Evolving consumer data analysis & how to apply it to your marketing

From data collection to intelligence.

Paul and James dive deeper into the application of audience-first marketing, and how data can be used for better communications, product planning, business strategy, and creativity. They also talk about how understanding an audience, whether a large FMCG or just a smaller startup, means you need multiple data points at scale. They also discuss the difference between consumer intelligence and social listening or like tools.


James Piecowye: Hi, my name is James Piecowye

Paul Kelly: I’m Paul Kelly

James: And welcome back to the ‘Know your audience’ micro mini-podcast.

Paul: Welcome back, James. feeling good?

James: It’s great to be back.  I’m feeling really good. 

Paul: It’s a great day. (James: It’s a wonderful day). It’s always a great day when we get to do our mini-cast. 

James: The whole point of these ‘Know your audience’ podcasts is to help demystify, help give folks more knowledge about how they can put into practice, really good practices of analysis, decision making, through looking at data and using data in an efficient manner. 

Paul: Yeah, that it’s exactly in the marketing context, I think. And the best part, I think, in what you just said, is also understanding that perhaps, how we have always done things doesn’t necessarily mean that’s how we need to keep doing things. 

James: And how many times have you heard that when I was a kid?…I’ve been doing this ….. 

Paul: This is the most ridiculous bias, 

James: Or “I’ve been doing this for 50 years, this is the way it works”. 

Paul: Exactly. And not here to disrupt any of that thinking, because learning, understanding from past experience is a great thing. But I think anybody in the finance sector will tell you past performance is not an indicator of future returns. And that’s certainly the case, I think, as the world evolves in communications, channels evolve, and everything changes at such a fast speed, that you need to really understand people’s behavior, what they’re doing, and how they’re interacting with the world around them to better reach them. 

James: And that’s exactly what we want to talk about in this episode is this whole idea that if we can understand the audience a little bit better, what they’re thinking, what they’re feeling, in all that that holistic aspect of their lives, we can message them better. And if we can message better, that is going to have a generally better effect on our marketing of ideas, our marketing of products, our marketing of everything. 

Paul: Exactly. And also, it doesn’t just make marketing better, it makes business more efficient, or products more targeted to their audiences and things like that. 

James: The bell just went off: ‘efficiency’. That’s what I’m interested in, efficiency. How can I be more efficient, not waste time, not waste investment of people’s efforts, and throwing it at the wall, the ‘throwing the mud at the wall’ model doesn’t work anymore? 

Paul: No, and it doesn’t need to work anymore. It doesn’t have to work like that anymore. And the reason for that is that, you know, there’s been a large period of time throughout history where we just haven’t known what’s worked. And that’s, to an extent, still the case, if your objective is to reach as many people as possible in a short amount of time as possible, then then you do need to have a bit of a scattergun approach. But even in that approach, you can tailor things like your billboards, radio ads, or what’s being actually said in the communication language, more effectively, to actually resonate with the person you’re trying to target. And that’s, that’s really the key, that’s where the efficiency comes through.

A great idea, (and we said this in series one) is always a great idea / great creative sorry, creative idea. So a great ad is always a great ad. And there are creative people out there, and this certainly not advocating to be only data lead data helps that process. It doesn’t, it shouldn’t hinder that process. And anybody who says that, doesn’t understand the value of what understanding an audience more implicitly and faster and better can do. 

James: Someone is going to be listening right now, and their eyes have just rolled into the back of their head and they’re angry. (Paul: They’re getting angry), and they’re going hold on a second. (Paul: Spittle coming out of their mouth). That’s it, they’re foaming. Where’s this guy? Where is this guy? There would throw their brand new mobile phone onto the floor and stomp on it except it cost them a whole bunch of change that they want to get into it, then someone is going to say, but hold on a second. Haven’t we always been thinking and investigating what the audience is thinking so that we can send any message to them? How is what we’re talking about in this micro mini-podcast any different than what’s already been done?

Paul: And it’s a fair point. I think. What a lot of strategists and planners and things like that in the advertising world have relied on is things like traditional market research methods, (James: surveys, telephone calls), traditional market research relies on using things called panels, which is, you know, bunches of people who, are, ready and paid to do this kind of stuff. 

The challenge is when people actually answer the survey, so forget about the randomized ones that you’re talking about. If you’re dealing with a panel, that’s fine, like it’s completely fine to pay someone for their time. The problem is that we have an inbuilt bias. The social desirability when we are trying to second guess what the other person wants to hear? It’s not necessarily a conscious bias, it sometimes is. But we sometimes think well, if I answer it in this way, and that’s why, you know, a well-designed survey, for instance, does take a very long time because it’s trying to eliminate that bias by asking the same question in about 10 different ways, right? So that you get to the right answer, and somebody who’s an experienced social researcher can do that. And it still works well. But market research in a lot of situations relies on face-to-face interviews, which sometimes relies on focus groups, where there is a peer group. A lot of those happen online now as well, not just in person. And the natural thing is that people will unconsciously lie. (James: Sure) Sounds harsh, (but no, it isn’t. But we do.) Whatever we want to say, we try and second guess you know, what, do you make ourselves seem better?

James: What do they want to hear? What’s gonna make me look okay, you know? Yeah, are you how many times a day do you week the exercise? …yeah, three. Yes. And how many times do you really exercise? None. But yeah, three.  

Paul: Yeah, and of course, I eat healthfully. Six raw eggs for breakfast, you know, and they’re just straight down doughnuts. 

And that’s just a perfectly human trait. It’s nothing bad. But what it can do is cloud the research. And there’s a great quote, there are lots of great books on this stuff. A great one to read is the choice factory by Richard Shorten, a recent book, he’s also a great follow on Twitter if you follow him. He has a great quote as the best way to understand (I’m going to butcher the quote, okay, go for it. My own words are the best), “the best way to get to a consumer truth is to watch what they do not ask”. 

James: Right. So the best way to actually understand what your audience thinks about something is not to ask them. But to take a look at what they’re doing and to take a look at how they’re talking about things (Paul: and behaviour), what they’re saying, and where they’re going to. 

Paul: Yeah, and what they’re interested in,  things like that, (James: which they don’t usually tell, you know), and if you ask, it’s gonna be not. Yeah, I mean, you might get seven out of 10 cases where they are telling the truth, but you will get outliers and things like that, in that response that perhaps clouds the end result. And there are a lot of examples through the history of product launches and things that have just gone horribly wrong. Because the researchers didn’t ask the right question. I think there’s a great one that I heard recently, might have been. I think Rory Sutherland was telling the story. He is the chairman, Vice Chairman of Ogilvy worldwide advertising agency, and he was telling the story about Coca-Cola and New Coke. Okay. And what the question that they asked, in the consumer research was the taste test and people like, right, alright, it’s great. Love it, you know? So would you buy it? Yeah. But yeah, but they didn’t say if we took away regular Coke, and gave you a new coke. And this is an 80s thing, right? So go and look it up. If you don’t, we won’t go into the full history here. Because we don’t have that time. But if I took away your regular Coke and replaced it with this taste, this new taste, would you still buy it? (James: No). Yeah. And that’s what happened. People thought that it was a supplemental. Yeah, it was just a new product. Would you buy it? Yeah, probably buy it once in a while I’m gonna drink my regular. 

So it’s a case of, firstly, slightly badly designed research, but also consumers, not necessarily saying the truth of what happened. And there’s, I mean, that’s the most famous, I think one of the most famous damaging corporate things, but there’s so many through history. And what we’re lucky in being able to do now in the modern world is being able to observe consumers, I guess, first before actually going toward to into the expanse or the detail of market research. And maybe it’s to better inform the brief of what they ask in market research or what they’re observing. But we can observe people first and get insight before that happens.

James: Which allows you to ask better questions, so that you can do the research, you can ultimately get the better messaging to your audience because you know, (Paul: better product, better launch or better design), because you have a better idea of what they’re thinking about. (Yeah). And to them as opposed to trying to, you’re speaking at them. 

Paul: Yeah, you’re speaking with them. Yeah. 

James: And I think they’re speaking with them. There we go. Yeah, it’s much better than to or, at. And that’s what you get through through this whole process is that you’re able to understand and build a better understanding of who those people are. Now this works on massive product launches, of course, but if you’re not faced with that kind of challenge, and you’re faced with an everyday challenge of perhaps getting people to engage with your content a bit better, you’re not going to go do market research. You know, a lot of its trial and error. It doesn’t necessarily need to be trial and error. And lots of tools out there that do that and, and give you a better audience understanding and better knowledge. But there is a difference in those in those kinds of tools is. 

James: So, can you give an idea of what the difference is? 

Paul; Well,  social listening, for example, one that everyone knows, we’ve talked about that they use keywords, monitor situations, conversations that happen publicly. 

James: That gets sold as “this is the answer”. And you can solve all your problems with social listening, and it’s affordable, and we can do it. And hey, we’re gonna give you all of this information. And we can do it fast and just, you know, sign on the dotted line, and it’s gonna revolutionize how you message your potential clients and your current clients.

Paul: Yeah, and I think there’s a certain amount of that is true, I mean, and if you’re able and have the skills of constructing complex Boolean operators and things like that, then it will be useful. But if you’re like 9 out of 10 people, and you don’t necessarily know that your experience with it is going to be very limited. 

And what social listening does for example is it really just what it tells us, as an example, might tell us that people talk more about soup on social media from November to February, right? I’m talking here in the northern hemisphere on techniques sensibly, not going to be doing that in the southern hemisphere. But, and that’s what I’ll tell you, it’ll give you that kind of insight. Now, there’s a, there’s a rich vein that can happen on top of that, which is what, like what we call but there are other people this is a common general term is AI-enabled consumer intelligence. And what that does is give us a rich vein of information. And our particular tool Sila is Arabic native, but these AI-enabled consumer intelligence platforms, what they’re able to do is tell us that maybe as a bit more detail that people maybe enjoy super, its warming quality during cold winter months, and they’re also interested in finding new, exciting soup recipes. Further, maybe in the run-up to the cold months, the colder months in January and February, people need help choosing the most nutritious ingredients for their families, particularly women aged 34 to 42 who have a solid affinity for fashion brands, (James: You can get all of that kind of information through an AI-enabled consumer intelligence platform?. It’s credible.) Yeah. And what it is, it’s the bridge between social listening, which is a fundamental thing that will just tell you a basic fact. Through to actually getting insight. And then perhaps, you know, very detailed market research, which may cost a lot more money. And what this is, I guess it’s a middle road, but it gives you a lot more information. So for you know, there are so many use cases for this stuff that enables you to really get a deeper insight. And that can really help you frame things like, parts of a good AI-enabled consumer intelligence platform will be able to do a lot of detailed analysis through different models. So things like sentiment, emotion, contextual analysis of imagery, personality, which is a good interesting one, because…

James: Personality? so talk to me about that. 

Paul: Well, what can happen, what does happen is that once you are able to observe an audience, you’re able to then segment them on shared interests, for example, or shared demographics, whatever the choice is, might be I want like I just mentioned before, women aged 32 to 40, whatever I said, and then we might be everybody who’s interested in traveling as students. And what we’re able to do then is apply a very specific personality model, Big Five personality traits, which is a very common thing, I won’t go too far into it now because of time, but look it up, we could also include it in our notes, some links to the article, what it is. And what it does is give us an insight into how that person behaves. So for a particular audience, for instance, it might be travelers from Saudi Arabia who want to go on vacation, so we might know that they’re idealistic and extrovert, I’m using, actually, (full transparency here), a working example, from something we’ve done recently, their top value, for example, as people was independence. So they’re not necessarily looking for power, empathy, they’re looking to travel, get out in the world see it on their own. So younger people, (I’m talking young people here, by the way), their knowledge or trigger is discovery. And they’re heavily influenced by a brand name, meaning it’s less about function, which therefore means less about price. It’s about the quality of the brand name, it’s about being able to discover things independently and that they’re extroverted and idealistic. 

James: And so I just want to clarify here because I think as someone is listening to this, that’s an incredible amount of information, and fairly targeted, and also really detailed. And this information is coming from observing, what is being posted, and what is being shared by people on their social media. Correct?

Paul: Yeah, and Twitter is the best source for this because people are typically more, let’s say, forthright or honest about their interests on Twitter. So as opposed to Instagram, for example, we live in perhaps a machine-live-filtered world. Yeah, it’s only natural. That’s what the platform is designed to do, right? Whereas Twitter is more about conversation, news. And that type of thing, so typically people, particularly when you’re doing a segmentation analysis, that the accounts that they follow are more aligned to their interests. So things like sports, clubs, news, things like entertainment, entertainers, things like that, that’s more who they are. As people. It’s not perhaps a fabricated version of themselves, right, which is perfectly natural. But because Instagram is designed to do that, typically in Instagram, we what the real insight lies in things like comments, rather than people’s own posts. But to do a personality analysis, you can get a really good idea of what a cohort or segment of audiences really like by examining… (And again, this needs a baseline needs about 3000 tweets, replies, interactions that are public, for each member of that. So you’re talking hundreds of millions of points data point to be able to get to this level of certainty)… but when you do, you get a really rich indication of who and how those people are. And it’s in the really interesting thing is when you sort of give the same sort of analysis or ask the same sort of questions to somebody, you get a type, you know, a very close alignment on that audience. 

James: I want to back up for a second because that was another piece of the data that you alluded to, you spoke about is that not only are you looking at, let’s use Twitter as the example. So looking at the post, and the physically what’s written in the comments and like, but you’re also looking at the images, because that’s another big piece of posting, and you’re generating sentiment from what’s in the images. 

Paul: Yeah, images, the text is large is a bigger source of sentiment, what images do is give context. Okay, context, yeah, so you’re able to better understand perhaps the context of a post by understanding what the person’s because social listening, just general, I mean, platforms also now do visual analysis as well. But it’s largely centered on the text. The better angle to run is using imagery for context. And then understanding the sentiment, and adjusting the sentiment based on that context, and the emotions and things like that as well. So you’re better able to physically get an idea of who that person really is, and what their intent of what they’re saying is because, in certain languages, even every language, sarcasm is a really hard thing to pick up from the machine. Yeah, you know, as it does, you know, it does take a lot of work to build models that understand or at least can understand the context that hey, maybe I shouldn’t analyze this because this could be sarcastic. Like, hey, the really good stuff that James? Like, yeah, you know, you missed a lot. That one that’s written, right? Do you only know that by the context or the emojis that are being used? And yeah, yeah. And even then emojis can be used sarcastically or, you know, crying emoji can actually mean happy. And so when you understand that context, you’re able to build a model like we have that understands the context, and you have better precision. 

James: What I want, what I’m curious is once clients have gone through this process, and they’ve started to use these tools, what are they saying to you anecdotally about their future messaging based on the messaging that they’ve done based on the insights that they’ve gained? 

Paul: Yeah, I think, depending on what the challenge at hand is, it just creates better briefs for their agencies. 

James: So so right back to the whole thing with the creatives, the better briefs of the creatives can do a better job of designing the imagery and the text and the cars and the consumer.

Paul: Like, yeah, exactly. What colors to use, what are the triggers? Imagery for that audience, you know, if it’s a food company is that it’s like a pizza with really stringy cheese? Or how do you make that more attractive to somebody, somebody’s looking for a more traditional approach to that, you know, where the the cheese isn’t necessarily brown where you know, or it’s like, really crispy or, you know, glazed doughnut, you know, like, you’re able to build a better picture and a better brief without defining what that is. But then you’re also able to apply it successfully to customer experience, as well. So things like actual sort of notifications from an app or those types of things. So you can use those Personality Insights, for example, to cater to somebody’s need for commitment. So you could, you could say something like, you’ve reached the top of the queue, as the first thing you say in a notification if they’re waiting for a chat or something like that. What those three things do is first is a commitment you’ve reached you’ve reached a goal. So if that’s important to somebody, even this is all subconscious, so you know, it just helps build a better experience. So you’ve reached something so you’ve committed to it and you’re going to follow now please follow through on that because you’ve waited this long. Please just stay a couple more minutes, that reached the top, help somebody’s ego, you know, you’re at the top of the queue. So if you if that segment has a need for status, that’s really important. Because I’m at the top, it’s just subconscious. Again, it’s not like I’m thinking, looking at it going, Oh, that makes me feel, you probably roll your eyes, but it gives you a slightly better experience. And that’s what it is about incremental improvements to experience. And so you reach the top of the queue, for example, the queue is social proof, oh, there are other people waiting. This is actually a really popular thing. You know, and I think some organizations and sectors, particularly the, everybody’s got bad experience with banking, where they’ll just say, you know, if I’m experiencing a high debt, load of loan calls, blah, blah, blah, that doesn’t cater to any of this kind of messaging. And they should probably be more forthright about how long things are gonna take. And perhaps, you know, there’s a high volume of calls here. Great. So tell me how many calls right, right. It’s not difficult to program this stuff to do this other thing? We’re experts are we’ve got over 100 people in the queue, do you want to call back later, or press zero? To leave a different phone number to call you back on something like that? Right, can improve the experience, but unless you know about that audience, you’re not able to build that stuff. 

James: So as you know, it’s really is that nudging activity. So yeah, just a message even, just a small change is gonna make you much happier. 

Paul: And you can fill that whole message, like that sentence that I wrote is the whole message. So it could be, you’ve reached the top of the queue for a callback for an upgrade on your upcoming, let’s say, Emirates flight. And, you know, we talked before about the need for the brand, acknowledgment. So even though they know they’re in a conversation with somebody like Emirates, and so whatever, the airline doesn’t matter, British Airways, whatever, the name is important to them. So you can reiterate that. So rather than just saying, your upcoming flight, you reiterate the name. And it just gives the subconscious nudges. And what that does is build authority just by using that brand name, because you’re able to, it’s the credible Messenger, 

James: And you’ve learned that through the analysis that you’ve done of all of this information that we’ve gathered in another way. Amazing.

Paul: It’s about providing reassurance because that’s that need. So you understand the values, the needs, the triggers, and those types of things that help you frame that customer experience. But it also, you can use that same insight, as you could well imagine, for copy on a billboard, write, copy on a website, all that type of thing, and you’re able to build that information from a deeper and richer understanding of the audience. 

James: So at at the end of this whole process, we’re not saying you’re not saying, Oh, we got to eliminate the way we’ve done it in the past, and it doesn’t work. This is another layer, enhancement, our tried and trued skills, and with information with data, and now we’re adding another layer of it so that we can enhance the effectiveness of our messaging. 

Paul: Yeah. And what that might do for your market research, for example, is make it way more efficient. So bring the cost down. So you don’t perhaps need to go to as many people you know, specifically the question you actually want to ask somebody. So you can design the surveys or observations or home visits all sorts of things a lot better. But also just in a digital customer experience or social media experience, you can use it to frame copy on ads, to frame the imagery to know where your specific audience is on what channel they’re on, and all that sort of thing. It just helps you tie things together. It’s sort of like the package around everything. And what it does is drives efficiency. So the amount that you might spend on something like this, then pays back, you know, three, four times over, at least, I’m talking small amounts, here, I’m talking large amounts on larger amounts, it’s much higher ROI. because your spend is getting more traction in whatever your objective might be, might be to get more leads. 

James: Everyone wants more effective spending, I don’t want to waste my money, no. Target, this is the way to target and get a better return.

Paul: It comes down to understanding terms, I think that’s the biggest part and understanding that, you know, one size never fits all and if you do go for a one size fits all approach, you got to understand the limitation going with your eyes wide open, you know, like something like social listening, know its limitations and use it to that, to the extent of what it can help you and not perhaps a bit like we’ve talked about before, what someone is selling doesn’t necessarily always reflect well, and wait, what you’re gonna get. 

James: What I love about this whole program, and what we’re talking about on this micro mini-podcast, is that it’s not saying we’re going to get rid of employees, we’re gonna get rid of people who do this research. It’s saying, No, we’re gonna make the people who are doing this research and analysis smarter, because they have the ability now to do better analysis, as opposed to the grunt work of collecting this stuff that they then have to try to make an analysis of, but they’ve got an incomplete collection method to start with. 

Paul: Yeah, or, or they don’t want to be guesswork, or they’re just using past performance of campaigns that for instance. 

James: Here’s the best way. I saw people at the coffee shop every day I go to the coffee shop and I see this happening. This has got to be what people are thinking. (Yeah, that’s it). “Every day I oh, five people that you see the same five people”. Yeah. “Same time, same behavior …and  I’ll become your, your base sample line”. The baseline of five. 

Paul: And the problem is, I guess, in a lot of situations that the problem at hand or the challenge or the opportunity doesn’t necessarily justify the spend on, you know, for market research things. So you try and use your gut instinct, I think. And we can just enhance that gut instinct by having a better layer of information. And that’s what really that’s what AI-enabled consumer intelligence does. 

James: And I’ve learned a lot, Paul. 

Paul: Thanks, James. I’m glad to hear it. I taught the doctor. 

James: I’m really excited because we really laid out some nice groundwork here. And again, go back and listen to the first series as well, it’s the 101. 

We’re now on the second layer here. And where we’re going to go in our next Micro Mini podcast is talking a little bit more about Arabic and language and misunderstandings and how this whole platform can help you navigate that. And maybe I’m going to use the word flatten, flatten the errors, but there’s probably a better word. 

Paul:  A better way to improve your accuracy. 

James: And there we go. My name is James Piecowye (Paul: I’m Paul Kelly), and this is ‘Know your audience’.