our views and our knowledge in analytics and other releveant topics

our blogs

How machine learning is reshaping marketing


Whether you’re an analyst, marketer or even an accountant, advancements in technology are changing the way that we work and live. As a marketer, it’s absolutely vital to have an understanding of modern technologies and the potential they have to improve the quality and efficiency of your work.

For a few years now, machine learning has been a regular feature in almost every “Top Marketing Trends of the Year” article but this hasn’t yet resulted in droves of marketers using sophisticated machine learning and artificial intelligence tools and methods in their strategies.

So, what does machine learning mean for marketing?

When it comes to refining and improving your marketing strategy, the data you already have about your customers will always be the main driver. The easiest way to think about machine learning is that it’s an extension of this with the focus of predictive analytics to identify possible future outcomes.

Although it’s easy to ignore the technologies that we’re not currently directly using, in reality, machine learning is probably present in a lot of the marketing tools you use on a daily basis. Here are just a few of the headlines about popular marketing tools implementing machine learning features:

  • “MailChimp Brings Automated, Data-Driven Product Recommendations to Small Online Stores”

  • “Google Launches Machine Learning-Powered Retail Ad Platform”

  • “HootSuite machine learning app launched to help businesses tailor social content”

Even if you’re not ready for machine learning right now, being aware of what’s possible and how you could implement it in the future could help you get ahead of your competitors.

Predicting conversions

Using historical data is the best way to accurately predict a website visitor’s next move in their online journey. Station10 recently partnered with Decibel’s digital experience intelligence platform to help leading tourism group TUI to better understand the behaviour driving their customers’ digital experience.

We used TUI’s web analytics data paired with data from Decibel’s Digital Experience Score (a framework that measures customer experience through  factors including navigation behaviour, frustration and form experience). The ultimate result, highlighted in the report (link to report) was a measurement framework that enabled TUI to predict £20 million in revenue growth.

You can download a copy of the report here.

Predicting customer turnover

Customer turnover or churn measures the number of customers who stop buying or subscribing to a business. Customer turnover is an important measure of buyer satisfaction and being able to predict it will help minimise it and place focus on increasing satisfaction.

By combining new behavioural data with historical data to uncover trends in user behaviour, it’s possible to create a framework to predict when a customer is likely to stop engaging with a business.

  • When was the last time they logged in?

  • When was their last purchase?

  • Has their purchase value or frequency dropped?

For example, let’s say that by analysing customer data you find a trend where someone often logs in and browses products three times a month and purchases once. This goes on for a while but then suddenly changes. Now, they’re only browsing once a month and purchasing even less than this. This should ring alarm bells that marketing strategy should change to re-engage the almost lost customer. Machine learning helps you do this at scale.


As one of our recent blog posts explores, personalisation and recommendations are becoming an expectation of customers and a necessity for businesses. In addition to recommendations on e-commerce stores, machine learning makes it possible to go beyond rule-based personalisations to offer customers a unique advertising and content experience.

Algorithms can predict which type of content will perform best with each visitor on your website. It’s possible for two people to view the same website at the same time and see completely different content. Netflix is a great example - every time you browse shows and click play or pause, Netflix is gathering data on your choices. Netflix then uses this data not only to inform the content recommendations it serves a user, but even to decide which artwork to display to promote content.


A Zendesk study showed that live chat delivers customer satisfaction levels of over 90%. That’s higher than for phone support or support given through any social media channel. On top of this, studies have shown that more than half of customers are more likely to return to a website if it offers support via live chat.

On the other hand, the problem with live chat is that it’s time consuming. It’s a huge task for any business to manage. This is where chatbots come in, a machine learning technology that imitates human interaction. Chatbots allow you to offer live chat on your website without needing the presence of human support operatives.

Chatbots are getting smarter and smarter and are being taken advantage of by huge companies like Domino’s and H&M. Machine learning features include:

  • Sentiment analysis to judge the mood of a customer

  • The ability for chatbots to pull data from social media to learn more about a customer

  • Chatbots can feed data into other channels to deliver personalised marketing


Machine learning is becoming more and more commonplace within both marketing tools and strategies and that’s not going to change. Although most marketers probably aren’t ready to throw out their existing techniques and proven strategies to adopt ML, it doesn’t mean they can’t find a competitive advantage by delving a little deeper into the world of artificial intelligence.