Deeper Relationships via Fast Machine Learning

Relationships deepen faster with a lot of interaction, listening, and learning -- because the relationship becomes more rewarding. Last week, I met with the chief data officer of a very large U.S. financial service provider (9MM unique visitors per month) and he, too -- is always trying to connect customer click streams back to his machine learning systems so direct expressions of customer interest can be used to train the classifiers used to determine audiences for marketing messages.  Unfortunately, advertising and email can only provide a small amount of information on positive interest, and almost nothing on "no interest".

Classifiers, those functions that divide people into groups based on their characteristics, improve faster when they learn from high volumes of both positive and negative information; learning that is impossible with ads and email due to banner blindness and spam filters.  For example, 999 people out of 1,000 don't click on a banner ad; is that because they aren't in the market for a product?  Or, perhaps 995/1,000 just didn't see the ad... who knows?   However, when 20%+ of all users interview with the financial institution on a monthly basis, the amount of information provided for training the classifiers is enormous and the rate of machine-learning is correspondingly high. High machine-learning rates translate directly into better targeting and more sales per square centimeter of digital real estate.  For example, look at the interview responses from a delinquency/skip-a-payment campaign below.    

 Table 1 - Example both positive and negative responses to interview questions.  

Table 1 - Example both positive and negative responses to interview questions.  

Note that we not only received information on who was interested in skipping a payment, the warm leads, but actually more information on who considered the offer but is explicitly not interested in skipping a payment.  That's important because, we assign both positive and negative values to these responses to train the our classifiers.    

Fast learners engage more and sell more.  For example, here are June 2017 results from three financial institutions, all below $10B in assets, using Micronotes to engage with their mobile and online digital banking customers.

 Table 2 - June sales results from 3 community financial institutions.

Table 2 - June sales results from 3 community financial institutions.

*Forecast sales is based on industry standard click-through rates for banner ads of 0.19%, a landing page conversion rate of 2.35%, and an identical conversion rate of leads to sales as Micronotes' leads.  

What's stunning about these results is not only the dramatic improvement in sales, which directly deepens relationships, versus advertising, but also the corresponding rates of learning from much higher click-through rates and the use of negative responses.  If you believe that deeper customer/member relationships reduce attrition and protect your customers from poaching, you would be correct.  

So, who do you think is going win the retention battle: the fast-learner or the slow-learner?  So, if you're not learning much from your digital banking customers/members, try Micronotes, and rocket to the top of the class.  You'll not only learn more and sell more, but -- you'll participate in the artificial intelligence and machine learning productivity revolution, without much effort.  

Try Micronotes free for 30 days, and learn about which life events your customers/members are, and are not going through -- and who does, and doesn't need your help.  A webmaster can install the trial software in about an hour so, it's a no-brainer.