Using Digital Behavioural Data for Insights

Click here to access the slides

Click here to access the slides


Transcript of recording with Mike Brown – generated automatically by HappyScribe which means it will be about 80% accurate – if you spot confusing errors, please email The timestamps are included to help you jump directly to a point of interest.


[00:00:08.840] – Mike Brown

Hi there, I’m Mike Brown of Green Bird Consultants, and with a career spanning 35 years in market research, I’ve worked in a range of large research agencies as well as 12 years sky in a mix of production strategy and sales roles. The route I’ve taken is to operations from data collection by telephone, face to face and online into managing, cleaning and processing reporting data for insights. More recently at Cantal, my responsibility was looking to the horizon or methodology and technology for insights from Quick Turn Survey Solutions, video chat bots, behavioural data.


[00:00:49.560] – Mike Brown

All were built and evaluated as being part of my role as director of technology and research in the Cantar Insights Division. I started Green Bird Consultancy in late 2019 to bring the range of skills to a wider set of clients focussing on utilising behavioural, passively collected data to benefit businesses and a personal waving the flag for location data. Of course, 20/20 didn’t help my early days when the world stopped moving around and sharing my own location within the lines of bedroom to bathroom, to kitchen, to home office to sitting room to well, you get the picture and you certainly live the same old as me.


[00:01:33.670] – Mike Brown

We did not need location based insights or did we? UK government introduced a lockdown, but not curfew. How effective was the messaging? One of my data and analytics partners could actually see how people are still moving around the UK real data showing distances, troubled times of day and locations rather than what people were telling research businesses through surveys. So although the world stopped as we know it didn’t. Anyway, that’s enough about me now I want to move on to the focus of this session.


[00:02:11.480] – Mike Brown

Let’s kick things off on setting boundaries so you’re all clear on what I’m talking about when I’m referring to passive or behavioural data. I’m not talking about harvesting data, which you might get from websites scraping or a pipe from, for example, a social media monitoring tool like Bramlage with that data. People are posting responses, comments, their opinions and attitudes. They’re actively sharing commentary between text, audio or video on the basis it will be available to view by the users the platform or a subscription community.


[00:02:46.970] – Mike Brown

In some ways, it is a different approach to directional quantitative research, where we pose questions and seek feedback for a survey. Much as I do believe in that approach for gathering opinions, today is not my time to be talking about these approaches. No, I’m talking about is data where we gather passively from users behaviour unencumbered to an extent by their opinions and biases. Whether it be the all surrounding world of acts, taxi hailing street Uber shopping through Amazon, consuming endless hours of comedy shows on Netflix, all the things we do with our phones, laptops and tablets can be gathered in a compliant and permission based way.


[00:03:31.950] – Mike Brown

As well as seeing activity, this could be search behaviour, websites visited and hovered around, essentially watching what people do without the need to sit next to him. And look over their shoulders, pumps are keen to do on the bus, although obviously this is slightly harder now with the spikes we are creating for each other through social distancing. And not only that, but we can also see what the is and very importantly, for insights where it is.


[00:04:00.330] – Mike Brown

And we know what information about the consumer exists in terms of demographics are to which can be overlayed. So if we have all this data available to us, how do we go about gathering it, managing it process and reporting it was several of you will be doing that already as you will have your own properties like a website and a. I’m hoping if you do have those in place, you’re getting the most out of the data you are collecting through those channels.


[00:04:31.030] – Mike Brown

However, that is within your ownership first party data collected from your own customers who are sharing their activity with you, like what they are putting in a virtual basket, what route they take to get to your contact details for feedback and complaints. How much time they’re spending on your property while using your products and even within your physical site. That is all good data and will help you understand your customers better. However, what about competitors? What happens when they never get away from your owned property, and how about the route they take to come back to your app or website?


[00:05:09.500] – Mike Brown

If I’m a quick term restaurant chain owner, how do I know what share of my customers time I’m getting and what share of revenue? Am I able to understand customers who might come to my physical site, but also order online through channels like delivery? And for those prospects who might be a more competitors set, how best to target more appetising to find them and persuade them to buy with me. Of course, traditional methodologies like surveys and qualitative techniques can help, but that relies on consumers being able to tell me about their lives within the constraints of a box we have to put together to gather the information.


[00:05:51.470] – Mike Brown

And remember in all those details that the routes taken to ordering from the restaurant, whether it be yours or your competitors, afraid to say that it’s not only hard, it’s not a worthwhile investment if your money. So if we want to research and insights need that can be better resolved by using behavioural data, where should we start? From looking at Mike Stevens website, Incyte Platforms in early June, a search for the word behavioural took me to 33 results in passive metering, behavioural tracking category from full blown behavioural panels down down into behavioural tracking technology views of Iot data.


[00:06:31.260] – Mike Brown

There were products, platforms and businesses to get your head around. However, this is thirty three results out of at the time, one thousand one hundred and sixteen in total across the insights framework. And gosh, if I type by tracking that produces 33 different solutions and what I would consider a niche area. So I would say that as much as we’ve been talking about using this behavioural data approaches for years of research is still to a major extent an uncovered opportunity if you can find the right key to unlocking sites that are often hard to gain from traditional quantitative and qualitative techniques.


[00:07:06.990] – Mike Brown

I tend to break down the sources and provide some insights into the space on the street and the different types of businesses all have their own degrees of excellence based on the length and breadth of data available, how it is collected in a compliant way, and importantly, what can be done with it in terms of informing insights. Making all this work in a lot of ways is like traditional quantitative work in that you have specialists in the participants screening them and getting them to install metering technology.


[00:07:37.600] – Mike Brown

Think here of consumer access panels like responding with survey panels, have a small proportion of people be metered. There are also panels set up purely as behavioural data businesses where participants do not get involved in what a survey taking tells, but can be asked questions specific to the data needed for the project. And with businesses here like measure protocol and maybe. Then military technology companies themselves are experts in the data tech gathering space like reality moyen. Beyond the normal horizon of market research, there were businesses like Bliss and Tomoko who take data feeds, had their own panels where people have agreed to share information by using an app, perhaps free with adverts free from adverts rather than paid for at free.


[00:08:27.220] – Mike Brown

And Homogenise and clean to get accurate estimates of location, that’s what blazin to do, and they create reports to demonstrate how consumers are behaving in the physical as well as digital space. Then mobile networks, they also work in a comparable way, taking customer data and enabling insights be copied from that base. In all cases, you end up with huge quantities of data, whether that be long so many thousands or millions of records or a wide data set hundreds of data points about an individual in their activity.


[00:09:03.580] – Mike Brown

So all in all, this is important to have a clear picture of where you’re aiming for. Why are you doing the research, what are the goals of the research, does that sound familiar? Just the same guidance would apply to any client looking to understand consumers. Here are some things to consider when approaching businesses in this space above and beyond what you might normally look for. So like countries or markets, I will talk to some of these questions now.


[00:09:33.910] – Mike Brown

So do you need servi alongside Pasic later? Well, if you already have a lot of opinion and actually data about your audience, choosing not to ask any further questions can make access to the Heidel data easier as it widen your choice of partner. Robert, how much data and how many users you date, as with any research project, try to be clear what is needed rather than ask for everything from as many people as you can? This will certainly help with your budgeting, not only for the sourcing of data, but also processing it and analysing it.


[00:10:07.090] – Mike Brown

And is the data or can it be historical? Some businesses will have volumes of data on a consumer audience, while others will be starting from scratch again, again, as with your traditional research needs, planned chronology set the timeline of what you need. Some further considerations, so browsers and search data, as I mentioned earlier, some technology will look historically. For instance, you might get a partner who can source search and browser activity by logging and can be consented to at a point in the survey.


[00:10:41.390] – Mike Brown

Be aware here, though, that although the technology might push past information, if you use as clear as search and set history, you may find very little activity. Then there is the whole engineering and management side of things. All in all, a long list of considerations when working in this space. My point is it’s worth it. Now I want to share some use cases with usage of digital behavioural data becomes valuable beyond the examples already mentioned earlier in the presentation.


[00:11:15.860] – Mike Brown

At the beginning of 2019, Lieberman asked to respond DeFrantz to design and conduct research on 18 to 25 year olds. Purpose was to get a better understanding of their online shopping behaviour, expectations and to focus on their interest, second hand products and the use of online marketplaces. A challenge for traditional approaches. That is why they opted for a two step programme. The first stage of the research consists of back tracking the online usage, the appropriate target group that young French Internet users from the list of users visited the corresponding timestamps, plus the activity on mobile apps and its duration.


[00:12:01.610] – Mike Brown

This step provided them with facts about shopping habits and customer journeys based on the reporting that their entire surfing sessions down to the old of pages browse. The next set consisted of an online community that combines live qualitative moderation by responder’s team with closed questions entirely, a quantitative assessment of the community’s insights and opinions. After a successful project, here are some of the conclusions and learnings that helped out in the bank, one that their proposition to make the consumers needs.


[00:12:38.500] – Mike Brown

Now entering some sort of Formula One motor racing. But many TV broadcasters. Formula One, a challenge with how to make their product as exciting, as engaging as possible for the fans. I now have so much choice of what to watch and when to watch it by streaming services or video on demand services. And rather than relying on traditional research approaches, using reported behaviours to understand their experience, the Insights Team F1 partnered with Walnut Limited and Reality Mind on a dual pronged approach.


[00:13:16.540] – Mike Brown

Firstly, by asking race fans in the US to download the reality show and pass it onto their various digital devices during two full race weekends, this allowed them to map exactly what Formula One related content was being consumed, at what times, from which channels and brands and on what platforms. It was analysis of this complex but highly valuable data that enabled them to work together to establish the key moments during a race weekend for certain types of content and which brands are currently fulfilling these needs and identify opportunities for F1 and at the same time in home, if not for using eye tracking glasses with a small number of fans in Germany, allow the team to bring to life exactly how secondary screens and information sources are used during a Grand Prix to support and enhance the experience of the fan.


[00:14:08.200] – Mike Brown

It got this feedback from the head of F1. Now onto an example measured protocol and one of their media clients, the client runs a weekly tracker, wanted to find a way for younger audiences to pop it, to take part in their brand tracker. They were struggling with low participation rates due to long questionnaires and incorrect and sometimes incomplete information. Mission Protocol captured all the subscription based video on demand through Netflix and Amazon Prime. As you can see, again, a positive outcome for measure protocols point, and I want to highlight the engagement with the audience, which we know is so important when we want to get through to authentic, valuable research.


[00:15:03.130] – Mike Brown

Next up, during 20-20, Szekely, Delta and Danish flight looked at the immediate impact of covid-19 on travel patterns in Denmark, using location data through the three mobile network, they built a tourist tourism insight’s dashboard to enable trying to look at domestic as well as international entry travel patterns and exit from the country. This allowed the clients at postcode level to understand a consumer’s appetite for moving around and within their country. So you see here, what do we see where people go, it’s where they visit their points of interest.


[00:15:43.730] – Mike Brown

And then my final example today is a location based insight’s example, again, from Tamako working with Kiah. That team worked on the data collected by an SDK in apps to understand the impact of advertising on driving footfall into dealerships, again showing the power of this type of location data not only to see activity around killing ships, but you can also look at activity at other brands and dealership. So emphasising this is actual rather than claimed behaviour. There is the ability to see who visit multiple dealerships, for example, and this was also shared through a dashboard, making it easy to slice and dice the impact across different demographic groups.


[00:16:31.370] – Mike Brown

So to wrap things up, there are several areas which will influence our ability to work with passive data for insights. Firstly, consumer participation and importantly, their consent. In an online access panel world, once permission is obtained, tracking continues until a consumer changes their preferences or at once their data is in the machine. It keeps on coming until they decide to stop it. Well, this is great for research as participants may forget that given this data mining is unbiased, however, is it right?


[00:17:06.350] – Mike Brown

Companies like major protocol, behavioural panels where individual every individual piece of work is consented to allow in. Their CEO laid it out very well with three points in a recent article on authority as giving consumers custody and control over their own data, breaking privacy into systems and processes to make what will certainly become increased consumer demands and providing a clear window into how data is being used in a safe way that doesn’t compromise privacy. And we look at tech companies who own a lot of this data.


[00:17:44.000] – Mike Brown

Apple recently gave more onus on our choice with how the data we create is handled, for example, trying to improve tracking pixels from email. Google on the same path and the cookie world is gradually being eaten away, in any case, building technology to stitch together the different websites, apps and devices use isn’t going to get easier. But again, with the right partner, remember whose business depends on handling these challenges, you can work through this process.


[00:18:15.470] – Mike Brown

Finally, what I want to reference our appetite to cross the line into the world of using digital behavioural data. The first project where I use webpages data was around 10 years ago. In the time since, I don’t think this has become as much part of research as talk is, it should. We know now how things can be done. The benefits of using the technology and the data for insights and the value they can bring to business. So I recommend we grasp the nettle of behavioural data and every thing it can mean to business, as well as keeping our eyes wide open for opportunity to benefit clients.


[00:18:53.860] – Mike Brown

It could even create a different differentiator from your competition. Now, I’d like to thank Ryan so much for giving me the opportunity to share my thoughts today and look forward to hearing and hopefully answering some of the questions now. Thank you.