Over the weekend, I got into my wife’s car as we were headed out to dinner, and her phone suddenly popped up a message that said we were 18 minutes from our friend’s house. But we were not going to our friend’s house, we were going out to dinner. Is it possible that the technology wrapped into my wife’s vehicle had simply made the prediction that we might be going to our friend’s house at 7 PM on a Saturday night? Whenever we do meet up with our friends, it is typically on a Saturday night and it is around 7 PM. So I was completely baffled by the fact that the technology in the car actually took a chance and had our directions pre-loaded to go to our friend’s house when in fact we were not. The technology took a chance. It was wrong, but it took a chance.

Are we taking chances with our data today? Are we taking chances with our analytics routines? As we continue on the journey of building a business about data, this is a concept we must not forget. Data will undoubtedly bring us down paths where we do not know the answer or perhaps the answer is that this group of members is the most likely to open a checking account, but there is no absolute certainty and that is ok. Taking the chance, and taking the at-bat (if you are into baseball…) is what matters most. Put yourself in the batter’s box and give yourself a chance. Put yourself in the data and send out a message. If you don’t send the message, someone else could be sending the message, or worse yet your audience may not hear your message.

What spaghetti have you recently thrown against the wall?

3 Replies to “Does spaghetti stick to the wall?”

  1. Back in the early 2000’s I was launching a new e-commerce site. We wanted to suggest “customers also bought” companion products everyone now takes for granted. Back then we didn’t have the technology to calculate this on the fly based on actual customer activity. So, we “brute forced” it and programmed them in by hand based on historical purchase data and educated guesses. We threw spaghetti against the wall, and you know what? It worked! Around 15% of orders contained these pre-programmed recommendations.

    Around the same time, Amazon was taking a chance with customized home pages recommending products they thought their customers would be interested in based on purchase history. Their algorithms seemed primitive however, as they only thought I was interested in toys, based on family birthday shopping. I wanted a big screen TV, but they were showing My Little Ponies. Amazon was throwing spaghetti too, but through the process, they were learning and getting better.

    I think the point is just get started. Start the journey now, with one step and then another and another. Pretty soon you’ll be serving tortellini on a plate and then… who knows!

  2. I like the idea that they just got started….so many times CUs think they do not have the TECH to look like they have the TECH. There are a lot of rules of thumb about what people like and what products or services go together. No tech needed, just bundle them in offers and create suggestion that they go one step further….they will give your tech the credit!

  3. Taking a chance on predictive analytics will eventually lead to prescriptive analytics. Credit unions have access to an enormous amount of data about a member and their life cycle. As we study trends of the data, it is very possible that we, as analysts, will begin to know what the member needs before the member. Getting in front of the member and consistently prescribing these smart solutions (via emails, web banners, online banking messages, Next Best Product, outbound calling etc) will eventually lead to that trigger of execution.

Comments are closed.