A Topline on Analytical Approaches (#604)

Bill: Hello, everyone. It’s Bill Hanifin. I’m the managing editor of The Wise Marketer. And I’m here today with Aaron Dauphinee, who’s the CMO of The Wise Marketer Group. Aaron, how are you?

Aaron: I’m well, Bill. How are you doing today?

Bill: Good, good. We gathered here for a purpose, right? And the purpose is The Wiser Loyalty Podcast series, which is sponsored by Paula Thomas’s. Let’s Talk Loyalty. And we’re so proud to have that partnership with her every, every month throughout this year. We’ve been tapping into a course in our Loyalty Academy Curriculum.

And we are going deep. We’re kind of, pulling out the, the most important parts of each course, like sort of the, the pillars and the, and the key points. And then we’ve been building on those. We’ve been challenging some of the old thinking we’ve been building and, and bringing things up to date and, and just telling a few stories along the way.

So this month we’re building around course number 107, which is our Introduction To Loyalty Analytics. And we’ve had two discussions already. We’ve talked about sources of data. So where does data come from? And then last week, we talked about zero party data, which we got into a pretty fun discussion about how all that originated the big discussion around it.

And and so this week, we’re going to talk about analytical methods and there, there, there’s a little bit of a difference between what I would call straight business analytics and the analytics that we do for customer marketing and for database marketing. And I think to me, there’s there and there’s 3.

Kind of areas of benefit, like, why do we do analytics? I hate to ask such an obvious question because there’s an audience, but it’s obviously we want to be able to rationalize the investment that we’re putting into customers, rationalizing investment that we’re putting into our marketing budgets. And we do that by optimizing.

The marketing campaigns that we execute, so we need to understand the customer. If we’re going to be able to understand why we’re spending money on certain things and be able to measure the success of those things. We also needed to identify the friction points that exist. So, if there’s areas of user behavior that you can see through analytics where there’s fall off, it could be shopping cart abandonment.

It could be a lot of different things could be attrition. You need to find those points of friction and be able to create a strategy then around it, obviously, to try to solve for a better outcome. And then the other, the big 1 probably is just create happiness. We want to increase customer satisfaction and loyalty.

So, you know, those 3 things, if I had to. Kind of narrow it down. Those would be the main areas where there are objectives, I guess, in marketing analytics and but there are different types. And I think it’s important to kind of walk through the different areas of analytics. It’s not just 1 bucket, right?

There’s probably 3 or 4 different areas of analytics. So, I mean, how would you walk us through that?

Aaron: Yeah. Yeah. And before I walk through kind of a lay person’s view on analytics versus some of the individuals who much more student this particular area in our, in our sector, but 1 of the comments I just talked about is.

You know, we talked about data driven marketing as, as a whole, as a body or a philosophy that we, we have as loyalty marketers. And, and some stats on that of why it’s important. I think Forrester recognized a while ago that data driven companies are 58 percent more likely to beat their revenue goals.

So there’s the, Hey, you know, this, this is why we keep tabs on the numbers sort of thing. And then IBM also came through with a discovery that 62 percent of retailers have reported gaining a competitive advantage. From information and their data analytics. So, you know, don’t just take it from us. Like this is, this is there are actually brands that are very, very astute in terms of the data that they collect and how they and the prowess that they have around their analytical capability.

And so there’s value in the capability. Making sure that your organization is customer centric and then by virtue of that, you’re collecting data to be data driven in the pursuit of providing relevancy for your customers. So that’s just kind of a little you know, don’t take it from us, but there’s others around there.

You can also draw and there’s many more stats, but I just wanted to draw a couple before I hopped into, talking about a bit of a few of the different types of of analytics that exists and and this is true for both business as well as consumer consumer analytics and and I think the the four and I’ll and I’ll keep kind of a high level lay person’s definition on these.

I won’t be too particular on them, but the 1st that I think about is, of course, descriptive analytics. So when you think about descriptive analytics, oftentimes you’re looking at historical data. Sort of information about your consumers to kind of come up with understanding what the trends are based on that past performance.

So, you know, in a business sense, you’re looking at KPIs in a consumer sense. You’re looking at the metrics that you want in terms of movement, whether it’s a frequency, recency or monetary value in terms of spend. Those types of things come into play in terms of being very descriptive, but you’re looking at past events essentially.

For future planning and decision making around your customers is number one. I think the second then is a kind of diagnostic analytics. I guess you’d call this is where you’re kind of investigating the reasons why these past outcomes have come into play. And so you’re looking for root cause, right?

That’s a good way to describe this type of analytics. You’re looking for the specific results. So, you know, Bill went into a store and actually, you know, not even that a better example is Bill went on a trip and he redeemed for his flight. What’s the reasons why, or what are the reasons why that Bill actually redeemed for that flight?

And can we have an understanding of, does he have a propensity for actually vacations that are more around, sun and sand or he has a propensity for vacations that are around walking around cities and seeing and touring or is it culture or something that like that. So, so that type of diagnostic to say, hey, we looked at bills last 3 or 4 redemptions and they all seem to be sun and surf.

Right? So, so that’s kind of a diagnostic type of application. I think the next one and I won’t there’s a ton here. I’ll only do four today. Just this kind of examples, you know, predictive analytics. I think this is the one that is a long been on the rise for for many loyalty marketers. And as we get more and more data and insight and getting to multivariate type of analysis that comes into place where we can take models and start to predict out future trends and behaviors of our consumers.

So you’re looking at the past historical data. Yeah. But you’re basically predicting it out for future events. So, and, and looking at the consumer life cycle and your customer life cycle and what, what, what are they going to do? Right. And so the easiest one here, I think that we can jump to is, gosh forbid, they aren’t happy with us and they’re starting to attract.

What are some of the indicators that we can have in our model that allows us to see that someone’s about to exit the program? And how do we head that off at the curb, so to speak, in advance? And that’s a kind of a negative example, but it’s a good one in terms of being productive about it and outcome.

And then last but not least is, is, of course, prescriptive analytics where, you know, we’re trying to, to take recommendations for the decision making to achieve a desired outcome. And so you’re looking at the raw data plus all the predictive stuff that you’ve done and the prescriptive analytics.

And you’re trying to say, well, what’s the best course of action for us in terms of how we achieve success in meeting the relevancy for, for Bill or Aaron as, as two of our customers. And so for me, I think in terms of an example on this one here, Is maybe getting down to we’ve taken all the decisions.

The outcome is we want to make sure that our store store planning is fully capacity because we just put a campaign into play and make sure that when Aaron and Bill go into that store, that the shelves are actually full. So, you know, the desired outcome is to sell more stuff through our campaign. And to make sure that the stock in those particular stores is in place, and by moving it around at the variability of what was purchased, we were able to be you know, provide strict recommendations for store on down the street, 2 blocks from Aaron needs to have the capacity because there’s a lot of lookalikes that like Aaron that are going to buy this particular product, but store B that’s near to bill.

He’s, he doesn’t have the same pickup propensity as Aaron. And so we don’t have to have as much stock in that particular store. So you can move things around. So it’s a stretch on that one a little bit to some degree, but you get the idea of it’s about the end goal is to sell more stuff. And so where do you maximize where your inventory is going?

Bill: No, it’s, it’s really good to have that framework. And so when I think about some of the traditional terms that we use, I mean, we, You and I both had a connection point earlier in our careers where we all we talked about was our found recency frequency monetary value and we had these what we called multivariate models that where we would we do a decimal segmentation.

We’d end up with being able to assign a cell number or a score. Probably a value score. Remember, we had a hybrid value score, so it took all three variants. It applied some value. We got a score and then you could clearly see, okay, here’s the quadrant that you should invest heavily in. Here’s the 1 that you should just nurture.

Here’s the 1. You should just maybe ignore a little bit less, you know, a little bit more. So, that was our orientation. It seemed like analytics were those and the segmentation and then we had retention modeling and so. Like, with the RFM, is that in the descriptive analytics bucket? Is that where that lands?

Aaron: Yeah, I think it is more in terms of taking a look at the historical data. So that’s, that’s true. I mean, that’s all analysis on a base segmentation. I think anytime you’re doing a segment that’s Looking at past history to create, like, what are the groups and that look like for whatever reason that you’re doing the segmentation on, like a decimal analysis, you take a variable and you divide it by 10, right?

Like that’s, that’s Pareto. We take the top two and we say, you know, 80, 20, 20 percent of the two decimals at the top where we’re happy with them. And that’s our best customers. Well, I mean that we’ve moved beyond because we know that that’s just not true and not falls out that way. It’s, it’s a good rule, general rule of thumb, but it’s, but it’s, you’re getting now into the, the era where.

The sophistication of the information that you have goes beyond just the transactional data and previous purchase behavior. And you’re getting into the essence of who the individual is, you know, who do they, who do they hang out with now? Who, like, what’s the communities that they’re connected to? Who, who are they influenced by?

You know, when they’re online and someone posts something and then they, and then there’s a response, like these are the data attributes that can start to be collected now. And so it’s become much more rigorous and sophisticated. Let alone the fact that, you know, we’re the what we’re talking about in terms of segmentation models was typically static, you know, set and forget.

Yeah, do you do the segmentation and you leave it for a year or 2? Gosh forbid if it’s 2, a year was a year was even a short term. Now, there are models that allow us for real time data to come together. Where you’re actually doing segmentation in the moment and near to real time. So in theory, a transaction that an individual has in a retail story for some, some brands could move them and bump them in a tier in the moment.

And you could have a, the sophistication to put an offer back to me in that moment, while I’m at the POS swiping through to say, Hey, you just hit the next tier, congratulations, here’s, here’s what’s going on. That’s what we should aspire to. Anyway, not everyone’s there, but. That’s what we should have.

Bill: Oh, I agree.

I agree. And then, you know, I was listening to you talking about diagnostic analytics and finding root cause of actions taken. And I’m curious to know if you think that that you can in a real strong way, determine the root causes without having some of that 0 party data appended to it, like, do you need the qualitative data to sort of confirm what the root cause?

Because you can see that bill usually takes a vacation on sun and sun and sand, but. Maybe you need to ask Bill and make sure, you know, is it, is it because I have young children, but I’ve always desired to go to the mountains. And so as soon as they get to be a certain age, I’m, I’m always going to be in the mountains.

I mean, I don’t know. So I’m always a

Aaron: big fan of if you want to know something, ask, like, if you want to know what Aaron, if you want to know if Aaron’s loyal to you, ask Aaron, like, and he’ll tell you, right. You know, there’s, and there’s some bias when you ask your, your customers to some degree in terms of what they’re on and redemption.

Behavior is one of those anomalies where it’s, it’s so emotive as well as rational at the same time, like, and in the moment, it can flip around. So, I think that’s 1 of the reasons why, you know, we haven’t dwelled into that type of analytics as much as an industry as I think we could, but, but to your point, I think you can have a model and it can pump out a suggestion and if there’s ever an element of, well, we’re not really 100 percent sure.

Then I think it’s, it’s run it by the particular customer because it’s easy to send you a note to say, Hey, Bill, we noticed in your last 4 vacations that you’re focusing on sun and surf. Here’s a couple that we think that are right in your wheelhouse, you know, holiday holiday B, but, oh, but just in case we’re off here’s holiday.

See, that’s more of a walking tour of, of you know, Milan or something like that. Because also like fast, given the, I’m going to go

Bill: I’m going to go all Tony Kornheiser on you, pardon the interruption. And you know, we have our part in the education series on the Wise Marketer. Here’s 2 lightning round 1 minute answers.

If you can do it to that, but we went to a conference lately, and it was populated mostly by retailers. And it seems like they were all living in the predictive analytics space because they’re trying to figure out what’s the next purchase going to be? What’s the adjacency of 1 to the other? Do you agree with that?

Is that where predictive comes into play for retailers in particular?

Aaron: Yeah, you’re trying to like predict when the individual come back into the store next or when they are induced them to come into the store next is actually probably a better way to describe it because you know that the, the, the, they’re going to be out of a certain particular product by a certain timeframe.

And so it’s, Hey, let’s push them back in and then come maybe pick up more, maybe bulk, bulk load, who knows, depending on the type of model. But yeah, that would be a fairly, I simple way to do that. Yep.

Bill: Well, you did that under a minute. You’re getting brownie points here. So here’s the last 1. I want to ask you about in perspective in prescriptive analytics.

Yeah, it to me is such an interesting area because it highlights the value that loyalty programs bring to the enterprise beyond just. You know, increase of frequency and purchase and things like that. We’ve been seeing large retailers, and, you know, we just did an interview with a CEO at 7 Eleven. They’re talking about how they’re using analytics from loyalty programs to change their range and assortment within the stores.

They’re customizing the inventory in the stores to meet the demands of people in particular neighborhoods. Yeah. And they were kind of touting that, you know, our stores are the hub of these different communities. Like, wow, really? From loyalty analytics? So. You have any thoughts on that?

Aaron: Yeah, I think the key here is that We’re looking at actual advice that you can give, like what are the next steps or the best steps that you can take to achieve the outcome that you’re looking for.

That’s the way I think of prescriptive analytics is always like, you know, you, you got to the decision making and, and, and here’s our advice of what you should do. And so, to meet, to meet a, a goal, either with the consumer or with the business. And so the example that you just laid there where you’re literally starting to change the merchandising structure of where you’re putting stuff as a result of your customer insights and where your customers will have a propensity to pick up, like you’re, that’s your, your loyalty analytical tools telling you, Hey, we need to change where the flow of our goods is going to help make sure that we, when we run a campaign, that there’s not a bad experience because there’s a, there’s an out of stock at a store in an area where we’re inducing a bunch of behavior for people to go pick up something.

Cause it’s a two for one.

Bill: As an example. That’s great. That’s great. Okay. Well, that’s a good way. You’re you’re just, you must have plenty of caffeine today because you know, that is really good. So, yeah, well done. Perfect. So, yeah, that’s that’s a good place to wrap it. We’re going to come back next week with some examples.

I think market examples. I’ll put in parentheses, if we can find some really good ones where we are sure that we’re doing a great job. Because we’re having a little trouble finding the retailers that are doing an absolute, you know, just bang up job on, on showing the consumer that they’re using the data, but we’ve got a few for you.

So, for now, at least, as always, anyone interested in joining this community of loyalty marketing professionals at the loyalty academy dot org we’ll be able to get all the information about the seal and P designation about the community of people designation, which now numbers very close to 1000.

Stand by for a big announcement. They’re in about 54 countries around the world. But if you want to, if you want to amp up your learning, help your career out. Add your credentials. Brag a little bit then check out loyaltyacademy. org and and see what’s available there and If you want to see any of our previous podcasts go to wisemarketer.

com Obviously go to letstalkloyalty. com and Aaron, you have a you have a thought final thought you want to share? Okay, so, I don’t know with that stay with us We’ve got one more episode in this series and we’ll wrap up our discussion of loyalty analytics and we’ll see you back here next week Thanks so much

Paula: This show is sponsored by Wise Marketer Group, publisher of the Wise Marketer, the premier digital customer loyalty marketing resource for industry relevant news, insights, and research. Wise Marketer Group also offers loyalty education and training globally through its Loyalty Academy, which has certified nearly 900 marketers and executives in 49 countries as certified loyalty marketing professionals.

For global coverage of customer engagement and loyalty, check out thewisemarketer.com and become a wiser marketer or subscriber. Learn more about global loyalty education for individuals or corporate training programs at loyaltyacademy.org.

Thank you so much for listening to this episode of Let’s Talk Loyalty. If you’d like us to send you the latest shows each week, simply sign up for the Let’s Talk Loyalty newsletter on letstalkloyalty.com and we’ll send our best episodes straight to your inbox. And don’t forget that you can follow Let’s Talk Loyalty on any of your favorite podcast platforms. And of course, we’d love for you to share your feedback and reviews. Thanks again for supporting the show.