When it comes to content, at Codehouse we come across two kinds of customers. Those that have been struggling to produce, relevant, high quality content. And those that have a treasure trove, but find that it’s hidden and scattered across their digital brand estate.

Whilst we love working with customers on the first type of problem, as developing a good content strategy and road-map for delivery is at the heart of customer engagement, working with a Library of Alexandria’s worth of content stretches our capabilities in a different way.

Drawing the dots

A professional services firm recently approached us with an interesting problem – the fantastic content their subject matter experts produced was getting lost on their websites.

 Despite having only one brand, they had created more than a dozen different websites over the last decade.

Looking at the brand and its standing in the market, it was clear that they had an opportunity to better use content to create deeper relationships with their existing and potential clients.

The challenge was how to surface the right content to their website visitors, at the right time? For example, how might they present contextually-appropriate content about GDPR or Brexit, both very important issues in professional services, to a website visitor over time?

This presented us with two questions:

  1. How could we meaningfully bring together the disparate pools of content from all of the company’s different websites?
  2. How could we drive powerful content recommendations to existing and potential clients across different marketing channels?

The volume of content the professional services firm had, meant that the overhead of an actual person manually going through all the content and thematically tagging it all was impossible. We needed the help of some very cool technology.

To approach the challenge, we worked with Lateral - one of our machine learning technology vendor partners.

With Lateral’s natural language processing (NLP) algorithm, we were able to run a proof of concept that pulled in over a thousand articles from nine of their different websites. Their algorithm is able to “read” the content making sense of its meaning much like a human reader would, and so join the semantic dots between the articles.

 For the first time ever, the firm was able to see a large swathe of their content in one place.

To test the algorithm, we created thematic slices of the content by running typical user searches and reviewing the quality of the results manually.

With this technical proof of concept, we demonstrated that large volumes of content could be analysed and clustered together with greatly reduced human effort. At the same time, we wanted to show the client’s marketing team the potential benefits of applying machine learning to help out with some of the things they do and, critically, prove how we could transform the customer experience using all the valuable content they create.

Letting content build the relationship

Any marketer who has fought to implement personalisation on their website will be familiar with the challenges and time investment required in consistently taxonomically marking up content and manually configuring persona profiles, even when using a best-in-class tool.

Configuring a persona profile in Sitecore 
Configuring a persona profile (example from Sitecore)

With Lateral’s clever machine learning technology, the reliance on tagging content manually to drive content recommendations is diminished. The tool can also generate tag suggestions should the marketer wish to include them.

However, what is often forgotten when implementing machine learning tools is that they will not work perfectly ‘out of the box’. The algorithms improve recommendation performance through increased usage and direct training. Basically, the more you use machine learning tools, and the better the quality of data they receive, the better they will get, and the more they will help you.

 Designing personalised customer journeys requires understanding the full shape of the desired end-to-end customer experience.

But how do you design a personalised content journey when the nature of your content is rapidly evolving, such as with Brexit?

Our solution was to enable the brand's customers to choose to subscribe to a topic as broad as ‘Brexit’ or as specific as ‘Brexit’s impact on Asian pharmaceutical companies’. Powered by Lateral and an email marketing tool, we could generate highly personalised regular thought leadership newsletters, and even alert emails that could be sent to a customer the moment an article is published.

The best part of this solution is that the quality of the content recommendations and customer engagement is tracked and drives the improvement of the machine learning algorithm over time.

And all of this is done without any direct input by the marketers.

Leading from the front

Until recently it was only the most sophisticated publishing companies who could conceive of, or could afford to, design and build this level of personalisation using the content at their fingertips.

Now, by combining the latest machine learning techniques with best-in-breed digital marketing tools, any brand in any sector can develop closer relationships with their existing and potential clients by putting the rich, thought-leadership content they produce in front of people’s eyes at the right time, in the right place.

Want to learn more?

If you are interested in exploring how machine learning could be used to get your content in front of customers and prospects in a better way, our team of digital experience consultants and technology architects will be very happy to help. Get in touch today.