Home > Columns > Executive Interviews

Chattermill Executive Interview

 Click the image below to download the ebook

Sheri Greenhaus, Managing Partner, CrmXchange, and Mikhail Dubov, Founder and CEO, Chattermill, discuss the benefits of Chattermill’s Unified Customer Intelligence platform. 

Please tell us about Chattermill.

Chattermill is a Unified Customer Intelligence platform, and we’re a global company headquartered in London – though increasingly, many of our customers are based out of the US.

Using machine learning, we help brands unify all types of customer experience data in one place and then extract critical intelligence from that data. The goal is to help companies increase things like loyalty and retention by actively listening to and learning from customers. Or as we like to call it, unlocking the customer reality.

We are specifically focused on unstructured data. Many companies analyze customer feedback and customer conversations. However, with traditional tools, this can prove difficult and time consuming. This is why machine learning and natural language analysis are so important.

Does Chattermill aggregate all the data and present the data in different formats? 

Most companies today have a ton of feedback. They have survey platforms, social media, email, forums, research, phone calls and so on – generating many different types of customer data. The goal for us is to unify all this diverse data under one roof. There are so many insights buried and waiting to be uncovered in these millions of data points. But all of this data can’t be analyzed with simple tools such as Excel. Tools like Chattermill are needed to find relevant insights, to find root causes, or even find interesting patterns.

Does a company need to know what they’re looking for or does your technology identify an issue and present the information?  

We serve both use cases. Sometimes the company has very specific questions that they want to answer, but most of the time, we build a program for the company that works across many different use cases. Quite often the case is to figure out what’s going on. There may be a question about metrics; why has my net promoter score fallen from May to June for example. They may want to be able to figure out how to improve the CX in a particular geographic area or how a customer is feeling about the checkout process. Whatever they’re looking for, Chattermill helps them find it.

Do companies have the ability to search for broad questions such as ‘how do I increase my customers’ experience’, or does the question need to be finer?

Typically, when it’s important to look across all aspects of the customer experience, the person responsible for the overall customer experience will need a birds-eye view. However, a manager in customer service may want to focus on agent empathy in a geographic area. Ultimately, we want to serve all of the important stakeholders in the company. But to answer this question directly, we cover both angles.

Do companies use their own survey tools?

Our customers would typically use another platform or do it themselves.

Do you think companies ask the right questions to get at the information they are looking for? 

The industry has moved in a positive direction over the last 20 years. Responses to surveys have massively dropped. Around five years ago a 1% response rate in a survey was considered a huge success.

Companies have now shifted to methodologies like net promoter score, where only one question is asked.  This net promoter score methodology produces a much better response rate. We tell people to keep their questions as generic as possible and keep them simple.

Machine learning does the hard job of finding insight and intelligence in that data. Provided there is enough volume, keeping the question generic and open-ended allows machine learning to cover everything, and for the customer to speak in their own voice. Consumers do not have the patience to respond to lengthy surveys anymore.

Some companies try to be too creative with questions and they go down a completely wrong path. A generic survey where the customer can ‘talk’ about what is important to them, provides much better information as it captures exactly what someone says. Machine learning pulls out what they are saying and looks for patterns. And from this, we can generate customer intelligence.

Does the technology look for patterns in the aggregate by whatever metrics the company wants to focus on?

Exactly. The issue is most companies don’t have the tools to analyze the data. It’s looking overall at the aggregate for certain words or phrases.

Companies collect so much data – can that cause them to be overwhelmed?

Yes. Most companies are already inundated by data and it’s getting worse. The average major brand today will have dozens of customer experience tools spread across different departments and different stakeholders. The number one problem is that you are only getting a slice of what is occurring because you can’t see everything. We unify all the data in one place and make it simple to work with. We bring attention to the right data.

We also do what is called anomaly detection. For example, let’s say that sentiment jumped 30 points or dropped 30 points during one month in a specific country. We draw attention to those types of fluctuations so the issue can be fixed quickly. We can help the company understand why something happened, rather than what happened.

From what you have seen, does how the company think the customer experience is matched to how the customer views their experience? 

It's very rare to see a company completely misguided about what the customers think about them. They typically have a decent understanding because even simple metrics like net promoter score give you a decent enough idea. Being objective is important. Everyone must be using the same data and the same methodology. But by listening to the customer more, you’ll always be able to build experiences that match customer expectations.

What type of companies benefit from your solution? 

We work with companies who think that their customers’ experience is very important. CX is a competitive advantage today. While they agree CX is important, these companies want to improve the customer experience on metrics like net promoter score. They may realize that something is not right and while they may have theories as to what the issues are, they don’t have definitive answers. These companies are looking for proof and data points of what they need to invest in. Chattermill helps them find it.

What do you say about companies that have tried this type of analysis before?

Companies may have tried this type of analysis 5 years ago and they received mediocre results. It wasn’t worth the money or the effort. The advancement in natural language processing and machine learning and what can be done with data has expanded dramatically. The quality of analysis and the insight you can get now from the data in real-time is dramatically different compared to what was possible only a few years ago. This is especially true for customer support data which tends to be very complicated. There is a lot of ‘noise’ in the data and the agent and customer data tends to be in very complex structures.

Just a few years ago it was impossible to access and analyze data to the degree that is needed to make a change in the customers’ experience. Unfortunately, many people are not aware of what tools that utilize natural language processing and machine learning can do. My advice is to try these new tools to see what they can accomplish. Because once you’ve opened your eyes to the customer reality, it will be hard to look back.