Customer acquisition costs continue to rise aggressively, and that trend shows no signs of stopping. Today we discuss broadcasting versus narrowcasting to explore alternative strategic approaches that get more with less. Moreover, Mike Ferranti will cover five steps that you can take now to address the realities of aggressive customer growth targets.

What follows is a lightly edited transcript of Episode 5 of the Inevitable Success Podcast with Damian Bergamaschi and special guest Mike Ferranti. (Listen Here)


Customer Acquisition in 2018

Damian: It’s just the beginning of the year 2018. We have Q4 behind us. What are you hearing from marketers about their marketing initiatives and plans in 2018?

Mike: That’s a very timely question. I guess there are a few things: Across the board—and I’m talking to lots of folks—there are a few major themes. For one, this is the year they’re going to get their data act together, and it doesn’t matter if you’re dealing with a hundred-billion-dollar conglomerate or the 50-million-dollar middle market business. One approach is to evolve it to a level that they always wanted to be at; a lot of brands have been working on this for the last few years. It started several years ago when the term Big Data first splashed upon the scene. They want to use that data to inform their relationships with customers to grow revenue and profit.

Another thing that many brands are thinking about is customer acquisition. It was expensive last year, and the year before that, and the year before that, and the year before that. It’s expensive this year, and probably more so as bid-based systems have taken over digital, and digital has continued to take over advertising. The more bidders you have, the more expensive it gets.

Damian: There were recently a few studies—put out independently on Google—that show that the cost per acquisition in each respective industry has been going up materially.

Mike: That’s not likely to change. Competition is going to be greater as a result. The bar is higher. It’s a wonderful thing if your name rhymes with either Google or Facebook.

It’s a big challenge when we have this perfectly competitive situation, just like when there are lots of participants in the stock market, P-E multiples tend to swell. In our industry as marketers, CPA, CPO, or cost-per-customer-acquired tends to swell as there are many more participants. That’s on everybody’s mind right now because in a good economy with a lot of access to capital any brand right now is trying to become more ambitious—because of course they want to grow when things are good. Unfortunately, that means they also have higher costs of acquisition. Brands are really trying to sort that out. They’re trying to use data on things like programmatic and CRM to improve their cost per order and improve the customers that they acquire.

Budgeting for Customer Acquisition

Damian: Do you hear that they are adjusting their plan based on their ideas about where the economy is now, and where it’s going?

Mike: Going back a couple of years, there was some boardroom speculation that things are good until they’re not. I think a lot of brands that we deal with—and this is an overhang since the Great Recession—have ambition, but they’re careful. They wish to couple that ambition with discipline. In digital and in measured marketing that means having great analytics, clear targets, and clear expectations of what marketing should do. That puts some pressure on the marketers as well.

Damian: So what are some of the things that these marketers are doing to deal with that pressure?

Mike: The first thing is that they being a lot more disciplined. They’re doing more testing, and they’re being more thoughtful about what they invest in. They make their goals much clearer than they have in prior years. It’s all about a “cost per X,” and in many ways it’s always been that way. But that discipline has permeated every brand today. A lot of brands we work with want very tight analytics around their Cost X.

Damian: Right and when you say Cost X you’re referring to price costs per customer acquired. Is that right?

Mike: The easy one is cost per customer. But brands are trying to be a little more sophisticated today. They’re beginning to look at things like cost per visit, cost per trial, cost per repeat purchase, and any of the engagement factors or experiences that the brand can quantify as being of value. We call those “actions of economic value.” Those are the things that they’re beginning to measure and assign an economic value to.

Damian: How big of a factor is new customer acquisition in hitting your revenue targets for marketers in 2018?

Mike: For a lot of brands it will play a role—for almost all of them. For some brands it’s going to play a major role. And for the latter, it goes back to questions like, “What is the size of your growth?” “Where are you today?” and “What is your level of sophistication when it comes to engaging with an existing customer base to extract incremental revenue?” We’re talking about the difference between your goal and your reasonable expectation of what you can get out of the existing customer base and existing traffic. That’s your net new customer acquisition. That’s what must happen.

Of course, the challenge there often becomes that you have a budget of X and you must fit Y number of acquisitions into it, which produces a cost per order or a “cost per X,” as it were. That is the challenge the market must face. We must make it work within that range, so that’s what has driven this improved focus and thoughtfulness. It is a greater strategic consideration of how we acquire customers.

Broadcasting Vs. Narrowcasting

Damian: Thinking strategically, I know you recently did a piece where you discussed the difference between traditional broadcasting versus narrowcasting. Do you think you could talk to us a little about what each of those are, and navigate through this topic for us?

Mike: That’s a good dichotomy. I think dichotomies like that are helpful in terms of thinking strategically about where you are and what you need to do next. On the one hand, we have broadcasting, the best example of which is broadcast TV Super Bowl advertising. That’s all about reaching the largest audience you can hope to reach, as efficiently as possible. It’s about breadth. At the other end of the spectrum, we have what is sometimes called narrowcasting, which is about quality and focus. Instead of trying to hit the broadest audience you can, you create a strategy upfront to identify the individuals that are most important to reach. You make decisions up front about how you should reach them. You target them and only spend on that rarefied audience. The reason you might do that is simply that it’s expensive to acquire customers. You need to take the steps allowed by your budget, and you must eliminate waste upfront.

Damian: What is the incremental value there?

Mike: You really can’t judge broadcast or narrowcast as good or bad. And that’s an important distinction because, as you know, digital people are digital natives in marketing. They might be inclined to judge the broadcast approach. But all media has value at the right price. When you buy very large tracts of media your unit cost tends to go down—at least if you buy it well. They used to call that the tonnage model: “Right we’re going to buy tons and tons of impressions but the unit cost is so low that we generate a cost advantage, because the amount of waste in that unit cost isn’t terribly high.”

For example, if I’m McDonald’s or Nike—that’s a large audience. Everyone wears footwear. So broadcasts can really work for a brand like that. It can’t be inherently good or bad. On the other hand, if you have a niche product—maybe you sell an add-on for drone hobbyists. You know you can’t broadcast that because even if everybody thinks it’s cool (which is probably not the case) as soon as you move to a high-priced piece of hardware for an enthusiast, that will be confusing for most other consumers. That’s a very different audience, and broadcast probably isn’t going to work.

That’s a good way to think about the spectrum of broadcast and narrowcast. Which one is ideal for you depends on the brand, on the customer, on the budget, and on the goals. That’s a little background.

I’m seeing a shift toward narrowcasting in many organizations that traditionally did more broadcasting. For example, Tapestry (formerly Coach) is now a portfolio of brands. I think one of the reasons that it makes sense to have that portfolio of brands is because it can help them take advantage of a broadcast approach. There is a brand for everyone—a style, a look, and maybe even a price point that broadens their appeal and makes for efficiency. But at the same time, as they’ve learned more about who the customer really is, the customer that makes repeat purchases, and converts at a better rate, there’s been a trend towards finding just that customer and the best place to advertise. The easiest example to think about that is in programmatic advertising, which has gone through its own evolution as of late.

Damian: So will there ever be an example or situation where you do both as a brand?

Mike: That’s a great question. We surely have brands that do both. The larger the brand the larger their footprint. If they have physical retail in many places they might do both.

A good example would be Amazon. If you think about Amazon you would say that because they’re so big and everybody shops at Amazon, they would have to broadcast—because their audience is everybody. At the same time, if you take a look at their very sophisticated organic search program, they’re doing very targeted narrowcasting. They are reaching their targeted products and categories. For example, they carry scientific equipment that they can only sell to a unique audience—industrial supply. Most people think of Amazon as a consumer brand. Industrial supply, while a tremendous business, is a very different audience. That moves them towards narrowcasting and having more intelligence about how they reach that audience.

When it comes to their existing customer base they’re very focused on narrowcasting. I think that’s another larger trend in general. Amazon, if you go back long enough, did a good job at upending. They were the first guys not to consistently do “batch and blast.” From very early, they were using the ability to look at what you bought, what categories you shopped, and consistently deliver recommendations. Amazon was the poster child for the first collaborative filtering, and they did a great job, as evidenced by the fact that everyone you know shops there today.

Damian: In a previous episode Gary Beck spoke in depth about how they came up with that and how it basically changed the game. It was very interesting.

…So you used TV as an example for broadcast. Are there ways that a channel like that can evolve to be narrowcast?

Mike: It’s happening slowly, but it’s happening. There is a consortium of cable operators who are collaborating on digital set-top boxes. Today there is cable ID. Those devices are on top of your wiring closet or wherever you have them, and the cable box is basically a Linux computer. It has an IP address just like your iPad, your iPhone, and your desktop PC do, and you can do addressable marketing to them. They’re also growing more interactive, so you can measure response as well. At a minimum, they’ll move off of true broadcast, where you hit an audience that subscribes to cable or has a certain channel package. They’ll move to demographic variables, psychographic variables, or intelligence that they’ve developed, and deliver it in the form of a video advertisement or even an interactive advertisement. (Interactive advertisement is still being tested in most markets.) So you could focus on a better buyer for you, and even measure response to it.

Implementing Your Narrowcast Strategy: Target Definition

Damian: So let’s say you know your brand and you want to go about executing a narrowcast approach this year. What are certain things you would start with, or even items that you’d look out for?

Mike: That’s a good question and it’s a good subject matter to unpack a little bit. If you look at online, programmatic was a great example of where we started adding variables. A few years ago our clients would do things like univariate targeting. So, I’d say “Look, we only want to target women. We want to target women heads of household between 35 and 49.”

Damian: When you say “univariate,” is it like a single variable?

Mike: Yes, they started with one and then they started building more complex targeting. There were issues of coverage, so we didn’t have all that information for everybody. But that has improved as more data providers have joined programmatic networks. Now it’s evolved to a point where there’s really a myriad of data sources that you can choose from to define an audience. That said, marketers then must wrestle with scale, the size of their goals, and the size of their acquisition ambitions versus how small they can get when they identify the ideal target.

Damian: When you say “small,” I’m assuming you mean how targeted, how focused?

Mike: Yes. Especially when your organization has the history and experience and the momentum of going big, that’s not a habit you shake so quickly. The first time you do a narrowcasting initiative or a test, you might have really good results but very small numbers. That’s not encouraging for someone who’s used to instant gratification, which comes with big and broad.

Damian: Is it more measurable, or is it like broadcasting, where it’s hard to attribute any sale directly to a campaign—like TV is notoriously difficult?

Mike: Of course it’s far more measurable online, and I think in the future it may get more measurable, but the jury’s still out on how measurable it will become. The challenge becomes to take that broad audience, if you are in a category or a product where your audience can be broad, and really identify the attributes of individuals in that broad audience that you can describe in the boardroom.

Damian: So how would you do that?

Mike: The answer comes back to data. You would have to begin with your existing customers. You might look at the median customer, you might look at the better customers, you might do a segmentation where you identify those who are very profitable more at the median (the median customer), and you might look at those that you lose money on. So, you know what not to buy in terms of the target. That would require a few things: One, you must know your customers. You must be able to personally identify them. That can be done with things like email, phone number, postal address, or you could use transaction data if you’re an online seller. There are programs that can match back to cookies, and there are a lot more options out there today than there ever have been.

What we’ve seen is that a lot of organizations have done tests, they’ve done pilots, they’ve done a Facebook, display, and the results have been mixed at best. Now to be sure there are lots of great success stories—and we’ve had our hand in a few of them—but a lot of folks listening will know it’s not quite a silver bullet, as cool as it may sound. Even then if you’re really on the edge, you’re now measuring, not only did I get a scale but what kind of scale did I get? and did I get a higher value buyer? Then you find out that it’s even more challenging because what you typically pick up is lots of folks in a trial; it’s their first purchase with the brand. Getting to that second purchase is much more difficult than is often expected. You need to use data to solve that problem as well, and to do that we recommend the same methodology—segmentation.

Look at your one-time buyers. Where is their heterogeneity and where is there homogeneity? What makes them similar or different? Then you can begin the process of peeling back, or as we say in the data business, “slicing” the customer base, or that one-time buyer base, and finding which of their common attributes were most predictive of whether they would make that second purchase. Then you fold that learning back into your customer acquisition.

Damian: When you’re doing this, are you using statistics? Or are you making intuitive assumptions by looking at each of these records and saying, “Yep, this seems to be the Upper West Side mom,” and you put a persona around it?

Mike: There is typically an evolution. Most brands start with what they understand. That makes sense because if you can’t understand it, it’s hard to do a good job with it. So yeah, they might look at the “Upper West Side mom” or the “Midwest family” and then they might test and see how successful they were with that. Maybe that becomes the benchmark. That’s not uncommon.

Damian: That sounds a lot like the univariate approach you were just mentioning.

Mike: Right, in that narrowcast to broadcast spectrum, maybe that’s in the middle. That might be a good way to think about it. It surely feels much better than what had been done in the broadcast world. But it’s not the same as true narrowcasting, where you might do a statistical model, to your point.

You mentioned statistics and modeling it out, and that’s something that we would do as well. You can use statistical methods, which basically begin with regression if you remember college stats and layer lots and lots of sophistication on top of that. But we’re trying to identify where individuals coalesce, using those factors or attributes around a highly desired target, and that highly desired target should be the customer that you want to acquire. This really brings us to the state-of-the-art in narrowcasting.

Implementing Your Narrowcast Strategy: Target Selection

Damian: So what are some channels that are starting points after you’ve found the customer segment that you really want? You have a model. What do you do next?

Mike: After you’ve completed the target definition, you move to target selection. That requires that you have an audience or channel as you put it, so that you can identify the individuals who have those criteria. When you do that, you’d probably use tools like Facebook, which is fantastic because of the amount of data they have for every user. Of course, it is also fantastic because they have almost everybody. Another example might be Google and GDN, because they’re watching all of us all the time and now listening as well, with the advent of personal assistants and intelligent speakers. But all of that information is being used for the purposes of targeting an audience. Audience definition—ultimately what that requires is that you go out and do some marketing. Now we’ve identified our target. We’ve defined it. We found them out there in the wild, as we like to say, and we must test it. This is an approach, by the way, that is not entirely new because direct marketers, direct mailers, have been doing some version of that for decades now. With the cost of direct mail being as high as it is they had no choice but to get very efficient at it.

Implementing Your Narrowcast Strategy: Test Your Audience and Your Message

Mike: These are the same types of methods that are now being applied to digital, and with some success. That really comes down to one or more tests. You’re not just testing the audience, because if you have a good audience, we like to think that we’re sort of fishing in the right pond, but let’s make sure we have the right bait. That bait, of course, is your offer, your messaging, your creative. This is important because this brings it back to what many marketers are most comfortable with: the brand, the offer, the creative, the message, as opposed to the statistical methods used to identify the optimal target in the first place. But by the same token, if you’re in the results and performance business, and you’re trying to squeeze yield out of that marketing in the form of net new customers, you must be careful not to lay it all on the target, because that doesn’t work either. We need offers that resonate messaging that connects. We need to be relevant—more relevant than ever, given the amount of marketing we’re all exposed to—and we must give them a reason to try us. In almost every case.

Implementing Your Narrowcast Strategy: Set Your Benchmark

Damian: Because the first time you do anything it’s very unlikely that it’s going to be the best way that you ever did it, what do you do after your first campaign? What are the things you look out for? How does this evolve so that it becomes a strategic approach that you can adopt into your marketing strategy in 2019 and beyond?

Mike: It’s a great question. The very first thing you want to do is set your benchmark, rather than high five in the boardroom because you think it’s good, or wring your hands because you think it’s bad. You make a good point that the first time you do anything it’s not going to be as good as the umpteenth time you do it. But that’s your benchmark. I think it’s important for your listeners to know that we don’t look for home runs or strikeouts. We need to look for “signs of life,” as we like to say. Did we get any customers? What was the average order size? What was the conversion rate? How does that compare to the most relevant meaningful benchmark we had before that?

So, we’ve had clients that say, “I’m converting traffic from my advertising at 8% and this test converted at a fraction of a percent. So, I guess it’s not working out so well. We didn’t get the target right.” Well, that sounds bad. It sounds like a reason for concern. But you must peel the onion a little bit and you must say, “Let’s see where that 8% came from.” When you dig into that you find out that it was heavily influenced, first it was aggregated. Most of that aggregate turns out to be brand search. That is, people who already knew the brand and folks who probably shopped it before. When you do this type of narrowcast target definition, you’re going to exclude anyone that may have bought the brand before. So it’s a very different target in that regard, and it’ll cost more because they’re not your customers already. They didn’t buy your brand before.

Damian: It’s probably a truer metric of how much your new customer acquisition costs.

Mike: That’s a great point Damian, because a lot of brands—especially when you get into the middle market—do not have the experience with what it really costs to acquire a clean net new customer.

To do that in a way that is scalable, no less, tends to increase the cost, because now you’re not going after an audience of 2,000 individuals. Now you’re going to a bigger, broader audience, but one that can and will buy. They have the means, the motivation, the lifestyle, the demographic, the psychographics, to legitimately be a buyer of your product in your category for your brand. So you’re not just benchmarking “Does this target work for us?” You’re also benchmarking your cost of customer acquisition today—and just knowing a number can be far more valuable than making one up or assuming one.

Damian: One of my favorite things to do whenever I work with a new brand is to figure out what does it cost you to acquire a new customer. I find it so fascinating and so interesting how the cost to acquire customers is very different at different organizations and brands. Even more so, I find it’s very rare that a brand really has the number correct, right off the top of their head. Meanwhile it’s probably one of the most important metrics for their ability to scale and go somewhere—and how quickly they get there. And they don’t know that number! So, it’s one thing I would challenge almost every listener to dodissect your true cost to acquire a customer. Be very intellectually honest about what that is. Because you will change the game and everything else you must do beyond that, because you’re dealing with reality.

Mike: You know the old saying, “Spend 90 percent of your time defining the problem and you only need 10 percent to solve it.” What you’re describing is, if you know the true cost of acquisition, then you could begin to define your problem in all its dimensions. Then you could really get to work on solving it. But if you’re trying to solve the wrong problem, you shouldn’t expect much of an outcome.

Damian: And, oh by the way, if you do that you’ll also be a lot better at attributing new customers that you acquire into the future.

Mike: That’s right.

Damian: Mike, thank you so much for joining us today. We have so many more topics—and we say this every time—that we’re going to go very deep, so you know we’ll get you to your own chair and your own seat here soon.

Mike: I think we have some good stuff for your audience to think about.

Damian: Thanks, Mike. If you enjoyed today’s episode we ask you please write a review. Or better yet, share it with another marketer. Be sure to subscribe to the podcast for new episodes. If you’d like to speak to someone about any topics covered in today’s episode, please start a chat with the BG team today.

The Inevitable Success Podcast

Host: Damian Bergamaschi

Special Guest: Mike Ferranti

Mike is the Founder and CEO of Endai, which brings 20 years of marketing, analytics, and technology depth. He has developed solutions and software to major brand clients and niche marketers alike. Mike is a recognized thought leader in the database, search engine, email, and direct response marketing. He provides commentary and analysis to the media, including Bloomberg TV, Brandweek, and DM News. Mike earned an MBA from the University at Albany and an Entrepreneurial Masters from the Massachusetts Institute of Technology.