Sam Daish, Head of AI & Data Science at Qrious, lets us know how to make the most of unstructured data


As marketers you’ve all been building your data powered marketing approach, but one type of data has remained largely untapped – unstructured data. IDG is predicting that 93% of global data will be unstructured by 2022. Whether it’s social reviews, customer service voice recordings, chats, or something else, there’s a huge opportunity for marketers to use this data to gain deep insights and enhance the customer experience.

But how do you tackle unstructured data? And why would it be useful for marketers?

My suggestion is that you run your unstructured data initiatives in parallel with your current initiatives that are using more structured data, such as from your databases. The combination of intelligence that you gain from both types of data will supercharge your customer insight and your ability to deliver enhanced customer experiences.

Here are my top three tips for taking advantage of unstructured data:

1. Find the right problem to solve

Understanding what the customer really wants can be elusive. They might tell you one thing in customer surveys, but your results are telling you something else. This is where unstructured data can help you fill that gap. I recommend breaking the process down into three key areas – capture, analyse, and execute.

Capture

You need to tap into the data sources where your customers are actually talking to you. What sources of human expression do you have access to? For example, can you access the data from social media comments, or call centre notes, emails or chats? Start with one of those data sets you can access.

Analyse

To decide what to analyse, consider a challenge that you’re experiencing or a question you need answered. For example, you might want to understand what the biggest frustrations customers have with their subscription. You’ll want to apply sentiment analysis to uncover how customers are feeling towards your product and topic analysis to uncover what their most common complaints are.

Execute

In the analyse stage, you’ll likely uncover a number of areas where you can improve the customer experience. It’ll be tempting to try and fix them all! But start with the one you think will have the biggest impact and go from there. For example, when we were working with a client on topic analysis from their contact centre data, we uncovered that there were certain issues that required mitigation within only a couple of days before the customer would cancel their account, but others where mitigation within a couple of weeks was fine.

Once you’ve gone through this process you can keep testing and iterating to ensure you’re hitting the mark and making a positive impact on the customer experience.

2. Connect the dots

It’s important to remember that the role of delivering outstanding customer service doesn’t just sit with marketing - it’s the responsibility of several departments across the organisation.

I’ve recently discovered a US-based beauty company, Glossier, who has managed to connect the dots between marketing, customer service, and product. Their customer service agents are integrated into the marketing team and the marketers are encouraged to listen to customers share their unique experiences and recommendations. They mine this data for key customer insight which gets fed back to marketing and product. By taking this customer-centric approach, in just five years the business has grown into a $1.2 billion business and has reportedly sold out products in 30 seconds. That sounds like success to me.

3. Embrace changing technology

The thing about unstructured data is that you need to approach it a little differently than your structured data. It takes different technologies and new methodologies to analyse it and gain useful customer insight. This is where AI and machine learning can help. And it’s not as scary as you think! You’ll all be familiar with the term “data explosion” and maybe “data volume, velocity and variety”. There is now simply too much data for human cognitive processing. We need to let the machines do the heavy lifting to uncover that critical customer sentiment and key topics, to give you the insight you need to transform your customer experience.

Just like how social media changed the marketing game many years ago, AI is going to do the same. It’s time to take the plunge and embrace AI in your everyday marketing practice.

So, what unstructured data could your marketing department access? What burning customer insight questions do you want answered? And what will you do differently?

We’d love to hear what your customer data challenges are and how we can help.