The future of analytics will accelerate your growth. Here’s how

Happen’s Head of Analytics Todd McLoughlin has been working in the world of quantitative data research since 1995, evaluating innovation concepts for the likes of Unilever, Reckitt Benckiser and Kraft. He joined Happen in 2018 after following the consultancies’ emotion analytics work with StarMaker. StarMaker uses big data to seek out online and offline conversations relevant to a business challenge, identifying expressions of excitement or frustration that explain consumer behaviour.

Here, McLoughlin discusses the potential and challenges of emotional analytics, and the future of the field.


What’s different about Happen’s emotion analytics tool, StarMaker?

For years I’d been using survey data and I was starting to see the limitations, especially in context of some of the questions clients were asking. I became aware of Happen and StarMaker and saw that it not only had the social listening element, but could also be used to analyse offline data – customer service transcripts, feedback forms and alike. In addition, it did something different that showed me it was far more advanced. I joined Happen because I could see that StarMaker goes a distinct step further by identifying specific areas driving excitement and frustration. That’s not the same as liking or disliking something. Innovators or marketers need to understand what pushes or pulls a consumer to purchase, as much as possible.

People talk about what they want online in a very clear way, and what we’ve discovered is that if you take the time to listen, you can fast forward the entire innovation process. It’s about getting to the right idea, first time.

How has the analytics market changed in the past five years?

Five years ago the gaps in knowledge, when it came to quantitative surveys, were increasing. People had busier lives, and that meant they had less time to do in-depth surveys. The industry responded by using shorter questionnaires with less depth, or accessing new and different pieces of data such as photos or videos. Forty-minute surveys can tell you how people respond to a category, but after that amount of time a lot of people probably mentally tap out. It meant lower quality data was being produced, and big data started to add value. This is because, unlike in some surveys, online data is not biased or guided commentary. It’s spontaneous and unsolicited, which makes it a really rich place to look for the answers people need.

"Getting data is one part of it – cleansing the data is probably even more important. If your data isn’t clean it will lead you down the wrong path."

Todd McLoughlin, Head of Analytics at Happen

What should anyone new to social listening be aware of?

You can’t assume that social listening is an exact replica for the type of sample you get from the general population when doing quantitative research. That’s because people commenting on a product online are buyers, and they’re more invested. They’ve gone online to share an opinion and are a little bit more engaged in the category. They want to be heard. In general, they are early adopters in a specific category. Demographically, things are pretty consistent though – everybody is involved in online conversations in some shape or form. Sometimes clients ask us if there’s enough data out there for their category, and of course we don’t just have to use social data – it could be from customer care lines or data from social communications they might be running. If I think about pure consumer commentary online, we have even run projects on frozen peas – and if we can do it on frozen peas, rest assured we can do it on any category.

As with any type of research, there are also a lot of comments that maybe aren’t quite so pure, or spontaneous. That’s why it’s important to take out the distractions to get to the voices that are pure consumer commentary. Getting data is one part of it – cleansing the data is probably even more important. If your data isn’t clean it will lead you down the wrong path.

How do you block out that background noise?

As the content out there increases, this will get harder. Right now StarMaker can eliminate multiple comments, or comments with similar wording which are clearly planted, be they from PRs or advertisers themselves. However, bots will grow in usage and there will have to be new approaches in the future.

What advice would you give a company that has never used analytics to make decisions?

Ask a lot of questions, and get comfortable with it. Learn about the depth of experience of the suppliers you are talking to, and the big data approaches they are using to see if it reflects your immediate needs. Talk to outside people to see if the suppliers have a good breadth of experience.

How are brands using StarMaker?

We started out by applying the insights we have around what people say and what people hear to category understanding and innovation development. We eventually saw there was amazing value it could add to tracking consumer reactions to new product launches, so our clients could course correct their communications early, and even get ideas for their next innovation based on what consumers were saying online.

Now, we are starting to leverage StarMaker to track consumers’ connection to a brand. We get thousands of online and offline comments for a category that provide a 360-degree, holistic view. We then develop a baseline for that emotional connection, breaking it down by brand, or even sub-brand for companies with large ranges. This gives us a great view of the companies that have the most or least ‘brand pull’, and why. You can even start to spot the small emerging brands that are starting to make waves in a category – which is critical for our clients since some of these small players end up becoming tomorrow’s toughest competitors.

Will we get to a point where machines do it all?

Human curation and creativity will always be part of process – but that doesn’t mean we can’t make it faster. At Happen we’re seeing if we can use past analyses to help inform future analysis, to get to answers faster. For instance, if we ran a StarMaker project looking at ice cream, and six or seven themes came up as being critical, we could start with those first the next time we looked into a category. By doing that we can create an AI system based on what we know is valuable.


Like what you see? Get in touch to see what StarMaker can do for you: