Down the Social Media Rabbit Hole: Fine-Tuning for Your Target Audience

Dean Bromwell

Dean Browell is Co-owner and Executive Vice President of Feedback Agency, a market research firm that uses hands-off observation of online behavior coupled with behavioral analysis by data scientists (psychologists, sociologists, and anthropologists). When people are in focus groups or other situations where they are aware of being observed, their behavior changes. By not engaging their audience, the digital ethnography research team is able to get a rich body of authentic, unadulterated data.

In the discovery phase, Feedback’s digital ethnography research team works to understand a client’s audiences and where they are going. (FaceBook, Twitter, YouTube, forums, niche message boards . . . and deep into sub-communities) The team then observes behaviors, identifies preferences and channels, assesses group discussions, actions, and sentiments, and evaluates all of this existing data to discern what “makes people tick” and what causes them to make decisions. Through the sheer volume of information, the research team can identify trends and quantify and qualify word-of-mouth impressions.

Using that information, Feedback may test responses to different messages. For example, they may place 70 small-space FaceBook ads and be able to report back to the client the 4 top-performing imagery types and messaging types. Feedback is not involved in the marketing implementation . . . they hand over the data to their agency customers for use as implementation guidelines and provide clients with detailed, research-based action plans for increasing customer engagement and sales.

Dean has found that human beings are far better at detecting meaning. Where computer programs may track words and bot can make decisions about whether words are positive or negative, humans can pick up the nuances of sarcasm and humor . . . and provide a far more accurate view of audience sentiment.

Dean recommends a couple of resources for targeted searches. Board Reader( searches forums and message boards, either by content or by forum focus. He also talked about local message boards, such as City-Data, which can provide a wealth of information about the selected community.

Dean can be contacted on LinkedIn at in/dbrowell/. His company’s website is

ROB: Welcome to the Marketing Agency Leadership Podcast. I’m your host, Rob Kischuk, and I am joined today by Dean Browell, Co-owner and Executive Vice President at Feedback Agency based in Richmond, Virginia. Welcome, Dean.

DEAN: Hi. Thanks so much for having me.

ROB: Great to have you here. Why don’t you start off by telling us a little bit about Feedback Agency and what makes Feedback great?

DEAN: You bet. Feedback is ostensibly a market research firm – and I use that word “ostensibly” because we approach things from a very different angle. We use what we like to call digital ethnography. Most people are familiar with ethnography. Think Jane Goodall and the gorillas; don’t touch anything, you just want to observe behavior.

We do that, but we do it in an online space. Instead of relying on bots or even complementing, obviously, surveys and traditional research methods, we’re looking at all of that data that’s already out there, all of that behavior that’s already out there, and using ethnographic techniques to discover what makes people tick, what makes people make decisions.

That’s the heart of what we do. The other real element that sets us apart is we use human beings to do it. For whatever reason, that truly is a special element of what we do. We use folks with sociology, anthropology, and psychology backgrounds and really sic them on the internet. It’s easier to teach them how to find where people are than it is to teach, say, a marketer behavior and behavior analysis.

ROB: This ethnography approach, is the premise behind that almost like in physics, the Heisenberg uncertainty principle? Essentially, if people know they’re being observed, then your observations are less trustworthy. You can’t maybe do both at the same time.

DEAN: Exactly. The way I like to say it is what people will say in a room with free pizza and a moderator is very different than what they’ll say to their neighbor over a beer in the backyard. What we want to do is basically try to quantify and qualify word-of-mouth, because it is out there for us to see now – and has been for almost two decades with at least some form of social.

To that point, actually, I should say when we say we’re looking online, yes, you can call it social media, but it includes any place that there is user-generated content. We’re talking niche message boards, we’re talking forums, we’re talking review sites. It’s not just Facebook and Twitter, although it would include those things as well.

ROB: I was going to ask, what are these native environments for internet humans? It sounds like social media and forums, and for certain businesses it could be Yelp, but then very boutique forms as well.

DEAN: Absolutely. We do a lot of healthcare work. I’d say about 30-40% of what we do is in healthcare, and that’s a perfect example. When you’re diagnosed with an issue, you might say something about it on Facebook and you’ll get lots of sympathy and support from family and friends there.

But that’s very different than going onto a message board for people with that issue and how you can talk about your journeys, talk about your relationship with your physician or your nurses, talk about your relationship with different brands, what you’re comparing to make your decisions based on.

We really do talk differently when it’s that niche peer group about an interest or an issue than we might on something very general, where it’s more of an online community that has all facets as a part of it.

ROB: You say you’re doing this with people. How much of the end work product is quantitative and how much of that is more subjective?

DEAN: Yeah, it’s funny. My background – I definitely am a qualitative researcher. It is absolutely a qualitative product. That said, when you’re doing work for a retail fashion client and they have 3,500 stores, and each store has an average of 200 to 1,000 reviews and we’re crunching all of that, it ends up being a very quantitative end product also.

We’re able to say with quite a bit of certainty what we’re seeing isn’t just something anecdotal, but trends that we’re seeing. “All the reviews say this, this, or this about your product or your service. Here’s the things that really rise to the top, here’s the outliers,” and really put it in perspective.

Our end product, which tends to be a presentation and then a set of appendices where we show all of our work, really does end up being both quite deep and rich qualitatively, but absolutely ends up being quantitative as well – even if that number is zero.

If we go on there to find something and we don’t find anyone talking about something, zero is a data point, and I want to know why. Why are we not finding that?

ROB: Right, which is probably an interesting thing in the medical space in particular, because there’s a lot of things that people won’t talk about. We’ve told our own kids, “Hey, don’t talk about Mommy and Daddy’s doctors’ visits.” A lot of people have a sense of privacy around that.

DEAN: That’s right. And they value anonymity when they do want to talk about it, and that’s actually why a lot of those niche forums and message boards – often those people are not revealing their names, but they’re saying, “Hey, I live in Poughkeepsie and I’ve got this particular issue. What are some docs around here I should go to?”

They don’t ever reveal their name, their family. You may not know anything else about them. They’re clearly valuing that anonymity but still want to be able to reach out and find some answers. They’re valuing that peer interaction, but not at the price of their privacy – actually, that’s a good example of why they don’t ask those questions on Facebook, where their names are.

ROB: Facebook is now increasingly and notoriously harder and harder to do anything with in terms of public data. Are you finding ways to break through that?

DEAN: A little bit. The way we put it is, if it’s private, we will be noting that there is a private wall here and we won’t be trying to go past it. What we want to try and illustrate isn’t just the answers that we find and the insights. We’re also trying to illustrate, what does someone who comes to the table with this question see?

For example, if it is healthcare, or even if it’s someone who wants to buy new windows for their house, what does someone who’s doing their internet research – someone who’s asking peers and looking for what someone else says about this – what are they going to find? It doesn’t mean that what is out there is capital ‘T’ Truth of the way everyone feels, but we have to deal with it as being the reality of what a novice or someone coming to the table is finding.

ROB: The internet is such a huge space. I think it’s easy for anyone listening to picture how they might approach Twitter, how they might approach Yelp. How do you find these other corners of the internet where there’s meaningful conversation, but maybe it’s a one-off platform just for that client? That seems like the real trick. There’s probably a few one-off platforms for each client or type of client that might not be expected just at a glance.

DEAN: You’re absolutely right. It truly is a little bit of an art. We break our process down into two parts, and it’s not 50/50. I’d say it’s more like 20/80 in terms of the time we spend on it.

That first part is what we call discovery. That really is understanding the audiences and where they’re going. We’re going to find all kinds of channels. This is where the ethnography part really comes into play, this snowballing or following people into rabbit holes type thing.

I’ll give you a great example, because this is probably the best way to illustrate it. We did a project for a men’s fragrance. We knew that there were lots of small sites about people reviewing fragrances – men’s and women’s. On Twitter, though, we found a guy mention a YouTube video that he had seen about a fragrance he liked.

We followed that rabbit hole, that YouTube video, and we ended up finding a whole series of YouTube videos that were unmarked. In other words, you couldn’t find them by search. It was an incredibly illustrative set of videos. They might be 20 minutes long – reviews of fragrances, talking about what to eat when you’re wearing them if you’re on a date, don’t sit near a fan if you have this one on, how long they last, when to reapply.

In the comments section of those YouTube videos was a link to a private message board that had about 4,000 to 5,000 folks on it with much more in-depth reviews than anything we’d found before.

That’s a great example where we found that from a tweet, but we followed the rabbit hole down to find the sub-communities. These folks, some of them are influencers. Sometimes we might look at that and say, “this is a very niche audience that isn’t going to influence the wider market” – but you need to know that it’s there.

You can imagine how in some cases, it’s good to know it’s there because in a crisis, those people may be activated either as your detractors or as your advocates.

ROB: Interesting. Dean, if we rewind a little bit, what led you to start the company in the first place?

DEAN: My background is actually a little more academic. We’ve got three founders of Feedback, and all three of us are different generations. I’m the Gen X’er of the group. Jeff Thompson, our CEO, comes from a very long line of agency background (Ogilvy & Mather, The Martin Agency) whereas I was in the agency world for about two seconds before we started this company.

I actually have a PhD in education (Higher Education Leadership), but specifically in how different generations interact online. My dissertation was trying to find a way to marry ethnography with the online space. Originally, though, it was to help improve the first-year experience at colleges. It was about what we could learn about what these millennials are doing online to improve everything from mentorship programs to allowing some anonymity in asking questions about the college experience, all kinds of elements like that.

When I left higher ed., I went to an agency that was working with higher education, so I was still working in higher ed. a bit there, just multiple clients. I met our two other founders at the time, and basically I started to say, “Why can’t we apply this same technique, but for every industry?”

Honestly, I actually said “no” when first asked whether we should start this little playground, because I’m not really the entrepreneur type. But eventually I said, “You know what, let’s dive in.” That was actually 9 years ago next week.

ROB: Happy anniversary on that. How has the agency evolved from the initial concept to where you are now?

DEAN: It’s funny; there’s actually been some constants that have never really changed. We deepened our deliverable with what we do, our training regimen for our researchers. Our researchers have very specific skillsets. Somebody from an anthropology background is a very different kind of mentality than someone with a psychology background.

So, we’ve really done a lot to evolve how we’re choosing the team to put on certain projects. I should also point out 50% of our clients are ad agencies who use us with their clients. Someone comes to us and says, “We have these personas that we built out of some web work we’re doing, but we know they’re not exactly right. We need them tuned. Can you go out and validate them for us?”

Our folks with psychology backgrounds are amazing at personas. Personas, archetypes, that is the basis of a lot of psychological teaching. We’ve gotten really good in the last 9 years about tuning the right people for the right project.

I think the other thing is we’ve added in another element of the research we do where we’ll use small space Facebook ads on the backend of a traditional – traditional for us – research project and do message testing. We take what we learn from the qualitative and then try to test the calls to action that we think they’re going to react to.

Then we can come back to the client with a giant matrix and say out of 70 small space ads, here are the 4 top-performing imagery types, messaging types, things like that. So they’ve got both the deep qualitative and then also what these same folks will act on.

We use Facebook because we can go by behavior much better, and targeting. It just has a little bit of an additive. But that’s something we added on about 3 or 4 years ago, and it’s really been great to, like you mentioned before, further help quantify the qualitative.

ROB: It seems like you have a very unique proposition in what you do. There are two things I would observe from the outside, and I’d be curious about your thoughts.

Number one, what you do and what you offer is hard to commoditize. It’s maybe sort of easy to do the really lazy version of it, but it seems hard to commoditize, it seems hard for somebody to bring in-house, and it seems like it would be a very unique opportunity for your team as well. It would be less probable that someone would do the usual “agency to vendor to brand and back again” dance that happens so often in the marketing world.

DEAN: Right, absolutely. I’d actually say we have more perceived competitors than we do have direct competitors for what we do. Of the perceived competitors, the number one thing are the dashboards. Radian6, Sysomos.

The irony is even they admit that they don’t really do what we do because they don’t even look into message boards most of the time. They’re just looking at Facebook and Twitter and outputting red light/green light, positive/negative brand temperature, if you will.

When we first started, the dashboard thing was really taking off. Our first client was The Martin Agency, who also was ironically one of the biggest investors early on into some of those dashboards and bringing them in internally. At the time, we weren’t threatened by that – in part because we were just starting out and we knew we were very different.

It’s funny that 9 years later, we now get more clients who bought into dashboards and are now turning around and saying “The investment we made for seats in this dashboard and the, generally speaking, junior folks that we’ve hired to run this dashboard – we’re starting to realize we’re hardly getting any real insights out of this.” Not to mention they’re not even sure they can trust the positive/negative bot decisions about sentiment, too.

So we get a lot of clients who take advantage of us – and might even keep the dashboard. One of our clients is CamelBak; they use dashboards as well, but they also use us to augment and help tune their dashboard experience so that they’re capturing more.

But also, CamelBak is making all kinds of water bottles for road cyclists or mountain bikers or runners. They know that those forums, they want someone like us looking into, because a bot is never going to understand sarcasm or really understand what their needs are so that they can make better models.

ROB: I think this is a really good question to dig into. Is there anything even pretty good in terms of automated sentiment analysis? Or is it kind of a hot mess?

DEAN: Here’s what’s funny. I think that there are some automated sentiment analyses that can get somewhat close. My favorite of the group – and we’ve actually worked with agencies that use it – is Watson, IBM’s thing.

What’s fantastic is – and this is Drake Cooper, who’s out in both Boise, Idaho and in Seattle, Washington – they have done some great stuff with us where they’ll have us do our work on the front end and then we will use Excel to categorize all of the sentiment we’re seeing. Then they dump that into Watson and it helps them then, from the emotional standpoint, build better campaigns for the next year.

I think there are some great pairings for it, but the problem is you would not want to rely on that for a day-to-day analysis of how your brand is doing. We did a study once, just a test for CamelBak. It was for one of their products, and it came back – I think it was over a month – and said there’s 1,000 positives and 1,000 negatives that came back based on this dashboard.

We looked into it by hand and noticed that every single one of the negative ones was actually positive, but it was someone saying “I finally got my CamelBak.” Because the word “finally” was deemed negative within so many words of the brand . . .

There’s also, to use the safer word, the “poop” problem, which is if someone uses the poop emoji, are they saying, “This is crap”? Are they using it sarcastically? Are they using it as, “You are . . .,” in a positive way almost? Emojis are a great example of how those bots can be fooled in a heartbeat by Gen Z right now. It takes someone with context to understand how and why people are saying what they’re saying.

ROB: For sure. We’ve definitely seen, especially some brands with some resources, have looked more at human scoring.

We, at one point early in our own life of our company, worked with a very big mobile phone manufacturer and piping in some Crimson Hexagon data around their brand. I think maybe 3% of the tweets had any sentiment whatsoever. How could that possibly be actionable? Yes, it’s 3% positive or 3% negative or whatever it is, but if it’s only 3% of the conversation, what does that even mean?

DEAN: You’re exactly right. Then you layer on top of that the fact that then agencies or brands will hire honestly some of the most junior people that do not have behavior backgrounds. They’re people working their way up the agency ladder. They’re just put in charge of managing the dashboard. So you don’t really even get that analysis level on top of that either.

ROB: Interesting. Dean, what are a couple things you’ve learned from your experience building Feedback that you would do differently if you were starting over today?

DEAN: That’s a great question. I think the biggest thing that comes right to mind is we went – and this isn’t meant to be like a, “My problem is I volunteer too much,” but we went international way too fast.

It had some fantastic benefits. We actually came out of it with a translation firm partner in Paris that we use today whenever we do anything international, or even something that’s domestic but has multiple languages because they use in-country translators. Fantastic partner, love them to death. Prima Scribe.

But that said, we dove into international because we realized how unique what we were doing truly was, and other people started recognizing it. We got stars in our eyes. I was doing speeches all over the world. We had a lot of international clients, but we really weren’t necessarily ready to walk yet, much less run, outside the U.S.

I wouldn’t say it was a failure by any stretch of the imagination, but I think it was something that – you asked earlier about how what we’ve done has evolved – I think that we probably could’ve stood to learn those evolution lessons first before having gone out so broad, so fast.

We’ve actually pulled back a bit a lot more domestically. We still do things internationally, and we have agency partners that will send us work, but we don’t spend as much time being out there doing that, because it really is rough travel-wise. It takes our eye off the ball domestically. I think we could’ve learned that lesson a little bit sooner.

ROB: That makes sense. Similarly, though, it does seem like you’ve been able to stay in your lane in terms of what you do, what the service offering is. What has given you discipline there, and have there been any brief dalliances outside of that that may have reinforced the need to stay in your lane?

DEAN: That’s a great point. I’d say the main thing is we started out working with agencies. In fact, while it’s 50/50 now, there were certainly points in the last 9 years where it was closer to 80% or 90% agency work.

When you do that, what we noticed very, very fast was we could easily become threatening if we weren’t careful trying to add on services. I will say part of it is this agency universe right now where everyone says “yes” to everything they’re asked to do, and then they figure out how they’re going to do it. We wanted to be careful not to do that, in part because we would eventually end up overlapping with someone and be threatening in that standpoint.

I think that’s helped. The other thing is we have done some – to call it “social media management” is absolutely not it, but we’ll do training where we say, “Hey, here’s what we learned about your audience, and here’s how to go about messaging them” or whatnot. We haven’t necessarily gone full into social implementation in a real forthright way, but we’ve edged along the sides of that a little bit.

But again, part of that is trying not to step on any toes on the agency side with that so that we can remain helpful partners. Even the brands we work with, they have teams internally. While we might be able to help them out in a pinch because there’s some turnover or something like that, we are not out there selling to tweet for them.

We’d much rather tell them how people are talking and whether or not they should even be bothering with Twitter instead of selling the service for a channel that we can’t say definitively that they even ought to be on.

ROB: Right. It’s probably pretty interesting – you mentioned you use Facebook ads as a tool, but you’re not saying. “we’ll do a bunch of Facebook ads for you.”

DEAN: Absolutely not. In fact, in every circumstance for the most part, we’re handing these matrices over to the agency and they’re then putting all their campaigns in the water from there.

But there is an advantage to the agency being the one that gets to bring in – they always say if you can bring in a white hair and glasses, a British accent, or a PhD, you could be saying the same thing you’ve been saying for years, but suddenly they’ll listen.

The agencies have used us like that with their clients. We’ll come in with our matrices and they’ll say, “Awesome, here’s our game plan now.” Now the client believes it because the research company told them that they ought to be doing it this way.

ROB: Do you have any British accents?

DEAN: In a pinch I could probably put one on. [laughs]

ROB: What are you excited about, Dean, that’s coming up for Feedback, or perhaps more broadly for marketing and research?

DEAN: For me, one thing I love that we’ve gotten more into in the last few years is the fact that now we have relationships with some of these agencies over 4-5 years now. Many of them have stuck with us this whole time. Those deepening relationships have only created cooler and cooler opportunities.

What I love about that and what I think it really signals for the future is to me, that’s how these agencies are also keeping their clients. In an age where the Agency of Record is sometimes a little bit of an extinct animal, we’re seeing that agencies that can consistently bring new things to the table – honestly, we’re being white labeled less; early on we were much more frequently white labeled. What I love about that is we’re able to do much cooler work.

We’re able to do our research, and instead of going away until next year, they’re inviting us in to be a part of the first couple brainstorming opportunities because we are experts on their audience now. It’s great to be that voice.

I’ll say this personally, too, as much as I would say this for our company: I love that because to me, it feels like the consumer has an advocate in the room. I love being able to be that voice. There’s something about that that I feel is very organic.

And to me, that’s what makes us beat our perceived competitor of the dashboards. There’s a warmth and a richness to what we find. For us to be able to be those translators for the agency, for the brand or whatnot, to really speak for them I think is a really special place. We’re really turning a corner right now with that being valued, and I really love that. I love what that means for the future for our industry.

ROB: For someone who wants to dip their toe in the water – I’d say there’s almost a layer of what you do where people don’t know what they don’t know.

DEAN: Yes.

ROB: Are there any tools or tactics you suggest? I could picture in my mind somebody using let’s say an earned media reporting tool to start discovering some of the places discussions are happening. I don’t know, are there any tools left that will do search on public Facebook, or is that dying off? What’s out there? What’s helpful for someone trying to get two steps into what you do?

DEAN: You’re right in that public search on Facebook is a little wonky. Even though Facebook wants search on Facebook to be better, it’s not there yet.

My favorite example and way to dip your toe in isn’t “How do people comment on news articles?” The lowest form of life on the internet right now are those who comment on YouTube videos and news articles. [laughs] The problem is that’s what people then think about as the public web beyond Facebook and Twitter.

But look at your localized message boards – things like City-Data. You’re in Atlanta; that’s actually a great example in Atlanta, especially in the South. City-Data is one of those websites that if you’ve lived in the city for any length of time, you probably rarely go on it.

But it is a perfect forum for people who have lived in the area for a long time in forming newcomers who ask about school systems, docks, “Where should I buy this?”, “What do you think about this?” “What do you think of this improvement in the area?” It’s a great example to dip your toe in that.

I’d say tool-wise, there’s a great site called BoardReader. It does not look at all message boards. I wish it did. But it looks at the ones that are based on the ProBoards code – which, not to get too nerdy, is a kind of message board that a lot of people use because it’s free, because it’s filled with ads. BoardReader skims those.

Now, it’s not perfect, but it’s a great example. If you’ve got a client – if it’s healthcare, maybe it’s a specific diagnosis; if it’s automotive, maybe a specific tire fitment, whatever, something you could put in that’s real specific like a Boolean search – and just see what comes up on where people are talking. I guarantee it’ll blow your mind.

ROB: Awesome. We’ll try and get that link in the show notes. Dean, thank you for sharing your journey with Feedback. I think we’ve all learned a lot about what you do and would love to learn a lot more.

A special thanks today goes out to one of our previous guests, Greg Brock from Firefli Agency, who told us Dean would be a great guest – and you have been. Thank you so much, Dean.

DEAN: Thank you so much for having me. Appreciate it.

ROB: Have a good one. Bye-bye.

Thank you for listening. The Marketing Agency Leadership Podcast is presented by Converge. Converge helps digital marketing agencies and brands automate their reporting so they can be more profitable, accurate, and responsive. To learn more about how Converge can automate your marketing reporting, email, or visit us on the web at

Leave a Reply

Your email address will not be published. Required fields are marked *