Glenn: Hi, everyone. I’m Glenn Gow, Founder & Advisor of Crimson Marketing. Welcome to Moneyball for Marketing where we talk about the incredible changes happening in marketing organizations around big data and marketing technology. We feature marketing technology insights from the top marketers in the world. The reference to Moneyball is from the story of how the Oakland A’s baseball team were able to win and win and win because they figured out how to use data and technology to their advantage. If you’d like to learn about how to use big data and marketing technology and marketing to help you win visit us at CrimsonMarketing.com or email us at info@CrimsonMarketing.com. And now on to our podcast.
Today I am very pleased to welcome Brian King. Brian is the senior Vice President of Marketing at Cycle Gear. Brian oversees marketing which includes advertising, sponsorships, PR, social media, and strategic partnerships. In addition, he manages CRM and catalogue for demand generation online and in and in-store. So how does Cycle Gear describe themselves? Cycle Gear is the largest and fastest growing retailer of motorcycle parts and apparel in America. So Brian it’s a real pleasure to have you here.
Brian: Thanks Glenn, thanks for having me.
Glenn: Absolutely. You and I were talking earlier about something that really fascinates me where some companies commit themselves to repeating past promotional activity because of the way they measure themselves, not necessarily because of the fact that it works. Tell us that story.
Brian: Well I think it’s more than some, I think the vast majority of companies do. It really comes down to the fact that we all build previous financial performance into our promotional plans. So every year we’ve got to do better than the year before. We might look at it on a daily basis, a weekly basis, monthly, or quarterly. And we always trying to lap out previous. So you can be running promotions at any given time frame and it can make a significant contribution to the top line and the bottom line of your business, give you an underperformance in terms of ROI. So a typical case, the big that everybody recognizes is that in grocery you can be doing circulars, or free standing inserts, and when you measure them through something like media mix modeling you see that you’re only getting 80 cents back in revenue for every dollar that you spend.
Glenn: Now let me stop you there. So there are two things that jump out what you just said. One thing is I would like you to explain for the audience media mix modeling, just give a definition for that. And then help us understand how anybody could be running a negative ROI campaign.
Brian: Media mix modeling is a tried and true methodology developed for CPG (consumer package goods) companies because they spend so much money. So think Kellogg’s, General Mills etc. SO what they do is they take multiple years of sales history and they do it of multiple geographies and overly on top of that all of the promotional activity going on. And what this does is passes out the apex and the synergies of all the different promotional activity. It can be pricing, TV, direct mail, email. All of these different things that are going in your market at different times. And it can pass out the contribution of each of those. So your baseline might be that you have zero activity in a given market at a given day or week or month, and that would be compared to other markets at that same time that have different activity going on. So the combinations over an extended period of time in a relationship to sales performance is how you identify the contribution of the different media.
Glenn: Got it so this is a sense of a way of doing marketing attribution.
Brian: Correct it’s called top down. We could get into a whole conversation about that as well Glenn, but it’s a way of seeing how different activities give you in terms of the dollar contribution the margin contribution, and then your return on investment. Because you also provide all the costs of those.
Glenn: So this brings us to the ROI conversation so tell us what’s going on there.
Brian: What it also shows is what the natural baseline of the business is. So it knows within these models you also provide store openings, store closings, and you can see what is going in the economy. They often use Federal Reserve information around inflation, employment and all of these other things. So it can separate out first of all what the economy is doing to your business, what your footprint is doing to your business, and then on top of that what marketing and merchandising activities are doing to your business. You could have a company that is doing let’s say a billion dollars in business. And fifteen percent of that might be contributed through all the new marketing and promotional activity let’s say. Within that you can look at how much you spend on all of those promotions or on those media and see how much you get back. So groceries have been circular and free standing inserts for decades. They know that when they stop doing those or completely cut them off, traffic dies, or slows down and you’re going to have a negative comp compared to previous year’s promotion. That though does mean to say that it’s given you a positive return on investment, you actually see something like it could be something like 80 cents on the dollar, but these circular and free standing inserts account for let’s say half of their 15 percent or even more that marketing is contributing. And therefore if you turn it all if that could be a 7 and a half percent negative comp that you can experience unless you can find ways of replacing it at the same scale, that it’s delivering the sales.
Glenn: Right and that’s really the key is finding something to replace. And that’s maybe in your example where the grocers are struggling.
Brian: Correct. And I use grocers because it’s the most obvious example. I’ve seen it in other retail and specialty as well. You have to be more numerate and smarter and test alternatives to be able to come up with those replacement strategies.
Glenn: Alright good. Talk to us about the barriers to making these changes.
Brian: Yeah so there are a couple of them. Finding programs that scale. So a lot of people think that digital has a lot stronger ROI than old fashioned media like direct mail and TV. The thing is they don’t necessarily have the same impact and get people off their couch and into your store and online. This actually is more applicable to bricks and mortar than it is to online. But my experience for example is that I’ll get four to five times stronger ROI from email programs, but then I’ll get no where near the same scale as out of a direct mail or catalogue program. So for example I could only get maybe a dollar twenty or fifty for every dollar I spend on direct mail, but I can if I have a large enough database say in one case twenty-five million active customers, I could deliver 17 million dollars in contribution to sales for the business. Whereas within active email base, say only 5 million. And much lower response rates, there was no way I could contribute more than like 5 or 6 million dollars to the business through all of those email contacts.
Glenn: Got it so what kind of return would you see in think you said 120 to 150 on direct mail, what would you see on email?
Brian: Well I would say 4.50 to 5 dollars.
Glenn: Well that’s a huge argument, that’s such a significant difference. Let me ask you about that though. Some companies that we work with would say hey if it’s 4.50 return I’m just going to put all my dollars in that basket. How do you look at that approach as opposed to even experimenting with something like direct mail, where you know you’re going to get a lower ROI? Would you consider that, or do you do that, and if so why?
Brian: Absolutely you do both. We maxed out the volume of direct mail that we could. We physically could not send anymore. And you get to the point where you’re spamming people, and your open rates go down, your unsubs go up. And even if your unsub rate stays the same, if you’re increasing the volume, you’re actually incrementally making your unsub volume increase. So then you go and say how can I find smarter ways of doing this, and that’s what I’m dig right now is looking to find ways of improving personalization within the email programs, and looking at ways about measuring a better cadence of content with customers.
Glenn: And if I my Brian, I know he audience is interested in personalization it’s a very hot topic, tell us a little bit about what you’re doing there.
Brian: Yeah so, in the motorcycle business you’ve essentially got five ride types, people who ride off-road bikes, people who ride touring, and I won’t get into all the details. But ride types are something of significance. If I send someone who rides off-road bikes information about street bikes, it’s probably less relevant to them. However, people have combinations of rides, and they change the kinds of rides they have. So we need to keep on top of primarily all the transactional data that we have on what kind of product that we think they’re riding. We should also be capturing make/model/year which can give us much more relevant information. But from there we have promotions in all of these categories all the time. So what we’re doing is working with a relatively small vendor that has a personalization platform that will parse out past transactional behavior, online behavior on our site, what they’re browsing and what they’re looking at to identify what promotions are going to be most relevant to them. We have like 200 items on promotions at any given time. So then what we are planning on doing is running an AB test where we include groups that look the same in our standard program versus groups of customers that actually have the content personalized and dropped into the email based on their ride type and we think they’re interested in buying at a given point in time. So that’s going to take probably 3 months before we start to get a read on it, and potentially six months before we get that honed. We’re actually using machine learning in the process, so we’re not predetermining with business rules what people will be presented with, but rather what the relationships are between people’s transactions and what we have available, or what has been presented in those emails.
Glenn: I love the concept of machine learning using data. And let me ask you, I’m going to imagine this is what you’re doing. You have some hypothesis going in but you’re going to let the actual behavior of buyers tell you what are the attributes that seem to respond to particular types of offers. And that’s really going to be the answer to the question to what’s working and what’s not working, did I get that right?
Brian: Correct, and the answer to what’s working or not is that we do it at the macro scale. So are we getting better revenue, margin generation for one type of treatment. So it’s going to be measured, it’s nearly impossible to measure at the individual level, we have to do it in aggregate. The other thing I’d like to point is we’re really trying to skip from doing segmentation down to doing individual context. The technology is there to allow us to do it. And if we put business rules in place we could markup or predetermine and provide the right information to the right customers. So we’re really doing segmentation 101.
Glenn: So that sounds really hard to do. So tell us a little bit about some of the technologies you might be using, or the approaches you’re using and how can you get to that level.
Brian: We’ve been working with a company called More Cloud.
Glenn: And tell us a little bit about how that works because I think a lot of companies get pretty good at segmentation, but then they get overwhelmed when they start to have too many segments, and when you get to a segment of one you’re talking millions now.
Brian: Exactly. Segmentation can be overwhelming and what you’re always doing is, segmentation is a way of trying to get away from the averages of the total mass. So you start to say there are some variables that are different from that group customers. For any segmentation you’re working with an average group of customers, and we use segmentation at levels typically 3, 5, 7 segments are pretty common. Because we can get our minds around that. And we can manage that. And we can target based on that. However, if we skip that level and say use the algorithms to tell us what people and individuals are responding to. Look at their personal buying behavior versus the buying behavior of the group and then present what that individual is interested in. You have to take a leap of faith to do that. To say can these technologies do what we think they’re doing. You can’t explain it simply to your executives or to anybody. So that’s why we do this as an AB test so that we can say hey I’m not going to explain to you what’s going at the individual or segment level, but I can show you a total when I treat people one way with this series of campaigns versus another way with this other methodology. This is the difference in my total performance.
Glenn: What fascinates me about this is you can gain a competitive advantage for a couple of reasons. One is I think you’re somewhat smaller than some of the big retailers. So you can be nimbler, but more importantly you’ve created a culture of experimentation and innovation. So you’re willing to work with vendors who can manage this big data in a way that helps you learn about what’s working and what’s not working. And that can give you a really huge leg up over the big competitors.
Brian: Yeah so we’re big in our market but we’re small as a business. What this does is also the way we’re structured it’s for performance engagement. So we don’t take a huge amount of risk apart from the time we invest. And it’s self-funding because if we get the margin we’re looking for and generate what we’re looking for, that will generate the cash we need for the program. So it means that we’re probably compared to a large organization we’re taking a fly around on this because we don’t have to kick out any other vendors. We can keep our current programs going. And it’s an opportunity with upside. The way we’re structured is supplemental customer database has been built, it connects to a platform that already exists with this vendor, and we’re basically putting the information back in two ways. One to help us with catalogue targeting and two to help us with email personalization.
Glenn: And I’m going to ask you a question I didn’t ask you but it has to do with look alike modeling. And I love the concept, because if you can test what works and doesn’t work inside your current database customers, then you should be able to build the perfect profile of what’s going to work when you go find someone that isn’t yet a customer and know exactly what kind of promotion is going to appeal to them. Tell us if you’re going to be able to do that yet.
Brian: We’re not concentrating on that yet. Where we would go with that would be look into digital. Because they would be having these online browsing behaviors which would look like our best customers. Definitely look alikes we want to get right, so 3 percent of the population are on a motorcycle. It’s really hard to find our customers, and it’s very expensive to recruit them. So what we’re trying to right now is get better value out of our existing base where we know people are riders because thy proven it by shopping with us. And then the next step is looking at look alikes online, but again I hesitated earlier because of the potential scale of that. We will still have to do traditional prospecting where we go on and rent lists, or we use friend to friend programs to generate new customer acquisition.
Glenn: Fantastic, well I’m impressed with what you’re doing and your drive toward personalization. Your drive toward a segment of one, that’s kind of Nirvana, and if you can pull that off I’ll be very impressed. So Brian we’re just about out of time so I just want to thank you for sharing some of the things you’re doing and helping us understand some of the complexities in pulling off this focus on ROI, personalization, and segmentation and I really learned a lot thank you so much.
Brian: Fantastic, thanks Glenn.
Glenn: Alright Brian, talk to you soon.
If you like this podcast, please subscribe and rate us on iTunes and tell your friends about us. You can also go to our website, CrimsonMarketing.com, and sign up for our free monthly newsletter featuring the very best of our marketing insights, featured Moneyball for Marketing podcasts, and one of our favorite features called, “Bad Marketing,” or email me at info@CrimsonMarketing.com. Thanks for listening to Moneyball for Marketing from Crimson Marketing. Have a great week and let us know if we can help you in any way.