Data and the customer experience: getting the best out of big data

Correctly leveraged, big data can offer huge benefits to any business, but where to start? We spoke to ESCP Professor Louis-David Benyayer and Glanceable co-founder and CEO Arthur Cohen about the potential – and pitfalls – of big data when applied to customer experience.

Louis-David, you have been the co-scientific director of ESCP’s MSc in Big Data and Business Analytics since 2018. To start off, can you offer a closer look at what big data looks like in a business context?

Unsurprisingly, big data is a big concept! First and foremost, we can divide big data into two main types: intentional and unintentional data. Intentional data is created when we as consumers undertake a specific action: browsing the web, commenting on social media, ordering online…

In contrast, unintentional data is gleaned as a byproduct of our digital behaviors, such as our geolocation, buying habits or choice of language.

We can further distinguish between human and machine data, which is another difference that’s beginning to impact the world of business. Previously, it was mainly human interpretation of data that drove decision making.

Today, machine-obtained and machine-deployed data is everywhere, meaning that – crucially – data is being captured and redeployed in real time.

Examples can range from automatic pricing algorithms, which continually adjust prices in response to market demand, to modern technology such as Uber, whose method of matching riders and drivers has now fully superseded what used to be a manual method of interpreting data.

Arthur, can you tell us what sparked the idea to create Glanceable?

Like Louis-David mentioned, the idea to create Glanceable stemmed from a problem we experience every day – how to bridge the gap between manual and automatic data interpretation in an age of information overload.

Unlike a generation or two ago, when one still had to put in real effort to find the information one was looking for, today, information is at your fingertips. The challenge now is filtering through it to extract the information that matters to you.

Take Deliveroo, the British food delivery business, for example. As a consumer, I’m offered a multitude of restaurants, all of which I must sift through to finally pick the one I want. How do I choose? I look at the feedback to make up my mind on where I want to eat, but unfortunately there is so much feedback that I spend a huge amount of time making a decision.

This begs the question: how much time and resources are lost by businesses attempting to filter through information to get to what matters? This problem, coupled with our passion for artificial intelligence, led us to create Glanceable, which uses artificial intelligence (more specifically natural language processing) to summarise large volumes of data into key actionable insights. 

Unlike a generation or two ago, when one still had to put in real effort to find the information one was looking for, today, information is at your fingertips. The challenge now is filtering through it to extract the information that matters to you.

Arthur Cohen

In light of this, can you explain how companies concretely use big data today?

Arthur: In terms of benefits, our clients use big data to better understand their customers, which in turn improves brand recognition and the customer journey both offline and online. Our clients also use big data as part of competition benchmarking and, in particular, CRM management by acquiring and leveraging data in real time. Glanceable starts off by centralizing all data from multiple sources such as Trustpilot, Twitter, Google My Business, and our client’s internal data (surveys, online forms, etc…). Once the data is on the platform our AI automatically summarises this information into key points for easy action – the best of human and machine data management, one could say.

Louis-David: To take Arthur’s point about the customer experience a little further, personalisation is another way in which big data can be of huge benefit to companies. For example, this could involve adjusting marketing messages according to a consumer’s personality and buying habits. With big data, this is done at scale, automatically – hundreds of A/B testing email messages can be generated and tested, fully exploiting the available data. Another way is using data to predict rather than to respond. An obvious example is estimating production volumes for a given market or anticipating trends and optimizing customer targeting to convert prospects into new consumers.

Arthur, from a client-facing point of view, could you share an example of how the big data process leads to tangible change?

Arthur: Of course. First, we ask our client to manually build a business overview depending on their structure and priorities – this could be region-based, product-based, brand-based… After integrating their data sources, they’ll be able to instantly see common themes running across their data points. If we take the example of a restaurant, they may see, for example, customer feedback relating to poor sanitation or COVID protocols. The overall rating of those locations goes down from 4/5 to 3.25/5, leading to reduced footfall. With Glanceable, they’re able to detect the problem early and implement a sanitation strategy, which they can also plug into Glanceable and monitor directly.

What are some of the other immediate benefits when it comes to using big data? And conversely, what are some of the biggest pitfalls and challenges?

Louis-David: Continuing on from the previous questions, in terms of benefits for companies, big data can also be leveraged in ways that aren’t immediately visible to the consumer. For example, in the fashion industry, big data drives logistics and delivery, trend forecasting, and manufacturing. Even something as seemingly simple as next-day delivery and other tangible customer benefits are made possible by data management behind the scenes. 

In terms of pitfalls, it’s worth remembering that big data also has its limits. YouTube is a great example – its user recommendation algorithm has evolved over the years to the point where many users feel it offers too much of the same content, thereby failing to renew the user experience. From a more technical standpoint, data storage seems like an obvious problem, particularly as the Cloud expands – and of course, the privacy concerns that that entails. 

In the fashion industry, big data drives logistics and delivery, trend forecasting, and manufacturing. Even something as seemingly simple as next-day delivery and other tangible customer benefits are made possible by data management behind the scenes.

Louis-David Benyayer

Arthur: There are several challenges one faces when dealing with big data. Louis-David is quite correct when he mentions data storage and privacy concerns, not just for consumers but also in a B2B context. Some datasets may only be used academically, some for research and some for commercial purposes, therefore doing your due diligence is very important. Aside from that, the key to generating quality insights is via quality data. In our case, we “clean” our datasets to make sure our insights are at a high standard, which is often costly (and tedious) if you do not have an efficient/automated system in place. 

In terms of technical challenges, at Glanceable we’ve certainly had to adjust things based on client feedback in order to better interpret their datasets. When we launched, we had initially built a very analytical platform, but our clients wanted to be able to act and strategise from the platform directly.

We realised that the main selling point of the platform is not the analysis itself but how one can act on the information on the platform.

From that information, we built the event tracker (to track the performance of your internal strategies on customer perception), the live feed (to respond to all feedback from one place), calendar integration and more.

I’d expect big data to be combined with human-driven interpretation moving forward, particularly in industries where brand perception is important, such as fashion and technology.

Louis-David Benyayer

Finally, Louis-David, can you share your thoughts about the future of big data?

Louis-David: I’m interested to see where big data goes in the years following the pandemic. In the first few months of Covid-19, the usual data prediction models became totally obsolete – we regressed to purely descriptive analytics, not reactive or predictive. Without a doubt, big data has facilitated the move towards the subscription-based service models we’re seeing currently, although it’s far from being the only factor. As a result, I’d expect big data to be combined with human-driven interpretation moving forward, particularly in industries where brand perception is important, such as fashion and technology.