Go Jump In A Data Lake

Subscribe Now!

Stay on top of all the latest news and trends in banking industry.

Untitled(Required)

Learned a new term the other day: “data lake.”

I’ve heard of databases, data warehouses, data stores, but never a data lake. The guy I heard the term from even had a picture of a square body of water to help us visualize the concept. Cool.

Although I have no better idea of what a data lake is after hearing the term than I did before hearing it, it may very well have been one of the more comprehensible concepts coming out of an IBM presentation on Transforming Banking With Analytics.

***

Of course, it could be me. Just because I…

  1. Have an MBA in Finance and Statistics
  2. Was a teaching assistant for a graduate-level statistics course,
  3. Developed a regression model to predict health care demand for my Master’s thesis,
  4. Created marketing and financial models back when BI was called DSS (decision support systems),
  5. Developed countless other models to size and forecast demand for financial products and services, and
  6. Worked for a database marketing firm alongside brilliant PhDs who developed sophisticated analytical marketing models

…doesn’t mean I actually know anything about analytics. Maybe the presentation was over my head because I’m just a dumbass consultant/analyst.

***

What did I find confusing? For starters, everything involving the word “cognitive.”

Apparently, IBM is focused on helping banks become Cognitive Banks, “integrating cognitive computing, a data driven digital network, and a secure and scalable infrastructure to drive value from business ecosystems.”

Cognitive computing, in case you’re too dumb to know — like I am — incorporates machine learning algorithms, natural language processing, and cognitive analytics. Cognitive analytics, I assume, should not be confused with other types of analytics like predictive and prescriptive analytics.

IBM went on to say that the Cognitive Bank “leverages Cognitive Operations, Cognitive Analytics, and Cognitive Engagement. Cognitive Operations is “a Cognitive core that coordinates a network of distributed business services and drives a simpler, leaner organization that can make faster decision and is closer to customers.”

Cognitive Engagement (as I’m sure you already know) refers to “cognitive-orchestrated systems of engagement that align financial services supply with a customer’s economic choices and optimize the customer interaction and experience with the bank.”

I spent the better part of the day before the IBM event with the executive team of a mid-sized bank. I tried to picture myself telling the team that they needed “cognitive-orchestrated systems of engagement to align their supply with their customers’ economic choices and optimize those customers’ interactions and experience.” And I couldn’t decide if that picture was the funniest thing I ever imagined, or the scariest thing.

What will be the impact of all this? According to IBM, “cognitive will transform the financial service sector.” After seeing that, I’ve come around to thinking that Coin has a better chance of transforming the industry.

The coup-de-cognitive-grace was the slide that asserted that business value increased as you move up the analytics food chain from: 1) Descriptive (what has happened?) to 2) Predictive (what could happen?) to 3) Prescriptive (how can we achieve the best solutions?) to 4) [yep, you guessed it] Cognitive (tell me the best course of action?).

Ignoring for the moment that “tell me the best course of action?” isn’t actually a question…how is “what’s the best course of action?” NOT prescriptive? And how is “how can we achieve the best solutions?” prescriptive?

I asked the presenter to clarify this, was given an answer, and was asked if the response clarified it for me. But I had to honestly answer “no.” Clearly, I was the dumbest analyst in the room, because nobody else seemed to have a problem with all this.

***

Communication Breakdown

At this point in the presentation — that is, 15 minutes in from the start — I was thinking about Paul Newman. Not because I have some kind of weird thing about the man, but because of a line from one his movies: “What we have here is a failure to communicate.”

What we have here is a communication failure — a marketing communications failure. To be more specific, the meddling of marketing into service delivery.

In an attempt to carve out a different positioning in a crowded space, IBM has apparently felt the need to commandeer the word COGNITIVE, slap it front of any and every analytical concept it can find, and place it at the top of the value chain in attempt to imply that only they can help banks achieve the nirvana of analytical transformation by reaching the pinnacle of….um….cognitiveness.

***

Speaking of “transformation”…

My favorite question/comment of the day came from Celent’s Bob Meara. One of the IBM presenters (my favorite presenter, by the way, because he was the analytics practitioner) talked about the path to injecting an analytics capability into an organization. He said — and most of us couldn’t agree more — that it needs to start with the definition of a specific business problem to be solved.

Bob jumped in at the end with the following (quoting as best as I can remember): “So the key to transformation through analytics is to address and improve existing business problems, which, of course, is anything but transformation?”

The IBMer’s answer: “Yes.”

And thus ends our speaking of “transformation.”

Measurement Malpractice?

So, if we’re going to focus on real business problems instead of aspirational transformation delusions, what business problems should analytics efforts focus on? How about increasing loyalty, improving customer retention, and increasing customer lifetime value? Sounds good to me.

What didn’t sound so good to me were the “potential outcomes” IBM claimed could be realized: 10-30% improvement in loyalty, 10-50% improvement in retention, and 10-25% improvement in CLV.

Really? Where did these numbers come from? Certainly not associated with any named banks. And of course they weren’t, because they aren’t real — or realistic — percentages.

The worst banks out there have retention rates of 80%. A 50% increase would put retention at 120%. Which isn’t possible, of course.

And the improvements in the other metrics? How exactly did IBM determine that a CLV improvement of 25% was due to analytics efforts, and not product design changes, or general market conditions, or TV advertising, or improved customer service capabilities, OR ANY OF THE OTHER GAZILLION THINGS the bank does?

Twitter Twaddle

And then there was the discussion regarding the use of Twitter data to identify life stage change signals, like identifying that someone is in the market for a new home, and making a mortgage offer to that person the next time he logs on to online or mobile banking.

I asked how a bank was supposed to be able to tie a Twitter ID like @KnowItAllAnalyst to Ron Shevlin, customer of the bank with account #12345678. The actual know-it-all analyst sitting across the aisle from me jumps in to answer and says “you ask customer for their Twitter IDs.”

Yeah, right. When I asked IBM how many banks they knew of that had done this, and had any meaningful number of Twitter IDs tied to customer records, they said “none.” (Don’t mess with me, know-it-all analysts!)

I went on to point out that not only can’t banks tie Twitter IDs to specific customers, even if they could, not everybody is on Twitter, not everybody who is on Twitter is active on Twitter, and not everybody who is active on Twitter is one of the obnoxious sorts who tweets every damn thought that pops into their head.

So, in the end, the percentage of prospective mortgage applicants you’re going to identify by tracking Twitter is infinitesimally small. If, however, you still think this is a worthwhile use of your marketing department’s time, go for it. Don’t let me ruin your social media wet dreams.

***

The unfortunate thing here is that there are probably few — if any — companies out there better positioned to help banks improve their analytics-related efforts. The company’s consulting resources, teamed with their research skills, and backed by the marketing and analytics technologies that IBM has acquired over the past six to eight years makes it a formidable competitor in the analytics space.

Why the company feels compelled to wrap their analytics offering in a shell of thought-leadership nonsense is beyond me. All this cognitive-this and cognitive-that left me with, well, cognitive dissonance.

Cognitive Dissonance

Bank CEOs don’t care about scaling the value chain of analytics from descriptive to predictive to prescriptive to cognitive. They just want to make better decisions and reap better business results.

One CEO I recently met with mentioned how one of his employees pulled some data to identify which customers had responded to a recent marketing campaign, and how the profiles of respondents differed from respondents to prior campaigns. The data, he said, was insightful, and gave them ideas for future marketing campaigns.

Nothing real special or sophisticated about that, is there?

In the context of IBM’s model, maybe this CEO should have done nothing with the data, opting instead to give it to some PhDs to build some high-falutin’ Chi-square prescripto-cognitive model to optimize and maximize the value of their analytics efforts.

***

Bottom line: The hype around analytics is obscuring the value that can be derived from simply making better, more data-driven decisions. Kinda makes me want to go jump in a data lake.

This article was originally published on . All content © 2022 by The Financial Brand and may not be reproduced by any means without permission.