Should customer analytics belong only to banks?

Register now

Banks tend to use customer analytics to cross-sell products to consumers and do things that benefit their own bottom line.

Fintech startup believes that to keep people happy long term, you have to give them the insights that are helpful to them, whether or not the messaging would help the bank. Doing so might not help profits, but it might make for more loyal customers, says's CEO Omar Green.

Green, who formerly worked at Intuit, wants to help consumers benefit from their account data, just as Wall Street firms benefit from buying anonymized data. applies multiple artificial engines to consumers' spending and banking transactions and gleans insights that it immediately shares with them.

Banks' customer analytics programs are typically set up to learn about customer behavior in order to market them certain products. Green recalled an event where he spoke on this topic — bankers in the audience accused him of being anti-marketing.

"I started laughing and said no, we don't tend to think of ourselves that way, but I can see how you could come to that conclusion," Green said at the event. "Think about if it were you: If you found out through that something about an external trigger made you susceptible to spending too much, wouldn't you want to know that before your bank knew that?"

[Digital identity is broken, and fixes are urgently needed. Learn how large financial service and healthcare companies are tackling the issue — to enhance customer experience, to stake out positions in their business ecosystems, and to manage risk — on our Feb. 23 web seminar. Click here for details.]

Some industry observers agree with Green and think the world will move in this direction.

"Banks have a moral responsibility to provide the best advice and not self-serving advice," said Alvi Abuaf, senior vice president and head of financial services consulting at Capgemini Consulting. "That's not to say they should find the cheapest possible product and give their customer the cheapest possible product recommendation, but nevertheless they should still not just recommend theirs."

Abuaf points out that with investment products, banks have a fiduciary responsibility that could be heightened in April, when the Department of Labor's fiduciary rule is expected to go into effect.

Luc Burgelman, CEO of customer analytics software company NGDATA, also shares this view.

"The customer is getting way more powerful and has way more options," he said. "There's a trend when it comes to data, consumers can actually reverse the power and easily switch from one service provider to another. That's not far off."

The Payment Services Directive (PSD2) is forcing European banks to share customer data with others, which will give consumers more control over who does what with their data. Some say this will eventually affect banks in the U.S., too.

How It Works analyzes millions of data points — bank transactions, payment transactions, location, app usage and more — to help consumers make better financial decisions and basically, overcome their bad habits.

For instance, one early user of began taking Ubers exclusively instead of taxi cabs.

"By looking at the data, we could see the number of trips this person was taking were no more excessive than they had been in previous months, but the cost of those trips had gone up significantly," said Green. Most likely, the customer was traveling during surge pricing times.

"What was clear from the data was that the behavior shift this person had made was harming him," Green said. generated a text message telling the customer if he had simply taken a regular taxi for every Uber that he took that month, he would have saved $160. Within a week, the customer's Uber consumption dropped 40%. will analyze how much consumers spend at a vendor and how often they shop there. It also will point out abnormal spikes in spending or account usage and ask if they were intended.

Other chatbots and PFM apps also point out unusually high monthly bills or spending patterns. But isn't rules based and it looks at a much wider set of data — from basic account data to human telemetry data like location and the last five apps the customer used on their phone. It could potentially glean information about the people the user spends time with.

Those friends who always tend to pick the most expensive restaurants? Not helping your finances.

"We would normally say, you spend more when you hang out with Tom and Mary, so you might want to think about that," Green said. "Having that kind of feedback that's honest and without an agenda, that's as close to an objective advocate that we can build." is working through the privacy issues around this and testing it with customers before it will make it a regular feature. The same goes for adding social media feeds to its engines.

"To the degree consumers are comfortable with that kind of feedback, we want to be the people that can help them get it."

Where AI Comes In

The company uses several different artificial intelligence engines to come up with its insights. One, for instance, analyzes when the best time to deliver messages is.

"That turns out to be really complicated to solve," Green said. "Is it better to give it to you at 6:00 in the morning, so it's the first thing you see? Is it better for that message to go out at 9:00 at night when your day is over?"

Also, picking the right time isn't always going to be the same.

"These kinds of ebbs and flows are a unique problem," Green said. "This has to be solved with machine learning because when to interrupt you for a moment to give you an insightful message is going to change every day because your day is going to change."

Artificial intelligence is used to determine which nuggets of wisdom will be imparted to users, which data sets should be analyzed, and which kinds of analytical tools make the most sense for each customer.

"If you're talking about someone who's single and in her 20s and is out in the world for the first time after a college degree, the places you might go to get data and the kinds of analyses you might want to do to create a corpus are going to be radically different than even that same woman 20 years later," Green noted.

The AI engines learn what kinds of messages to which each person responds best. One early lesson learned is it's essential to give positive as well as negative feedback.

"The system needs to be able to give you an 'attaboy' at those points where you feel you should have one," Green said. "We built that into the infrastructure that creates messages so they won't all be, 'You really shouldn't have.' "

The Payoff

While's altruism is commendable, who will pay for this and how will make money?

Green says his investors signed on knowing this idealistic mission was part of the package and that high returns might not come right away. But they also accept that gaining the confidence and trust of consumers could have long-term value. can also be used to provide general business intelligence to businesses, to tell them how well they are doing with customers and even to analyze loan performance. Banks have asked the company to use its technology to find more mutually beneficial loan products for customers. Green declined to share client names.

Wealth management is another potentially lucrative area for Highly personal suggestions generated by the AI engine would give wealth managers an excuse to call the customer.

The company also sees opportunity in helping other startup and fintech companies, to make their products more engaging.

"They think about paying for a solution like to deliver a daily bit of value to that customer and give them another opportunity to converse with them," Green said.

Editor at Large Penny Crosman welcomes feedback at

For reprint and licensing requests for this article, click here.
Artificial intelligence Technology Predictive analytics Business intelligence