Mid-way through #money2020 and in my fourth back-to-back chat about Fintech and leadership and core systems renewal and machine learning and and and … I am suddenly taken back to my roots. My roots are firmly in data. I guess that’s because I started in technology looking at office automation and workflow, process re-engineering and straight through processing. That evolved into data warehousing, data mining, data analytics and propensity modelling. Showing my age, those areas were my roots dating back to the 1980s. Thirty years later, we are still struggling with my roots.
The reason this cropped up is that I was talking GAFA and FATBAG with a couple of folks, and they wondered why banks cannot be more like an Amazon . Amazon is a great data mining firm. Amazon has over 150 million active accounts, 1.5 billion items that can be ordered and delivered through 200 fulfilment centres. Their online catalogue of products receives more than 50 million updates a week and every 30 minutes all data received is crunched and reported back to the different warehouses and the website.
Why can’t banks be that agile, nimble and deep?
Facebook gets 10 billion messages updated daily. 4.5 billion of those messages get a Like , many of them on the 350 million pictures that are added every day by the 1.5 billion users. Facebook knows who our friends are, what we look like, where we are, what we are doing, our likes, our dislikes … in fact, Facebook has enough data to know us better than our therapists (for those who have them).
Why aren’t banks as knowledgeable about our financial world as Facebook is about our social world?
Google answers 40,000 searches a second which equates to over 3.5 billion searches per day and 1.2 trillion searches per year worldwide. Each of those searches uses 1,000 computers to retrieve an answer in a fifth of a second, with 1 in 5 searches never having been asked before.
Is a bank that efficient with its data?
AliPay processes 85,000 transactions a second supporting 450 million users. These stats make a bank’s typical data usage look trivial.
So what does a bank do with its data?
Not a great deal so far. Much of the data is distributed amongst the silos and fragmented systems in the office.1 Those that have been consolidated and rationalised are ok but, even then, are often underutilised.
Machine learning, artificial intelligence, deep data analytics and algorithmic mining will change all of this of course but, today, it’s still hard. Forrester reckon that 99.5% of most corporate data is not analysed. I’d be surprised if the numbers were so high, but not that surprised. After all, banks are like telco’s, retailers and other high contact companies. They may get 1, 2 or 10 touches a day through mobile apps, but do they know what those touches mean? Can they enhance each touch with personalised data to make the customer feel special? Can they enrich each touch with context to grow their breadth and depth of relationship with those individuals every time they make contact?
This is what the enhanced world of data analytics combined with machine learning will offer, but I’m guessing that most existing banks, retailers and telco’s are still stuck with the same problems they had back in the day.
Back in the day, they built their systems to reflect their products. Their products are structured in different systems and lines of business. They call them silos, but they’re really just segregated structures run by baronial crown princes and princesses1. Try to destroy those fiefdoms and consolidate all the data into a single enterprise view is like trying to unite all the kingdoms of Westeross under one throne. It ain’t gonna happen.
The reason why Amazon, Facebook, Google and AliPay can leverage data so effectively is that they all began with a single data view and have added leverage to that single data view over time1. When Google was founded in September 1998, it was serving ten thousand search queries per day. By the end of 2006 that same amount would be served in a single second. In August 2012 , Google’s search engine was monitoring more than 30 trillion unique URLs with crawls over 20 billion sites a day, and processing 100 billion searches every month or rather 3.3 billion searches per day or over 38,000 thousand per second.
Platform is the model; scale is the key; and consistency is the focus.5 We do see this in some of the transaction banking areas of finance but, generally, for customer account data – retail, commercial and investments – the banking silos still rule their data structure. A hangover from the past, a reflection of product focus, and a lack of ability to invest in consolidation and rationalisation. However, if data is the competitive battleground, it’s just another reason why banks need to get an enterprise data structure in place sooner rather than later.
After all, if GAFA and FATBAG are the competition, are we ready and fit to compete?