That quotation from Peter Drucker goes back a few years, before the advent of the internet and online systems, when customer playing data was notoriously inaccurate and subject to the whims of the people rating the players. Slot machine players were practically ignored. Laboriously inputting machine data into spreadsheets was not a fun task, was prone to errors, and did not tell anyone very much.
The advent of online communication has increased the availability of data exponentially. Every day 2.5 quintillion bytes of data are created, that is 2.5 times ten to the power of eighteen. In fact, there is so much data that we are swamped with it; parsing and making sense of it is extremely difficult.
In an online retail environment, we now know who bought what, when, for how much, what else they bought, whether a specific ad or offer triggered them to make the purchase, what they looked at before and, probably, what they looked after making the purchase. Retailers can test different promotions or page layouts and see in real time the impact of these changes. Customers see homepages with offers tailor-made for their particular profile. Walmart is said to collect over 2.5 petabytes of customer data a day, the equivalent in the analogue world of over 50 million filing cabinets full of text. This amount of data is impossible to analyse without some very sophisticated computational tools.
Why is big data important? Through analysis it allows us to see patterns that were are otherwise not obvious. By measuring we can be much more precise in our understanding, make better predictions, and improve decision making. Making decisions by gut or intuition should be a thing of the past.
I have seen too many companies that are still run on gut instinct. Decisions are made on the fly based primarily on past practices. For example, a company might use actual win for a customer rather than theoretical value in determining the amount of comps that the customer should receive. I’ve experienced companies that use the HiPPO approach: the highest paid person’s opinion sways the day.
Some companies use data in minimal ways. For example, when trying to decide which machines to replace on a machine floor, any machine that has a daily gross win less than the average is a candidate for replacement. Too many times I have seen slot purchasing decisions that result in revenue moving around the slot floor rather than generating a genuine sustainable increase. And some managers make decisions and then back fill their decision with cherry-picked supporting data.
Caesars, or as it then was, Harrah’s, spent a great deal of time and money building a large data warehouse and the analytic capacity to study large volumes of data so that they were able to understand machine performance and customer behaviour. They got to the point where they could predict the life time value of a customer after only a few hours play. Caesars, probably more than any other gaming company, has a rigorous approach to data analytics – decisions are not taken without data and the supporting analysis. Their Total Rewards program is second to none and the company’s culture revolves around the collection and analysis of data in decision making.
Scientists are taught that first you understand the challenge, create a working hypothesis, design a trial that will test the hypothesis, and have a robust definition of success before the trial begins. If the trial doesn’t succeed, learn from it and design another trial. Some companies follow this approach except for how they define success; that allows for wiggle room later on. But without a rigorous approach the outcomes will almost always be suboptimal.
You might argue that this approach is okay for very large companies that can afford a staff of data analysts and computer programmers but is problematical for smaller companies. These companies cannot afford to have the needed expertise on tap, in house. But there are plenty of companies that offer these services for all size and shape of customer. Online communications allow for data to be transmitted, analysed, and presented almost immediately in an understandable format, highlighting the important matters. That allows managers to concentrate their time on things that need paying attention to.
There is also one other ingredient that is needed to successfully implement a data driven approach: the right culture. If companies are to gain the benefits that data analytics can bring, they need to embrace the idea that data analysis will not always give you the answer that you expected and certainly may point to one that is contrary to gut thinking or intuition. Senior managers need to be willing to admit when they are wrong. Far from being a failing, this will help others in the company understand the value that is placed on data and its analysis.
I have worked on consulting assignments that have not had the support of senior management nor the culture to accept change. I have made my recommendations and returned to see how they are getting on the with implementation. Sometimes the experience has been a little like punching marshmallow: you punch, your fist goes in, you remove your fist and slowly the marshmallow returns to its original shape. Without buy in from senior management, people will go back to their old ways of doing things when the pressure to change is removed.
What does this reliance on big data and sophisticated analysis mean for people like me, consultants who bring an industry-wide view, analytical thinking, and understanding of best practise to apply to the circumstances at hand? In a way those approaches can be summed up by calling them intuition. Does this mean that we are fast becoming obsolete, replaced by big data? Far from it! (You would expect me to say that, of course.) What we bring is an understanding of context and the ability to identify the right questions. As Pablo Picasso said of calculating machines, “But they are useless. They can only give you answers”.
