Digital Strategy No 4: Data, analytics and the big but

April 2016

It was the great St. Thomas More who first referred, in 1532, to the problem of finding a needle in a haystack (though he said ‘meadow’). Nearly five hundred years later, whilst still much quoted, it isn’t generally a problem we have anymore. Because we have powerful magnets, in the form of data, algorithms and computers. To find a needle today you need to first hypothesise what type it is, and then decide, and acquire, the data that you can interrogate that will lead to it. A more modern saying comes from Google: ‘in God we trust, everyone else bring data’.

Data is though a troubling mistress: it can represent correlation or causation. Knowing the difference is a skill worth learning and a source of strong competitive advantage. Which is why understanding data is such a key part of any digital strategy.

So how do you use data to boost your business? First off you need to be alert to its importance; once you set your mind to thinking ‘could ABC data help me achieve XYZ outcome’ you will start to be aware of all the data that surrounds your day to day activities. Instead of bemoaning that we are all drowning in data you will start to think of it as fuel, something that can power something else. And as the use of data permeates our lives we are naturally becoming more attuned to what is possible. For example, who doesn’t just ‘look it up on Google’ when they need to know something, or check their train is on time via an app, or order that book that someone is talking about on the radio as you lie on the sofa.

It is data that is removing friction from our lives, enabling us to spend more time doing something than planning to do something. And in a nutshell, for businesses, that is the second point to address; what are our customers after that data could help us provide in a more frictionless way? How do we make this service, product, interaction, building, city smarter? And by smarter we mean more attuned to the needs of the consumer, not the provider. Well today, one of the key ways is through data.

How do you make this happen:

1. Start with an hypothesis. If we knew X would we be able to predict Y? Or rather, could we say Y is 80% likely to occur if X happens? Because computer prediction is actually about probabilities rather than forecasting the future, and whilst not perfect, high probabilities can save you an awful lot of time and money.

So, in the context of real estate, one might ask: ‘when will this homeowner move’, ‘when will this company need a bigger office?’, ‘what asset would this investor buy?’, ‘how much will this building cost to run?’.

2. Once you have your hypothesis you need to define the metrics you would need to answer the question? So you might ask ‘what is the average time someone in that street stays in their home’ or ‘show me every asset that this investor, or investors like them, have bought in the past’ or ‘show me historical costs for a similar building, with similar tenants’.

Nobel Laureate Daniel Kahneman has told us all about ‘Thinking, Fast and Slow’ and the relevance to business is great. Citing real estate again, the industry loves the notion of ‘Fast’ thinking, which is really our instinctive reaction to any situation, our gut if you like. ‘We are a people business’ is symptomatic of ‘Fast’ thinking; through our networks we ‘just know’ the right person for that space, or the perfect investor for that building. Such ‘knowledge’ is indeed a valuable asset, but the future belongs, I think, to more ‘Slow’ thinking. Being asked what is 2+2 is an easy thing for our fast thinking mind to deal with but what is 28,489.2 times 6,870.306 requires the different skill set embodied in our slow thinking mind. In real estate, we need to be looking at how greater use of data, and analysis, can enable us to answer more questions, that are more complicated, and provide answers that are more optimally attuned to real needs, desires and aspirations.

With better data, we can define better metrics, and with better metrics we can do a better job, both for ourselves and our customers. Hunches have their place, but now that we have more data, more sophisticated algorithms, and faster computers, we should be making use of them.

3. Ask the right question, use the right metrics to answer, and you should have an outcome of insights that are valid and verifiable. These of course may well disprove your hypothesis, in which case you can move on to test another. Insights that lead to you not doing something can of course be just as valuable as those that lead to action.

This stage is a prime example of why ‘the robots’ will not take all our jobs, because it is for you to discern whether your insight really is causation, or just correlation. Have you the data you need, and enough of it, to take this answer as final? Do you believe the answer? Does it surprise you?

It is an absolute truth that data is the great enabler of the modern world. Without data all the computational power at our disposal, all the machine and deep learning algorithms at our behest, are worthless. The more you have the better, but there is a big but. It is still our brains that need exercising in asking the right questions, and judging the final validity of what the machines are telling us.

The world is moving from running on oil to running on data, and knowing where data fits in your digital strategy is up near the top of ‘must have’ skills.