data lake and, suddenly, there’s scale that helps a credit union begin to mine data to guide better marketing and decision making.

Ablack believes that credit unions are uniquely positioned to understand the value of data sharing. He pointed to the industry’s ATM sharing – such as offered by Culiance – as a reason for confidence that credit unions will hop aboard data sharing. “Credit unions got together, they collaborated, and they created a great network,” said Ablack.

Similar can happen with big data, he said.

Are there issues that need resolving? Definitely. Everything from ownership of the data to anonymity of member data to what happens when a credit union decides to withdraw will need to be sorted out.  

Multiple companies now are in the early stages of creating credit union shared data lakes.  Again, shop carefully.

But – very probably – data lakes will become mainstream, as a smart way to give smaller financial institutions parity with the megabanks.

When will this become mainstream among credit unions? Ablack’s belief is that in 2019 big data will really take off.

And if that’s so, credit unions will have a bright future because – very probably – the future will be brightest for the institutions that have the most mastery of data. And that future now is within reach.