Episode: 438 |
Rob Ristagno:
Rob Ristagno on Customer Segmentation:


Rob Ristagno

Rob Ristagno on Customer Segmentation

Show Notes

Rob Ristagno is a Harvard and McKinsey alum, he has shared expertise on digital media and e-commerce with CBS, Forbes, Adweek, and Fox. He is now the CEO of Sterling Woods where Rob and his team developed a data-enabled service, Scout X, that has created over $500 million in Enterprise Value for middle-market clients. Learn more about Rob’s company at sterlingwoods.com.

  • 00:43: Five common flaws in customer segmentation
  • 12:39: The process of B2B segementation
  • 15:53: The three main categories of customer behavior
  • 29:04: Developing the survey
  • 36:13: Building the segments

One weekly email with bonus materials and summaries of each new episode:


  1. Rob Ristagno on Customer Segmentation


Will Bachman 00:01

Hello, and welcome to Unleashed the show that explores how to thrive as an independent professional. I’m your host Will Bachman. And I’m excited to be here today with Rob Ristagno, who is the founder of Sterling Woods group. Rob, welcome to the show.


Rob Ristagno 00:18

Thanks. Well, it’s great to be here a big a big admirer of what you do.


Will Bachman 00:21

Thank you very much. So Rob, I’m excited to speak with you today about customer segmentation, which is an area I know you have done a lot of work in. And I wanted to start with, what are the most common flaws you see in customer segmentation?


Rob Ristagno 00:43

Well, I see it about five things pretty regularly. Most companies feel like they’re they’re doing segmentation, they feel like they don’t have any problems with with segmentation. It’s, it’s pregnant butter, it’s marketing one on one, of course, we have segments. But we often find one of five things is sometimes not in place, or usually not in place. The first you just tick through them one at a time. Absolutely. Okay, the first one we call anyone with a pulse, that sometimes companies don’t really do segmentation, that they that they say, Hey, you know, who would make a great customer for us, anyone who’s breathing, anyone who has who has the ability to pass. And maybe not say that directly, but kind of behind the scenes, that’s what’s really what the thinking really is. And it kind of reminds me of a company. They’re in the professional development space, focusing on CEOs. And they actually leaned into this anyone with a pulse, and the sales team was organized that you get anyone to, to pay for our services and join us. And when they finally said, You know what, we have a new new CEO and new cmo. And it’s time to take a look at our market and actually go out and do some research, we actually found that there were four types of CEOs that were actually interested in consuming the service. You know, things like the peak performer or the CEO who likes to crush their numbers and be ahead of the competition. Or the silent struggler, the the CEO, who is a bit of an insecure overachiever that maybe made it to the CEO suite, but feels like they have to fake it till they can make it. And there were some others. But you can see, probably the wheels are spinning already, that if you know that those are your target CEO segments or personas, how you’re thinking about positioning your product and what marketing channels you use, and, and what benefits your sales team play up are totally different than than anyone with a pulse. And the company was actually able to reduce their acquisition cost by 78%. As a result of this, and they, they dropped some money to the bottom line, but reinvested most of the savings, and we’re able to get some great top line growth as a result.


Will Bachman 02:54

Yeah, and what’s the problem with just saying, like, Hey, we will sell to anybody that is willing to buy our service, we’re open for business doors are wide open, will sell to anyone and your anyone, what’s the problem with that?


Rob Ristagno 03:11

It’s just an inefficient use of your limited sales and marketing dollars. So you’re gonna get a lot of nose. And if you have, especially if you’re small or medium sized business, this is, this is a huge problem. I guess, if you’re an enormous company just want to throw, throw a bunch of resources out there, it’s still inefficient, but at least you have the resources. And I know it’s scary. Actually, as an entrepreneur, myself, I was a little guilty of this, when I first started is you want to cast as wide of a net as possible, because I need the paycheck. 90, I need the income. But it actually is countered, it may be counterintuitive, that the more focus you are the better. And if you try to be everything to everyone, you’re going to be wasting a lot of your time and or money when it comes to sales marketing.


Will Bachman 03:55

Yeah, because you can spin your wheels chasing customers that are really want something much, you know, cheaper, or more affordable, or lower end than you offer. And you’re, you know, if you try to serve them, they’re going to push you to discount and it’s not really what you do and, and you’re gonna and then if you go for customers that want sort of the luxury all in Mercedes Benz, or Lamborghini or whatever, you’re gonna, you know, maybe you can spin your wheels and try to put a proposal together or, you know, meet their needs, but in the end, you’re going to waste a lot of sales time and not get the deal anyways. And so, it’s it’s about figuring out who you can serve and serve well with your offering. So that you can, you know, serve them much more efficiently. Yeah, exactly. Exactly. Right. Okay, so so so customers. Number one flaw is just companies not even really doing a segmentation and just saying we’re open for business, okay? what’s what’s flaw number two.


Rob Ristagno 04:57

But number two is just taking a rudimentary cut at your market, usually what happens is you come up with some sort of Tam, total addressable market for your business. And you look at a very high level vector slice of the market. So for example, you’re only if you’re a consumer business only selling to a certain dem thinking about certain demographics. So you’re thinking about certain age group or certain gender or certain income level. If you’re b2b, you’re thinking only about a certain geography, or certain industry, hey, we have segments, we were focused on financial services, healthcare, and ad tech in the US. Yeah, doing something like that is better than anyone with the pulse segmentation. But it’s you’re leaving money on the table, you can be way more refined thinking not only about demographics, or firma graphics, but more about attitude behaviors, and what actually motivates customers to buy. So I think a lot of companies fall into this bucket that What are you talking about? We have segments, we built a Tam, we, we know that we’re going after these industries or these these demographic groups, but that’s really insufficient.


Will Bachman 06:09

Okay, what’s flaw number three.


Rob Ristagno 06:13

Number three, is we made it up in the conference room. So there’s basically no data or validation behind the segmentation. So they got the memo, that you need to think about things beyond the demographics and the firma graphics. Usually, this happens when you hire a marketing agency to redo your website. And they tell you that you need three personas. So people spend an hour in a conference room and pick them up. We’ve done some some research on this and found usually out of the three segments that that you make up. One of them is a good segment, it’s usually your most common customer, though, it’s not necessarily your best customer, but it’s your most frequent customer. And then the other two are one of them is usually wishful thinking, it’s not actually a segment that exists in your in your market, or it’s not actually a segment that wants your product. It’s just something that you wish wanted your product. And you’re not if you focus on that segment, you will not be successful. And then it’s a bit of a coin flip on that third segment if it’s actually valid or not. So there’s one, another professional services firm that came up with seven personas, when they redid their website, they were all excited about it. But then they realized that none of them were, I guess five out of the seven really weren’t yielding any fruit. So when they stepped back and actually did a more rigorous data driven, research validated approach to building their segments and personas, they got a 30% growth rate their first year and a 50% growth rate their second year. And again, it came back, it’s a little bit related to the earlier point, they’re able to focus their limited sales and marketing resources on the right opportunities so that they could grow their business.


Will Bachman 07:53

Okay, what’s number four?


Rob Ristagno 07:56

Number four is averages lie. And this one is actually my favorite month. Because some companies feel like, we don’t know what you’re talking about here, Rob, we, we we look at more than just the basic demographics. And we survey our customers every year. So we have all this data, telling us that this is the right segment for us to focus on. And usually, if we find that averages lie, what do I mean by that? So for let’s just, there’s a little bit of having to explain math via audio only. But let me let me give it a crack here. For a company in the in the media space that we’ve worked with, we found that on average, if you look at the survey average the entire population of everyone who took the survey, the two things that the average wants the most raw data about the industry, and practical case studies about the industry. And those four years have become as a top two things that that the market wants when you look only at the averages. But when you segment the audience, you realize that there’s a segment that only cares about the data. And they don’t care at all about the benchmarks and best practices. And there’s a second segment who is exactly the opposite. That they all they don’t care at all. They just want the benchmarks and best practices. So if your sales and marketing efforts are are trying to sell this combined data plus best practices package, you’re not being as efficient as possible, because some people like well, I want half of that, but not the other half. And I kind of feel like I’m paying for something I don’t want. So if you can segment your audience, that’s what I do. My averages lie is that on average, it looks like those are the two most popular things, but in reality for one segment, it was amazingly what they want. wanted and the other segment, they didn’t want it at all. Does that make sense?


Will Bachman 09:51

Yeah, no, that’s clear. So if you’re trying to sell a bundle of these two things, then you’re gonna disappoint everybody. And yeah, exactly. Yeah. No one’s excited. And there might even be some things where maybe some smaller segments that really want one particular thing that doesn’t even show up and your average overall. Okay,


Rob Ristagno 10:11

yeah, that’s a great point. And what was often a, we call them a hidden whale, there’s often when you look at the data, there’s someone there who’s probably, maybe they’re only making up 10% of your customer base right now. But they’re very attractive customers. They’re relatively inexpensive to acquire. And they stay for a long time. And you can upsell and cross sell, and they become favorable word of mouth. So while they’re only 10% of your market today, that might be a huge growth opportunity tomorrow.


Will Bachman 10:39

And what’s the fifth law?


Rob Ristagno 10:42

The fifth law is doing nothing about the segmentation. So going through all this effort and doing everything the right way, and analytically validated with data and research and, and machine learning, but then it turns into a PowerPoint that sits on someone’s shelf. And I know you and I both used to work at McKinsey. And that those tiny love doing was was helping clients segment their market. And it would take us five or six months, and we’d go out and do a bunch of focus groups and collect a lot of data and, and slice and dice it and deliver a beautiful 120 page PowerPoint deck. It was great work. But what I would call the client back a couple months later, they would say, Oh, yeah, that was kind of interesting. And maybe, yeah, we tweaked the product a little bit. And we maybe made a few small changes are creative. But we really kind of no one uses a segment names yet. And it has didn’t really get internalized. So that was, that was a huge disappointment. Because of all the time and money the clients spent on doing things the right way. But then they didn’t actually drive to action right away. And actually, I found that if companies go through the process of properly segmenting their market, if they don’t make changes in the first 90 days, the whole thing gets wasted.


Will Bachman 11:58

Now let’s go through your recommendation on kind of the the right way to do this. We’ve talked about some flaws. Let’s talk about maybe we sort of focus on sort of the b2b category. And you can, you know, come up with a sort of a sanitized example type of company, and walk me through the process of, you know, you know, what you would do if you’re running the project of, you know, how do you start a segmentation project? What are the steps involved? And then let’s all go all the way through once you’ve come up with a segmentation, let’s talk about some implementation pieces, and what and what maybe the final state looks like of a company that not only has a segmentation, but has implemented it? Well?


Rob Ristagno 12:39

Sure, sure. So it starts with the premise that we’re looking for a segmentation, based on attitudes, behaviors, beliefs, problems, that we solve jobs that we do, not just the firma graphics and demographics. So to that end, it always starts with around management hypothesis generation, we’re gonna go out and do some research, that we need to be able to have some specific hypotheses to validate or invalidate. So how do we what different segments do we think exists within our market or audience? What do we think our best segments are like? What do they want from us? What are they like? How do we deliver value in a way that no one else can? So then, after you move out of that phase, you do some in depth interviews? Oh,


Will Bachman 13:28

hold on, hold on, let’s, I’m a little slow. So I want to I want to focus on this a little bit more so. So this initial so let’s say this is sort of phase zero, or phase one of the management hypothesis generation piece of saying, before we go out and actually start doing some research, the people in the room probably know something about the business. So let’s get the ideas on the table. Walk me through a little bit more what that conversation looks like that workshop looks like. What questions are you asking? How would you do that session? And what what should some ideal output of that session look like?


Rob Ristagno 14:08

I think that the questions to ask are, how would you describe our ideal customer? What do they want? What do they like? What problems do we solve for them? How do we do that? And that’s cheaper in a way that’s cheaper, better, faster, less effort than any of our competitors. Then we’d like to ask Who do you think are some of the hidden segments in our audience? Who are the sneaky segments? Those hidden whales that I alluded to earlier? Who are people that you think make up a small percentage of our, our market today or our customer base today, but we have some real expansion opportunities there. And yeah, then coming out of that, we would come up with a list of hypotheses. So we think that these are the types of problems that That we’re best at solving these the types of people who have those problems. And this is why we’re better than anyone else. So that we can go out to the to the actual customers and validate or invalidate those hypotheses.


Will Bachman 15:10

And you mentioned some other things earlier. So you’re looking at, you’re trying to get these executives to share, what are the different behaviors that each of these types of customers might have? So maybe you are and what are the different beliefs that they have? What, what are the different problems they have? And it might be difficult to say, well, there’s just one persona, who has this one belief in this one behavior in this one problem, and so forth. But it sounds like you’re almost trying to come up with a list of different types of problems, different types of behaviors, different types of beliefs, are there other categories that I’m missing there?


Rob Ristagno 15:53

Those are the three main ones, although we started doing more with values. And this is something that’s a 2021 innovation. For us, we found that companies, you can find customers that share some of the same values as a corporation are much better customers over the long run. So when you think about this, your your company stand for integrity, or hard work or creativity, innovation, just kind of think about what your company stands for. And is there a connection or mismatch with your ideal customer segments or not?


Will Bachman 16:27

I’m curious about that one. So everyone would say, Well, I stand for all those things. I mean, they’re all good, like creative creativity, sure. Innovation, sure, hard work, integrity, diversity and inclusion, you know, treating our people, right, when our customers, right, being nice, the environment? sure we’re How do you tease out and force people to kind of make tough choices? Or do you have some kind of, you know, rubric, or something that would say, Well, you know, which do you value more highly, like, following the rules or in procedure? Or, you know, being innovative, right, like, which would you have to do? Do you have some mechanism to kind of force do some forced choices there?


Rob Ristagno 17:07

Yeah, I say, the easier question for people to answer is actually the opposite. So what are after, and you can focus on the actual employees. So what are attributes of an employee, that would be clear to you that they’re not a fit here, they should probably leave. And if you can get some consensus on what that looks like, maybe by sampling some, some key managers, or even key key individual contributors that are there that really understand the business. And then I think, figure out what the opposite are, is figure out what the opposite is, of all the negatives that rise to the top, and you can’t just put a knot. So So if people say something like, people who don’t like change, don’t do well here, really force yourself to not just say we like people, we want people who like change, like maybe a good opposite of that would be intellectually curious people. And now you’ve have a value of intellectually curious. And then you can screen when you get to looking for clients. Do we really want to work with with customers that are setting their way and refuse to change? I mean, that’s going to be very difficult consulting engagement, if you’re trying to open up people’s eyes to, to, to new ways of doing things, new ways of thinking, and you’re constantly getting met with. But here’s how we always did it. I think not only would that be a rough employee experience, that would be a rough client engagement as well.


Will Bachman 18:30

What are the different types of we talked about? There’s beliefs and behaviors? What are the different, like, examples of what the output might include for those different? Like, what are the different? What’s the range of choices?


Rob Ristagno 18:44

Yeah, so for behaviors, it’s a it’s a lot, I just might call these trigger events or activities that someone does that signals to you that they’re in the market for what you offer. So what are they reading online? What are they reading offline? What are they listening to? Where are they? What are they doing to free time? Are they engaging with your you in your brand? and How so? Are they coming to your events? Are they taking calls from your sales team? Just thinking about all things that could signal? Hey, I have this problem, and I want you to potentially be someone I consider to solve it.


Will Bachman 19:19

Oh, interesting. Okay, so that’s a little bit more straightforward. Yeah. Okay. So that is that it’s huge. It’s a little softer. Yeah. Okay. So the behaviors is not any questions about, yes, the behaviors is not like, just every behavior in general, like, Oh, this company pays their bills on time or whatever. But it’s more about what are the indicated kind of behaviors that would signal that they are looking that you could go say, Oh, they visited our website, or Oh, this person is searching on this search term, or the person is going to go to this conference. So you know, those are sorts of things where you could go view visible see that someone doing those things which would alert you to maybe I should you know, maybe this is a For the possible customer for us. Yeah, exactly, exactly. Okay, now that helps me understand much better. So behaviors and then beliefs, what are some of the beliefs?


Rob Ristagno 20:10

Yeah, beliefs is a little softer. But I think this is a little bit related to your point on the behaviors. It’s what what attitudes or beliefs Do they have toward your category? So I think that a mistake that I see happen a lot is people have general beliefs. So just because your Facebook can tell you if someone is religious and just making something up, but just because someone is religious, it doesn’t mean that they’re necessarily going to buy your product. Unless you have a product related to religion. If but if I’m selling aircraft engines, chances are the fact that someone’s religious doesn’t really affect me. So it’s not general psychographics. It’s not general preferences and likes, it’s attitudes related specifically to your category. So do they view what you’re offering as something that helps them develop professionally? Does it help them save? Save time? Did it does it help them make more money? Does it help them feel more important? Does it just kind of just think about some of the attitudes they have towards toward the categories? It’s something that they’re excited about purchasing is is something that they they they dread purchasing? And and really understanding that helps you craft the right. Go to market approach?


Will Bachman 21:26

And the right okay, and the right language to speak to that belief? Yeah. Okay. Yeah. All right. Great. Okay. So that we kind of doubled down on this section here. So but I really helped me understand what that workshop would actually look like. So So now we’ve we’ve completed sort of phase one of doing the the management hypothesis driven piece, what what’s the next phase in your in your customer segmentation journey.


Rob Ristagno 21:54

The next phase is to go and talk to actual customers. And what I’ve learned over the years, I’ve been doing some former market research for 21 years now, is that the best, the most accurate source of data is in depth one on one interviews, in depth interviews, it is for short, I found that focus groups. And by the way, I’m very focused on segmenting customer bases for medium to high consideration. products. So things especially in the b2b side, things that people have been thinking about, they just don’t buy them on an impulse. And even on the consumer side, things that are are higher commitment of higher price tags or require more effort from from the consumer to make the decision. So we’re not talking here about Wrigley’s and gum and the checkout line. For the for those cases, maybe focus groups has have some some benefits. But what I found is that one on one interviews gets you the most accurate data. And the other thing I would say is, you don’t have to do a lot of them to be to make sure that you’ve covered all your bases. And I have a bit of a math background and always are concerned about sample size. But there’s some evidence to prove this out there. But depending on what problem you’re trying to solve, usually five to 10 interviews are more than enough. And that sounds like a shockingly low number to many statisticians. But, but when it comes to in depth interviews, I’ll give you a different answer, by the way when we get to the quantitative part of the study, but for the qualitative part, we’re just trying to come to write a good survey, that’s really the goal of in depth interviews, is to write a good survey, meaning that you thought about a few things you thought about what customers actually what words customers actually use to describe the hypotheses that you’re validating. So maybe you and I just just talked about intellectual curiosity. Maybe a single customer doesn’t say intellectual curiosity, but they say things like innovative or creative or left brain, right brain, so you’re just kind of capture verbatims of how your your customers are describing their attitudes, needs, beliefs, values, activities, actions, so that you’re reading a survey and their terms, not ours. And then we’ll always find that there’s probably some things that the management team missed, that there’s some opportunities outside the boundaries of normal that there are other problems that we solve that we didn’t realize and solve. Or conversely, we think we’re solving this problem, but no one else actually feels like that’s a huge pain point. Or sometimes I feel like you have that pain point. But that’s not how your product works. So it’s these, these interviews are very critical, but I always want to underscore the goal is to make sure that you write a good survey, you’re not going to draw any final conclusions from from doing five to 10 in depth interviews.


Will Bachman 24:50

Okay, boy, I totally agree with you on the power of touching the market and actually just talking to people who are close summers, I wonder if it’s a little bit personality driven, it might be that my personality is more drawn to that form of engagement, which is maybe how I ended up, you know, having a one on one kind of podcast. But, but I certainly feel that way myself where you learn so much so fast when you just start talking to people and just drop the secondary research. how, you know, there’s obviously different ways of getting at, and setting up those sorts of discussions with customers, you can ask the company to make warm intros to some of their customers, but then it might be a little biased. You can use expert network like alpha sites, or GLG, or third bridge or index dot one or one of those. You can go try to source them on LinkedIn, what are some of the ways that you found most effective at setting up those customer interviews?


Rob Ristagno 25:55

So we have, I would say, a non traditional point of view on this. And I think textbook marketing researchers might be a little concerned at first. But what we found is for this round of interviews, it’s always best to focus on your own customers. And you’ll lose alluded to at some potential bias. You’re right, there’s a little bit of circular logic that there already are customers, and what are we missing out by talking to people to people who already bought in, but what we found is that risk is outweighed by the fact that your customers actually know what they’re talking about. And sometimes you talk to people who have limited experience with that with certainly with your products, but potentially even your category, especially if you have a more unique niche that you’re playing in. So we’ve found for this round, it’s ideal to start with your own customers for the interviews. Now, what we usually do is, we do a round of research on our own customers. So we interview and survey our own customers to actually build the first pass at the segments. And I’m getting ahead of myself in the process here. But I just wanted to, to answer this point. And then if there’s some skepticism about Did we miss anything, or or not, that’s when we, we redo the process with external parties. And that’s when we bring in the the the recruiters to find us people to interview and take our surveys, the firm’s that you alluded to, but we find that it’s more predictive and accurate, the data you collect from your own customers, and then you can manage that slight bias that there might be some circular logic by validating it externally, after you’ve already gotten some really rich, robust, detailed and insightful feedback from your own customers who know you and love you.


Will Bachman 27:41

Yeah, no, that makes sense. I’d say, Yeah, I probably would have the point of view that if you can talk to, you definitely want to talk to obviously customers who know what they’re talking about and know the space. And it certainly is nice if you can talk to people that considered your client, you know, that were, you know, considered them that, that did an RFP or did some kind of process and considered the company you’re serving, but then chose someone else? Because, because then that gives you real understanding of maybe a segment that you could be addressing but somehow aren’t aren’t met aren’t measuring up. But yeah, I mean, the companies that you’re already serving, will tell you maybe, particularly if you have a third party, like a consultant talking to them, they’ll sometimes open up and just be very transparent about what they like and what they don’t like, and, and so forth. So okay, so we’ve identified some cosmetics,


Rob Ristagno 28:38

right? Well, by the way, I should, I should have expanded it to include existing customers and prospects. So I think you’re right, I do agree with you. It is great, even in this first round, to talk to people who rejected you, or didn’t buy from you. But they do are familiar with you. Yeah. So I think that’s thank you for bringing that up. That’s an important point.


Will Bachman 28:55

Okay, so let’s, um, so, so you do a few interviews, and then it sounds like what’s your next phase? It sounds like you’re going to get quantitative on us.


Rob Ristagno 29:04

We’re going to do quantitative next. And we found that the simpler and shorter the survey, the more accurate it is. So at this point, your job as the consultant is to make sure you’re managing expectations that if we send out 100 question, survey, especially to busy b2b people there, but even consumer people, they’re just going to click, even if they finish it and take it, they’re just going to click through and the data is not going to be very accurate. So this is where we need to be very economical, I guess about about how many questions we’re adding. We’d like to see a few mix of questions on the survey. We like to keep it to between 10 and 15 questions, all multiple choice. We’d like to see a couple questions around all the things that we’ve just been discussing. So a few questions around. What problems are you facing in the space in general And then kind of maybe probe a little bit on some some of the hypothesized problems that they’re facing. And then we kind of get into how they look to solve those problems in general. And then we get into our own their perception of our own brand. So we don’t talk about ourselves. So question five or six, it’s really just about what problems you have in general, how you’re solving those problems in general. Okay, what do you think about us, and then depending on the aim of the of the project, those about it the way can throw in other even though I kind of poo pooed demographics and demographics, you still need to know some of them. When you come to your your media spend when we get to the implementation phase, so it’s okay to throw in a couple of segmentation or firmographic, demographic segmentation questions?


Will Bachman 30:54

What are your favorite ways to ask their size? Or what would you ask?


Rob Ristagno 30:59

I think on the b2b side, yes, size is the most usually the most predictive. We rarely known Billy’s name, but we rarely see when we ask industry, we really see statistically significant differences in segmentation based only on industry. Almost all, sometimes there’s some correlation with size. But geography and industry, I’m not saying don’t don’t think about industry, but but usually the segments, the problems that you have in financial services are the same problems you have in healthcare. So yeah, maybe you need a sales rep who can who understands financial services, so they can talk the talk with the CIOs of financial services, companies are different than the CIOs of, of healthcare companies. But the underlying root cause the root problem that your company solves, typically is the same no matter what industry they’re in. And that’s a little surprising, because most people think that the first thing they segment by is industry. Yeah, well, I mean, probably, again, I’m not advocating you don’t think about industry, but but when we come to this use case segmentation, there’s very little correlation with industry.


Will Bachman 32:02

I mean, it probably depends a little bit on what product we’re talking about. Like, let’s say that, I mean, if you sell like, Oh, we sell the inner layer cloth that goes into conveyor belts, you might say, you might find out Well, there’s like the mining industry customers, and like manufacturing plant customers, where they have conveyor belts are two very different segments. And we actually don’t serve any financial services. Right? Or, and maybe like a third one would be like, logistics, or warehouses or something. But, you know, so it probably depends on like, the product or service, but it sounds like, you know, for some for some things like, Oh, we have like HR software or something that might might not matter so much. Good point. Good point. But, okay, so we talked about little demographics, and what are some of your other like, specific favorite questions to ask and, you know, just have any, not just from a graphics, but in these kind of surveys, what are give us some sort of sample questions of how you phrase different things to ask either about what’s your opinion,


Rob Ristagno 33:08

like leading off a sort of like, what’s your biggest focus this year? Or what are the one or two things that are gonna make or break your year, this year,


Will Bachman 33:14

and then you give like, you give you give a checklist of options to choose from?


Rob Ristagno 33:20

Yeah, it depends on, on which, what but what your category is, but just an example is we just did a segmentation project for a company that sells to marketers. So you know, we asked what is what is your biggest focus over the next 12 months? Is it brand health? Is it campaign performance? Is it figuring out martech and data? So there’s just a few flavors there, but just obviously very different. Those are things that all marketers are concerned about some some degree, but it’s important to force people to prioritize, say, like, what’s most important to you added these choices. I think that’s actually another good point to make is that we’ve had a lot more success and a lot more predictable results, when you ask forced people to prioritize. So I know from taking a lot of surveys, things that are not as effective questions that Select all that apply and let you pick all of them. That’s not super helpful, because you’re not forcing the respondent to make trade offs. And the other thing that doesn’t work that well, from our experience is, I think it’s called a Likert scale, asking people to rank on a scale of one to seven. Yeah, so instead of asking that question, what’s most important to you? Is it brand health or campaign performance or martec, and data. If you said Select all that apply, your chances are everyone will pick all three. And chances are, if you asked to rank each of the three on a scale of one to seven, they would all like average out to like 5.8 or something like that. So forcing the choice is a really important mechanism when you’re writing a survey and also helps keep it simple. You Don’t ask, it’s that would really be three questions in one, if I asked them to write one to seven right now just ask for one choice.


Will Bachman 35:07

Right? Right, like so either pick your either rank your top three choices, or pick your top two choices, something that forces some choice. So they don’t just say, Okay, well, everything’s a six or a seven. And then Exactly,


Rob Ristagno 35:19

yeah, and there’s a rule of thumb is, if there’s less than five choices for thinking that you have to force them to pick just one. If there’s like five to seven, or eight, maybe pick two. And if there’s a to 10, you can let them pick three. But try not to have more than 10 choices, because then it just gets to be a long, long survey, or a lot to read. And, and you This is why the hypothesis generation and the interviews are so important, because that helps you cut off paths, like there might be your 12 things you want to test or 12 things you want to ask, but you only heard six of them in the interviews. So you know, unless there’s one or two that you really are interested in, don’t bother asking the five or six that never came up.


Will Bachman 36:04

Okay. Okay, so we do the survey. And now let’s say we got the survey results back. So what’s what’s the next step?


Rob Ristagno 36:13

The next thing is to apply AI, I’ll be jargony here for a second unsupervised machine learning to the results, to build the segments, and and outside what that actually means in plain English. It basically let there’s algorithms out there, for those are data scientists in the audience here. Things like principal component analysis, k means clustering, there’s techniques out there to actually let the computer build the segments. And we highly recommend that because you take all management bias and your own personal bias out of the decision making as to what what the segments should be or are. And what these models do with these algorithms do is they’ll just cluster people, group people, based on the questions that have the most level of the highest level of variance, the most differences between them. So they’ll be able to fall apart the averages that I picked on CES so strongly at the beginning of this interview, and really build 34567 segments that are statistically significant, statistically significantly different from each other. And so that, you know, these are real, different customers. So you let the computer do do its job. But then you have to bring it back to life and personify it for to be actually meaningful to the client. So let’s look at for each segment, and how do they How do they sort out some problems with that question? And


Will Bachman 37:38

this is, so this is something that I not familiar with at all. If you are a machine learning, newbie, is it possible? Is there software out there where you could say, Okay, let’s say I have the results in Excel, that you can go to some, you know, get some tool to do this? Or do you need to, you know, hire someone off of Upwork? Or something who’s, you know, data scientist or machine learning person? Or how would you actually go do this unsupervised machine learning? To Yeah,


Rob Ristagno 38:12

you’d have to, you know, hire a firm like ours, who has the algorithms all built, or you could go on Upwork, or any of these other platforms to find data scientists or statisticians, that they’ll know exactly what I’m talking about, they’ll be able to do it for you,


Will Bachman 38:28

and what kind of person like what would they even call themselves? Is this something that a data scientist would typically do? Or? Or? what’s what’s Yeah, I


Rob Ristagno 38:36

think I would, I would, it’s kind of an aging term. But statistician is, is, is good enough for you don’t really understand statistics and can do this data scientists. It could also do it, but you might be overpaying for them. If they’re true data scientists, not only can they perform these statistical techniques, but they can also understand how to wrangle data and write the code and stuff like that. So it might be I mean, people might all statistician may be calling themselves data scientists at this point, from a branding standpoint. But But you need you need someone to give us statistics. And if you say principal component analysis, and K means clustering, and then most people with statistics degree, you should be able to know what that means.


Will Bachman 39:15

All right, cool. So let’s say either, you know, your firm has done this for someone or you hire a statistician. And then they come up with Okay, these are the six groups and they tell you these 30 respondents are in Group A, and these 40 on groupby, and so forth. But then how do you go from that to an actual persona, to identify what what’s what’s, what’s current, what’s the common characteristics of this group that this, you know, unsupervised machine learning as bundled together for me.


Rob Ristagno 39:50

So how we bring it to life is through a workshop and it’s, I would call it a gamified workshop. And the idea is to A little bit bored, that they actually like it in the end. But basically say, look, we’re gonna walk through 100 charts today. And on each chart, we’re gonna go one by one, all the questions that we asked for each of the, let’s just say three segments to keep it simple. And for each chart we’re gonna see is a few things that I want you to keep your eye on. One is how that say how that segment responded on an absolute basis. So if we ask, what’s your what’s your your biggest pain point? What percentage of them thought? Just to use our marketing example here? Again, what percentage thought that their biggest opportunity was in brand brand strength? How did the segment score? Do they care most about brand strength, they care more about performance of their campaigns, or they care more about marketing data. So you want to look at the absolute highest scores. But it’s equally important to look at the difference from the mean. So on every slide, we’re going to tell you, not only how did this segment, respond to this question, but how each of the responses compared to the population average, and it’s an index. So if this segment, says, For example, most of them care about brand house, but every other segment also cared about brand health. And that’s not a dimension that we’re going to be segmenting on, if we found that this segment, really cares about brand house twice as much as any other segment, then there’s a good chance that this is going to be a segment that is focused around getting brand health, right? Does that make sense?


Will Bachman 41:31

Yes, it does.


Rob Ristagno 41:34

So then what we found is 10 times out of 10. When you walk everyone through the math, they are making it extra brilliant. It’s my favorite part of the project, everyone makes connections in their head. So what might first sound like a daunting statistics course, ends up becoming a lot of fun, because by the end there, they’re throwing out names for the segment and connecting the dots and say, Oh, yeah, that’s just like bill that I talked to, and he, he cares, All he cares about is his brand. And he’s not looking at the data at all. And he’s just fine by the seat of his pants, and you know, stuff like that. So really that methodical, gamified approach to say, we’re going to go through this methodically, slowly, chart one chart at a time. And then we’re going to end with a creative brainstorming session to bring these these segments to life. I found that the most effective way to not only communicate the segments that we found, but also get the team excited and bought into them. So that they’re already thinking about activation, which is going to be the next part of this interview, I bet. Yeah. But like they’re already thinking about how they can grow their business now that they understand who their segments actually are.


Will Bachman 42:38

On that note, let’s talk about activation. So what are all the things that a firm needs to think about? Once they’ve come up with segmentations? That they’re all happy with, in terms of, you know, what you do with your website or other marketing cloud or your pricing? Or your your sales team dividing it up? Or just like, what what’s the range of things that we need to think about doing now?


Rob Ristagno 43:02

Yeah, so I say just, I was just thinking about a marketing funnel. So think about what are all? I’ll give you an example in a second, let me just see the funnel, though. Let me just think, what are all the things we can do to generate more leads in the segment? What are all the things we can do to convert the leads we have in the segment? What are the things we can do to increase the customer lifetime value of the other customers we have in the segment? And then maybe, you know, what are new products or brand extensions that we can launch that are appealing to the segment. So you got the sales, you got marketing, you got product development, and then just have a rapid fire brainstorming session with all those four components. So for example, I alluded to the professional services company targeting CEOs at the beginning of this interview, and one of their segments was kind of the silent circular, the insecure overachiever. So for example, top of the funnel, how can we get generate more leads here? But what about white paper that we promote on LinkedIn called first time CEO Survival Guide? In terms of converting them What about we create a new sales script for our sales team? Once we identify someone as a silent struggler? How do we ask them? Here’s the questions, we asked them to show that we can empathize with them. And that they really understand Yes, you get my pain points. And here are the here’s the value proposition tailored to the segment, that we should be communicating to them when we’re on the phone to play up the benefits that they actually care about. And then just the retention, how do we make sure that within the first 60 days of purchase, that they feel like they’re more secure and more confident in what they’re doing? And what does that onboarding process have to look like that customer experience have to look like to so they don’t get post put what’s called post purchase dissonance that they were they were told all the right things by sales and marketing, but the product didn’t deliver. And then what are some other rotors and brand extensions? How can we think about other products and services that would be useful to the segment to help them be better CEOs. So you just do a bunch brainstorming there. I’ll talk about how we prioritize all the ideas. But why don’t I pause there to see if you have any questions?


Will Bachman 45:06

No, that’s good. So, yeah, let’s, let’s talk about how we prioritize and then kind of work to bundle it all and implement it.


Rob Ristagno 45:15

Sure. So we found there’s three things that matter. When, when ranking ideas, I think we could have a whole episode about this. Well, if you want, how we came to these three dimensions, but that’s a different topic. One is, is it is it? Are we certain that it’s going to that it’s related to what customers actually want. And if you’ve paid attention during your whole market research project and segmentation project, nothing should be on this list that doesn’t meet this criteria. But just make sure that just in case, someone had a creative idea, if it’s in conflict with any of the research, that you just did cross it off, because it’s not going to work, because it’s not validated by any of the research and data that you just collected. The second one is feasibility. And we’ve we’ve created a persona called the dreamer. And it’s an executive who has great ideas, and they’re brilliant, and they’re going to do awesome stuff. And if they were venture backed, and in Silicon Valley, God bless them. But you have to also be realistic about what resources your company has, what skill sets your company has. And back to the point earlier that if you don’t take action on your segmentation, in 90 days, it kind of withered on the vine. So think about things that are practical and actionable and accomplishable, with the resources that we have in house today. Either existing employees or existing consulting relationships, existing agency relationships, or even potentially, as you talked about finding people on Upwork, or through Umbrex, or whatever, they can help us get some of this stuff done. And then the third dimension is newness. So we found that, if you’re doing the same thing that you’ve always done, or the same thing that your competitor is doing, you’re not going to get the results that you want, versus doing something that’s novel. So really put on cap and come up with ideas that have I’m not talking about sending a person to Mars, like it doesn’t be that crazy of an idea. But they should be tactics that are different than what you’re already doing. Otherwise, they’ll just get the same results that you’ve already gotten. Okay.


Will Bachman 47:17

And then any any final thoughts about you know, how company should go ahead? And, you know, get this in place?


Rob Ristagno 47:25

Yeah, I mean, it comes down to the scoreboard. So the last thing to do is once you’ve prioritized all these initiatives, keep score, like what is what is the goal? And see if see what’s what’s working, and double down on that. And then, if it’s not working, look at the data and ask yourself why it’s not working and figure out how you can course correct. And the idea is keep the eye on the segmentation over time. So continuing, continue to do a small amount of surveying and sampling to see two things. One, how are the segments trending in terms of size? Are you getting more segment A people? Are you getting more of these silent strugglers or not? And be an avid NPS each segment. And that gives you that’s a leading indicator to see which segments are becoming more and more loyal to your company and which ones you may have an issue. You’d have to go back to those four levers. We talked about figuring out if you have a, you know, if you have a product problem or an experienced problem or marketing problem or sales problem.


Will Bachman 48:24

Yeah, and that’s NPS Net Promoter Score, of course. Rob, thank you so much for joining today. If folks wanted to find your firm online and follow up with you, where would you point them? Come visit us at Sterling woods, calm. Fantastic, Rob. Thank you for joining. Thanks well

Related Episodes


AI Project Case Study

Paul Gaspar


AI Project Case Study

Astrid Malval-Beharry


AI Project Case Study

Julie Noonan


AI Project Case Study

Markus Starke