Roddy Knowles, Dynata

Thank you for meeting with us Roddy, can you start by introducing yourself?

Sure. I’m Roddy Knowles, I work at Dynata and sit on the Research Science team where I support our Product team from a research and methodology perspective, ensuring all our products are created with a sound research foundation. I also support the company from an innovation perspective, keeping an eye on where the industry is going and tactically in things like bringing on new data partners or bringing on new data sources that can integrate with our products.

My role also involves a lot of thought leadership to support our products, as well as getting into a lot of nerdy stuff too – like research on research.

Why did you get into research?

I sort of backed my way into research. I’m a social scientist by academic training; I studied religion at a PhD level. I didn’t do theology, but rather took a social scientific and historical approach to the topic. That may sound far off from a career in data, but there was a lot of data to look at.

During my PhD work I was presented with the opportunity to do ethnographic market research – and at the time I didn’t even know I could apply my academic training in a meaningful way. Long story short, I did a lot a shopper insight work in and after grad school, doing quant and qual data collection, analysis, and wearing a lot of hats. Then I made the decision to leave academia and do research full time… all downhill from there!

Can you introduce our chosen topic of ‘Data Integration from multiple sources’?

Of course! Depending on who’s listening or reading, they may think that data integration is a normal thing and what’s the big deal because of course, we integrate data from multiple sources. However, Market research isn’t usually lauded as being the most forward-thinking industry, and although I think we have made some massive leaps in the past few years, historically we tend to do one thing at a time, like qualitative interviews, or quantitative online surveys, or look at CRM data, with a tendency to look at things in isolation. We have been pretty siloed, historically.

What market researchers and insights professionals have done increasingly over the past few years to stay relevant is pull in data from various sources. It could be pulling in traditional research data or third-party data, or it could be combining different types of research data. It’s thinking about things from a more holistic perspective.

One of the questions we should be trying to answer is what data best allows us to address the problem or objectives, instead of saying ‘this is our methodology, so this must be our approach.’ Then figure out where or how to get to the right data sources.

Where have you seen Data Integration being done successfully?

One of the ways I’ve seen Data Integration done successfully is taking data from traditional survey data and combining it with observed data from other sources. It could be advertising data like ad-exposure data, purchase history data, voter data, or something that’s going to tell people more about what they are asking about.

If you’re trying to understand the impact of ad campaigns and marketing on how a brand is doing, knowing when they are actually exposed to an ad, or which channels work well is critical. If you want to know about a shopper’s journey, pulling data related to their shopping or purchase history would be beneficial.

It’s important to think about where you can get relevant and, I’m not sure if ‘clean’ is the right word, but data that most directly answers the question being asked – without necessarily needing to actually ask a question because again, this data can be observed and appended. As researchers, we have a habit of asking every question we can think of asking, because that’s what we do as researchers – we ask questions. But this often isn’t a good thing.

What are the biggest challenges you have faced with Data Integration?

I think the biggest challenge for most, is getting lost in all the data and with all of the shiny things you could work with! So, the first challenge is getting over that hump and planning how you can approach your research more effectively with data that actually matters.

You can look at all the different sources you have, your primary research data, survey-based data, qualitative data, etc. and then wanting to pull in more data such as from third parties and then getting lost by integrating too much.

The data sources themselves can pose challenges. Some survey researchers aren’t used to dealing with a lot of these types of data, some are structured, some unstructured so it’s easy to get lost. It’s usually best to start small and think about what type of questions are not really being well answered by the survey data you are gathering. What sources are going to be a better fit to gather that data? You can then find one or two sources, rather than collecting too much to pull in and get lost within it.

Would you say there’s a limit to how many data sources you should integrate?

It depends on the tools you have available to use, and the humans using these tools, and what type of data you are integrating. It can also depend on if you are leveraging AI, Machine Learning or other ‘buzzwordy’ tools to integrate, which is something I am seeing a lot more of. I wouldn’t say there is any sort of a limit, it just really depends on the tools, someone’s skillset, what that person’s approach is. I know data scientists are comfortable with a lot more data than what I could ever work with!

Is Machine Learning and AI something you are seeing more in Market Research?

It is something that is talked about a lot more now than a couple of years ago; those are hot buzzwords that are getting thrown around a lot at the moment. I think if you had a booth at a trade show and wanted to attract people to it, you could just put the letters AI in big letters. But in all seriousness, there are a lot of cool applications out there, allowing us to do more with research data whereas historically we have been a little more limited with our analytics tools and people are realising that there a lot of practical applications and gains in efficiency with AI and Machine Learning.

There are also a lot of companies doing cool things by leveraging AI, especially working with unstructured data. Also, with “Big Data,” dealing with massive data sets about online behaviour for example. You can use AI to look at unstructured data in terms of qualitative too, or use AI to moderate qualitative interviews and then analyse and structure to add meaning to the data. There are a lot of interesting things going right now with the buzzword du jour.

Can you lend some advice to people who conduct Data Integration? Is there anything people should consider before doing this?

I think this goes back to what I said earlier and it sounds sort of basic and this is the researcher coming out in me! It really comes down to making sure you know what question you are trying to answer and focusing on that. That will help you to be strategic, starting at that high-level. Instead of thinking ‘we have all of this data, what can we do with it?’, rather you should say ‘what are we trying to accomplish or what problem are we trying to solve?’ Think about things logically and from a business perspective or from a project perspective. What are you trying to tackle? And then, only then, start to think about what types of data you need to collect.

What is Dynata currently doing with Data Integration?

We are doing a lot of cool things at the moment! As a global company of our size with a first-party relationship with millions of people, our panellists, we collected a lot of data. Our panellists opt into sharing data via surveys and other means We have profiling data that we collect on an ongoing basis in addition to survey data. We also have other datasets that we can link to through our partners – and there are a lot of them relevant to a number of verticals.

We also often work directly with brands to link our data with theirs – what we know about people and can gather from them via directly engaging, essentially surveys, and what they know about people through their CRM or other datasets. We’re focusing a lot on how to make it easier to access our first-party data so stay tuned for a product release soon on that front.

We are doing a lot of interesting things in advertising measurement space, looking at how we can use ad-exposure data for tracking studies, for example. This capability sort of builds upon the core of data we collect with our Ad & Audience measurement products. And on the other side, we’re working with companies to help them advertise to the right people through using our data and relationships to help them build models and better target their ads.

We’re constantly looking for ways to allow our clients to connect – or integrate – more relevant data, whether that is data we collect directly from our panellists or not. Being open to leveraging your own data in combination with other data – and making these connections as easy as possible, is one of the keys to success in our industry right now, and in the future, in my opinion. And to be candid, it isn’t just us out there doing that. more companies than ever are trying to do this sort of thing because they recognise that this is where the industry is going.

Are there any situations in which Data Integration wouldn’t be beneficial or should be avoided?

Yes, I mean there are quite a few situations when you don’t have data that is relevant to what you’re doing and won’t fit – a sort of square peg round hole situation! For example, say if I have a situation which I want to understand a shopper’s behaviour in a certain category I might have certain information on shopper’s behaviour or point of sale data, but what I really need to understand is are the shopper’s needs states for what is driving her to buy a product. I can’t answer those questions from that type of data. I can hypothesise, maybe based on what other people have purchased, but you really need to engage directly to understand that.

I would be wary of taking data that is similar but not fitting and working with that rather than focusing on what’s really going to answer the question. And really, Market Research in many cases is focused on doing just that, working directly with people. Some people think that Big Data is going to solve all our problems and we don’t need to ask questions anymore, but I think that in fact it’s quite the opposite. Maybe we need to ask smarter questions and engage more people rather than ask every question that we think of. I’m a bit of a broken record, I know. But that’s because it’s an important point.

Where do you see Market Research going in the next few years?

In the next year, I think we will carry on the same path. But looking into the next few years, I think that a lot of companies that are focused on the traditional data collection part may find it challenging to keep up. I think if people who are doing online market research and are relying solely on survey data you collect directly and are not leveraging outside sources, staying relevant will be a harder thing to accomplish.

I also think there are going to be opportunities for companies who are only collecting this observed data to leverage more engaged data, by talking with someone directly or conducting research with them. So, I think there will be more partnerships.

Our company for instance, are not experts in everything. We have a lot of capabilities that continue to grow but we still leverage external partnerships for a lot of the stuff we do to service our clients. In our industry, I think we will see a lot more ‘coopetition’, where companies may be competing but collaborating at the same time. If you’re trying to offer the best you can for your customers it might not be what you can do, and you might not have all the data, but if you can access it and partner with another company that can help then you’re doing all you can for your client and building relationships as you’re breaking down barriers.

If you were to start your career again is there anything you would change?

That’s a tough one! I don’t know if there’s anything I would change. I spent 10 years in a PhD programme which may have been a little long, so maybe I should say that! But…in some ways it had no practical bearing on what I’m doing now but I learned a hell of a lot, so I don’t know if I would change that. Some of the soft skills and technical research methodologies helped but the subject matter doesn’t really. I don’t study religion anymore and it has nothing to do with what I’m doing now. The things I’ve learned from teaching and writing, and those skills are so beneficial. I don’t think there is anything I would change.

If I can flip your question around and answer a different one, I can tell you one thing I wouldn’t change. I think one of the most valuable things that I’ve learned and something I urge people to do as a researchers is to have some sort of understanding of qualitative research. It doesn’t matter if you’re a quantitative researcher or a data scientist, so often people are doing work with human data and trying to understand consumers and people. Working with people, directly or indirectly, and having in mind that they are human too really helps you with your work and research. It helps you create engaging ways to research. It helps you when you are trying to think of the type of data you are collecting and understanding what context people are creating this data. It may sound simple, but people forget how beneficial it is, talking to people, it is completely invaluable. An example is a brand manager is talking about the voice of her customers she should be able to go out to her stores and talk to her customers and hear and understand what the voice of the customer is!

What has been the biggest challenge you’ve faced in your career?

This follows from the answer I have just given, but one of the challenges is trying to get people to conduct better research because we often forget we are doing research with humans.

I think the biggest challenge is trying to get people to shift their mindset to think about the fact that there are different ways we can collect data and integrate data from different data sources, it’s been a real challenge and an ongoing one.

Another one would be trying to get people to conduct research that is mobile-friendly. I spent a long time trying to implement this and I’m not sure how much change it’s made.

What advice do you have for someone starting out their career in Market Research?

It would have to be to conduct some form of qualitative research and gain some knowledge and skills around this, even if you have no formal training. Having a wider reach of different approaches to Market Research is important but focusing is good and to have a focus or speciality is important. So, if you find an approach or niche you like, dive in.

Attending events and conference is also extremely valuable. I try to go to as many as possible to improve my knowledge as much as I can. It’s also worth noting that it’s great to go to events and conferences in your space and interest, but exploring outside this is an interesting way to learn new things and get inspired. The most impactful ones on me this year so far have been the ones related to, but not directly focused on, my industry. And if you can’t physically make it to events, there are no shortage of webinars available to feed your mind. So, my last advice is that if you aren’t learning new things regularly in your routine, you need to step out and force yourself to do so.

Learn more about Dynata