9 mins to Thane Road!! Using Segment-of-One Profiles to Transform Your Business
By Jon Pearson, Sales Director, Transformational Digital Journeys, Banking @ Feedzai
It was Thursday evening at five past seven in the evening, as I left the house and walked up the drive towards my car, a message flashed up on my phone… 9 mins to Thane Road!
What on earth could it mean?
As it happens, I was heading to Thane Road to play 5-a-side football the same as I do every Thursday, and running late, as usual, it was very useful to know that the traffic was clear and that I would arrive in time for 1915 kick off.
Let’s examine what happened here, a unique profile of my previous behavior existed and my phone decided to push some useful information to me.
Firstly, it knows the time, this is critical, had it been another time or another day I wouldn’t be going to Thane Road.
It knows my location and that I am walking from my house to the road.
It knows the traffic conditions on the route I normally take.
Think about this for a minute, in the very near future with driverless cars, does this mean that I wouldn’t need to order an Uber? Would it be there waiting for me in anticipation?
What if it was a Thursday when I wasn’t able to play? Would it then be able to track, adjust and pick me up from the station instead?
The possibilities amaze me!
So what could this mean for your business?
Let’s assume you are a bank, a customer visiting an ATM could be offered the ‘usual’ as if he was ordering a pint in his local pub.
“This is what segment-of-one profiles do, they bring you closer to your customers.”
To figure out how to formulate a plan to incorporate segment-of-one profiles in your future strategy, I have included an introduction below.
Segment-of-one profiles, an introduction for the uninitiated.
The advent of segment-of-one profiles has coincided with the introduction of new technologies that enable the creation of low latency, high throughput streaming data environments. This evolution in technology is the key to unlocking the shift to segment-of-one profiles.
Let’s take a look at an example of credit card payments.
To get an accurate view of a payment transaction, there are many different expert profiles required to form an accurate view of the risk occurring in real time.
Your “credit card profile” tracks many things:
How much money you typically spend on travel, on restaurants, and on shoes.
The time of day and day of the week do you typically shop and on what.
The seasonality of your behavior and its variations.
Finally, each one of these profiles extrapolated over different time frames, last 12 months, Last week, last hour, last 5 minutes.
This creates a group or an ‘ensemble’ of profiles, each an expert from its own perspective, all of which ‘vote’ and allow an expert decision to be made on behalf of your business and your customer.
Segment-of-one profiling, in combination with machine learning, enables you to predict a customer’s immediate needs ‘in context’, which means that you offer your service when they need it the most.
If my bank used this, I would stop receiving their balance transfer mailer 4 times a year and they’d know that I pay my card in full each month rather. Were they able to construct my profile, they’d perhaps figure that a charge card would be more suitable for me
In the context of a home buying experience, imagine being able to geolocate your customers driving between home properties currently for sale and push them with a pre-approved mortgage offer to their banking app based on the average value of the properties they were looking at!
Which part of your business could you transform to amaze your customers by helping them when they need it most? The only limit is your imagination.
Feedzai helps its clients do this very thing and transform their business in this new age of AI. In one instance we generated over 500 profile features and 180 models in 6 weeks to transform a business process. In another case, one of our clients scores 25 million transactions a day with hundreds of profiles all in less than 10 milliseconds.