Daniel Wigley

Managing Director

thought leadership
Demystifying Machine Learning

One of the biggest challenges for business leaders in 2020 is how, in an era of digital disruption, we stay connected with customers and remain a relevant, essential part of their everyday lives.

Most companies hold gigabytes of information on what they sell and who they sell to, so much in fact, that it is impossible for any individual human being to take it all in. The challenge is finding the right tool to harness that data to make better informed decisions that power your business.

The answer lies in machine learning, which finds patterns in customer data to predict what is going to happen in the future. This is the very tech that drives the profits of Netflix and Amazon, allowing them to both learn from and sell to their billions of customers a set of individual services tailored to them.

Giving the customer what they want is basically a model that has existed for hundreds of years. A corner shop owner would personally know every one of its customers and make recommendations based on that knowledge. Machine learning is simply doing the same thing, only at a speed, scale and accuracy that cannot be matched by human beings.

One of the great myths about machine learning is that it is out of the reach of your average Small-Medium Size Enterprise (SME). In fact,  it is an inexpensive, simple process that takes weeks rather than months to implement and rewards businesses with a transparent return on investment through improved sales and customer experience. 

It should be the first port of call in digital marketing. The sheer waste of targetting the wrong person with the wrong product, runs into hundreds of millions in the UK.

The biggest barrier to harness the power of your data is not the price or technology but culture. It is no coincidence that the kings of machine learning, Amazon and Netflix, did not have to overcome the baggage of tradition that exist in a 150-year-old business.

Yes, it takes a leap of faith for leaders to behave like a start-up and start with a clean sheet. Embrace ‘zero-based design’ to reassess how your operating model is configured around changing consumer needs. 

However the first steps on that journey are relatively straightforward:

  1. Great data on your customers. The starting point is to connect clean data sets such as Google Analytics and Salesforce, something many businesses already have.
  2. What’s the issue? Work out what business problem you are asking machine learning to solve. For instance it could be identifying customers with propensity to buy a certain product, or working out the lifetime value of a customer.
  3. Start small. Don’t overcomplicate. Once you’ve got something working and plugged into your business infrastructure you can prove the value of that machine learning solution then reinvest and continually improve. 


Machine learning shifts the focus of businesses from just hitting that month’s sales targets to becoming truly customer first. Yes, it is new technology but it adheres to the time-honoured principle that has driven profits for businesses of all sizes, from tech giants to corner shops - that of putting the customer at the heart of everything we do.