A tough (Wal)nut to crack?
Ian Bobbett shares his experience of bridging the gap between data analytics
and market research with sister agency, Walnut UNLIMITED
Analysts and data scientists who dismiss research are missing the point – there is huge potential to add richness and depth through the amalgamation of disciplines.
Realise has been part of Unlimited for more than a year now and time has certainly flown. It has been an amazing journey, but one of the best things that has happened is the ability to scratch a long-standing itch by working with the wonderful folks at Walnut Unlimited. That ‘itch’ is in part based on my early career where I started out in market research before moving to the dark arts of analytics and data science. But it is probably better encapsulated by a conversation that I had some years back. Here is a (heavily paraphrased, partly redacted and certainly warped by time) flow of the conversation:
Researcher: “We have done some research and discovered that users of our online platform averaged almost 12 visits over the last three months. Can you please validate that?”
Me [Analyst]: “Of course.” and a short time later “We’ve looked at our data and our users have only used the platform 4.8 times on average over the last three months.”
Researcher: “How can that be – we’ve asked our users? We know that there is always some error but not that much as it averages out.”
Analyst: “We’re looking at this the wrong way. Both datasets are correct but they are looking at different truths – one is behaviour and one is perception. But this is exciting – we know that usage drives revenue for us. Maybe they believe they are visiting our site more frequently because they feel guilty about overuse. If we put a counter on the site, they could see they are using it less than they think, which could have a significant positive revenue impact for us. And we can look at the visits that they are remembering as being recent but were a long time ago – are there certain interactions that happened on the site that resonated with users that we should try to amplify or repeat?”
Researcher: “I think your data is wrong.”
Researcher: “We have done some research and discovered that users of our online platform averaged almost 12 visits over the last three months. Can you please validate that?”
Me [Analyst]: “Of course.” and a short time later “We’ve looked at our data and our users have only used the platform 4.8 times on average over the last three months.”
Researcher: “How can that be – we’ve asked our users? We know that there is always some error but not that much as it averages out.”
Analyst: “We’re looking at this the wrong way. Both datasets are correct but they are looking at different truths – one is behaviour and one is perception. But this is exciting – we know that usage drives revenue for us. Maybe they believe they are visiting our site more frequently because they feel guilty about overuse. If we put a counter on the site, they could see they are using it less than they think, which could have a significant positive revenue impact for us. And we can look at the visits that they are remembering as being recent but were a long time ago – are there certain interactions that happened on the site that resonated with users that we should try to amplify or repeat?”
Researcher: “I think your data is wrong.”
There has long been a disconnect between research and analytics – when research doesn’t tie back to hard metrics, there is a perception that it is somehow wrong. But to me, this information is a critical counterpoint to behavioural data. Data science has long been able to tell us the what, when and how things happen – and that information is getting wider, more granular and a lot more interesting. But we’ve never known the why. We can often infer, but we don’t know until we ask – which is the power that market research brings.
Thankfully, that trend has been changing over recent years as data scientists and analysts embrace more diverse data sources (research included) and I think there is a genuine opportunity to bring these disciplines even closer together to better understand customer emotion, motivation and action in a holistic solution.
Which brings me in a full circle. Working with Walnut is scratching my itch by bringing together the what and the why. And it is leading to some exciting propositions and products – hybrid segmentations to name just one. Identifying the reasons and opportunities between behavioural data and research data has led us to our perceived-reality solution and there is plenty more to come. It’s been a great journey – long may it continue.
Which brings me in a full circle. Working with Walnut is scratching my itch by bringing together the what and the why. And it is leading to some exciting propositions and products – hybrid segmentations to name just one. Identifying the reasons and opportunities between behavioural data and research data has led us to our perceived-reality solution and there is plenty more to come. It’s been a great journey – long may it continue.