#147. Using RFM Segmentation to Personalize the Customer Journey | Podcast

Vira: 0:16

Hello everyone and welcome back to Email Einstein, the podcast by Flowium. I’m your host, Vira Sadlak, and I’m a marketer at an agency called Flowium. And here at Flowium, we really love what we do. That’s why we want to share our insights with you. And guys, you are listening to episode number 2 of our Segmentation May. Not segmentation, but segmentation. And I love this smart little wordplay. But basically, the entire month of May, we will be talking about all things email marketing. Marketing segmentation, all things segments. We’re going to have a few really exciting guests. For example, last time we talked about the importance of market segmentation, and today we’re going to go in depth of using RFM segmentation to personalize your customer journey. And for those of you who don’t know, RFM is recency, frequency, and monetary value. So these are the 3 main segmentation strategies that we will be discussing today with our guest, Andriy Boychuk from Flowium. Andriy is our CEO. You’ve met him a lot of times in this podcast, but I’m really excited about today’s episode because, I mean, it’s a basis, right? The basis of email marketing segmentation. And we’ll be discussing some of the foundational principles. If you will use them, believe me, you will be able to improve your customer journey and personalize your email marketing strategy a lot. So let’s do this. So Andriy, tell us a bit about this RFM segmentation. First of all, what does it mean and what does it stand for? RFM.

Andriy: 2:17

So RFM segmentation method. RFM stands for recency, frequency, and monetary value. So this method was created in the 1970s by direct mail marketers. So when they were sending physical mail to people’s door to market, to promote either their service or product. And as you can Can understand right now we are sending emails and they cost us close to zero or like just a few cents to send out the emails. However, when you send direct mail, it’s very expensive to send. It might cost $1, $2, $3,$5 per direct mail. So when you send to thousands, the amount adds up pretty quickly.

Vira: 3:07

Right.

Andriy: 3:08

So the segmentation method was created to target people better and also to save money on, on the actual mailers.

Vira: 3:17

That’s pretty cool. That’s pretty cool. I didn’t realize that it started like back in the ’70s. And honestly, if those like email, emails costed more, I feel like people would be more, more smart about the way they segment their list because that’s a popular rookie mistake. When someone have a list, when they have like 1,000 people on their list, they’re like, why don’t we bombard them all with the same message? The more people, the more money, the more fun. No, it doesn’t work this way. Unfortunately, we’ve tested it like gazillion times. The better segmented your list is, the better segmented your campaign is, the better revenue you get. And that’s— I still can’t like explain it per se, but it works every freaking time.

Andriy: 4:01

And there’s a bunch of ways how you can segment your customers, your audience, and actually your imagination is the limit. But this is one of the most popular ways to segment your list.

Vira: 4:16

It’s the easiest one, the most logical one. And we kind of discussed it before in our podcast. We talked about it in episode number 29. So go back, you guys, and listen to that episode. But let’s talk like in depth about those like quantifiable factors. Recency. How can we build segments based on recency and like what does it mean in the world of e-commerce email marketing?

Andriy: 4:45

So how often do they buy? So let’s say we have 2 customers and one customer spent$2,000 with you and another customer spent$1,000 with you. So we are based only on that information, which customer is more valuable to you,$2,000 and$1,000?

Vira: 5:09

Well, 2 grand, obviously.

Andriy: 5:12

Of course. However, if we add another criteria, recency, for example, the customer who spent$2,000 last time they purchased was 2 years ago, and customer who spent$1,000, they bought 4 times during the year, 1 year. So if, if we add those criteria, which customer do you believe is more valuable?

Vira: 5:37

Yeah, it’s getting more interesting with the$1,000 guy because in most cases, the more recently a customer has interacted or like transacted with the brand, the more likely the customer will be like responsive to our email marketing efforts, right? It’s like, this is, this changes everything pretty much.

Andriy: 5:57

And as we know that it’s extremely hard to convert one-time buyer to two-time buyer and so forth. And in this example, Even the person spent$2,000. So this is lifetime value of the customer. In the second example where we have 4 different purchases, each purchase like totaling$1,000, we can expect they will make other purchases in the future because it’s easier to convert them for 5th purchase, 6th purchase, and 7th purchase.

Vira: 6:28

Right. So recency is more about like how much time has elapsed since the customer like last visited your website or last had some sort of like activity. or transaction with your brand, but frequency is more about like how many times they purchased, correct?

Andriy: 6:45

Correct. Yes.

Vira: 6:46

Okay, Andriy. So we kind of covered recency, basically how much time has elapsed since the customer placed an order. We kind of covered frequency, meaning like how often has a customer transacted or like interacted with your brand during a particular time or year or whatever. And then what is monetary? Like, what is monetary factor?

Andriy: 7:13

It’s how much money customers spend with your brand. The easiest way, like, first of all, you need to know what is your average order value is. And in the last 2 months, we produce a lot of content on YouTube as well in our blog posts about AOVs, defining AOVs, how to improve your AOVs. But basically, your average order value is like your middle point and anybody who below that or above that, it’s—

Vira: 7:45

It’s a different group of customers basically.

Andriy: 7:47

Yes.

Vira: 7:48

Yes.

Andriy: 7:48

Which you can segment.

Vira: 7:49

You can segment and you can like treat them differently in a way. We’re going to definitely talk about it and then how important it is to like segment those people out. But like generally speaking, why should we care about like recency, frequency, and like monetary factors?

Andriy: 8:07

Like how So main benefits of any segments, not only this segment, but segments in general, is first of all to increase sales revenue, improve marketing, and increase customer retention, also known as decreasing churn rate. If we are able to communicate to the customer the right way on the, on the specific step of their journey, they are more likely to take action, which we want them to take, either buy from us, engage with us, leave us a review or whatever we’re asking for. So that those are like 3 main benefits of segmentation.

Vira: 8:55

Yeah, no, definitely. And again, going back to that example, I kid you not, we’ve tested it like so many times when we were sending the content to like pretty much like a master segment versus the segmented campaign. So the segmented campaign, very similar content delivered like 4 times the revenue. And that’s like the example that I showed to my colleagues and they all were like shocked with like how the segmentation works. Like, again, I don’t necessarily understand the, like the mechanics behind it because like logically, like the more people you email, the more people your email will reach, the more revenue it should get. But I assume that it will affect your deliverability as well, meaning that yes, maybe this time you will deliver and a lot of people will see your email, but in the future, your deliverability will be hurt and your email will end up in a spam folder or a promotion folder and no one will see it at all. So Andriy, how do I go about performing the analysis of that RFM? Is it available to everyone or do you need some special tools to perform that RFM analysis?

Andriy: 10:09

There’s a bunch of videos on YouTube which teach you how to do it on in Excel. There’s formulas. I think anybody can do like in under 10 minutes. However, you can do it only once, so it’s not dynamic. So I personally do not recommend. I believe in something what, what is dynamic. So if you use your email marketing service provider, Nowadays, almost all of them can do it. Since we exclusively work with Klaviyo, I’ll be giving examples from Klaviyo. There’s 2 ways how to do it. So first way, we can create a bunch of segments and combine them.

Vira: 10:52

Mm-hmm.

Andriy: 10:53

Next time you send in campaign, you can include and exclude some of segments. Or second way, you can create automation in Klaviyo or any other platform and which will be applying scores to each customer. And what do I mean by score? So let’s say, let’s say, let’s talk about recency. So let’s say, and we have a score from 1 to 5 and we have to apply a score either 1, 2, 3, 4, or 5 to all of our customers. So let’s say if somebody purchased in the last 30 days, we’ll give them score 5. This is the highest score, is the best score because they the most recent. Somebody who placed an order a year or more from now, we’ll give them score of 1. So everybody will have, everybody will have score either 1, 2, 3, 4, or 5. So one way to do it, as I said, you can create 5 segments. This person placed order at least once between day 0 and 29 days.

Vira: 12:00

Mm-hmm.

Andriy: 12:01

Or placed order one time, one time, at least once, 365 days, like, ago. So it’s like a year ago. So you will have those 5 segments. So this is the first set of data. The next set of data is defining your monetary value, and each business is different. So that’s why I said it’s very important to define your AOV, average order value, and your average order value will have number 3, score 3, because it’s in the middle. So average in the middle and anything which below that 2 and 1 will have lower monetary value and anything above will have higher monetary value. Same thing. We’re creating 5 different segments with those definition, how much money they spend, what their revenue. The last thing is to create 5 segments for frequencies.

Vira: 13:04

Mm-hmm.

Andriy: 13:05

So again, you need to define what is the average number of orders. Let’s say it’s 3 just for easy math. So 3. So if somebody plays, let’s say between 3 to 5, it’s a score 3. Anybody who plays between 2 and 3, it’s a score 2. And anybody who plays only 1 order is a score 1. So now When we have those segments, like, I know it’s a little bit a lot, like 15 different segments. You can reduce it if you want. You can do from 1 to 3, like to have less segments, but 5 would be better. Like it will be better segmented, that list.

Vira: 13:48

So basically all of this like segments that you assign them like the number or like the grade, they are all dynamic, meaning that people are like moving from one group to another depending on their behavior, right?

Andriy: 14:02

Yes.

Vira: 14:02

Okay, that’s cool.

Andriy: 14:04

So, so this is step number 1. Step number 2, let’s say we want to target somebody, somebody who has the lowest frequency, which is score number 1, and somebody who placed order 1 year ago or like even further in the time than who, let’s say frequency was score number 3. So on average they placed average number of orders, which is score number 3. However, they had the highest score for monetary value. So it’s the score they will have 1, 3, 5. So the lowest frequency, the highest monetary value, and in the middle frequency. So what can we send them?

Vira: 14:53

Yeah, that’s an interesting— well, I don’t know. I would think that to get them back, we would need to give them some sort of like discount, maybe like additional discount. Because like in my world, it’s like easier to convert those people who have higher frequency and bigger like recency or not much time elapsed since the last time they purchased something from us. So to those people, I usually don’t give the discount. But for people who shopped with us like long time ago, first of all, we need to remind them of who we are because it’s been like a lot of time has passed. So we probably need to remind them about who we are, why should they care about us, and then maybe try to win them back with some sort of like discount. That would be what I would do. Andriy, what are your thoughts?

Andriy: 16:07

I like what you said. If I, if we’re limiting, limited only to email marketing, I agree 100% with you and I’ll do the same thing. However, remember, like email marketing is not one, only one channel which you can leverage. If those people are like, have a score 5 for monetary value, I’m not sure exactly how much money they bring, but they are your, let’s say, top 1%. Maybe it’s worth your time to pick up a phone and call them. Or maybe you want to invest$5 in their service handwritten or Ignite. We interviewed the guy.

Vira: 16:46

Yeah.

Andriy: 16:47

Yeah. Ignite Post. Maybe you want to spend $5 and send them handwritten postcard to win them back because they already— you are extremely profitable on this, on this client already, it would be cool and beneficial for you to win them back. So this is just one, one example. Let’s talk about 555, somebody who has highest recency. So they just purchased in the last 30 days. They have a frequency is the biggest. So they place the most orders in this short period of time.

Vira: 17:22

Avid, avid fans, right?

Andriy: 17:23

Yeah. Number 5, the monthly value is the highest, that client already is doing everything that you want to do. So like discount unnecessary. So but maybe you can add value somehow, somehow else, maybe automatic emails from the founder, like, I don’t know, like, or maybe you can set up the notification to the founder, say, hey, this person has an RFM score 555, maybe you want to write them personal email from yourself or maybe add loyalty points.

Vira: 17:59

Actually, yeah, actually, remember we were discussing that on one of the conferences, there was this brand called Ministry of Supply. I remember like they are like the apparel brand or something and they have this thing where the best performing customers are like the customers who spend the most money with them, the customers who have the most interaction with them, they get to talk to like a CEO. of the company, they get to have like an interview with him or her. And sometimes they even like invite them to their company retreats, which is like super cool because you get to meet your customer, you get to talk to them, you get to ask all of the important questions about the brand. You can hear their opinion and you can honestly like thank them for being like so loyal to your brand. And when I think of this customers, like what we do internally with our email campaigns, sometimes we try to like utilize those people get reviews and user-generated content from them because they are your most loyal fans. They are the people who love you, who want to talk to you and who would be happy to talk about you. So that’s like how we utilize this group of customers, but I like how you sort of like segmented them in this like easy to understand, almost like formula. It’s super cool what you’ve done. Yeah.

Andriy: 19:19

And for those kinds of fans, maybe there’s also, it’s You can create something like we created for Flowium Podcast socks, something that you don’t sell, but you can exclusively send it to those customers. They will appreciate it. They already love your brand. They are true evangelists or how it’s called in marketing.

Vira: 19:41

Yeah.

Andriy: 19:42

They are true evangelists, ambassadors of your brand. So there’s no more convincing needed. You don’t need to provide them crazy discounts, but something special or like create some sticker and send it to them. Like people will appreciate that a lot.

Vira: 19:57

And you don’t want to train your customers to expect the discount codes from you all the time, no matter like what your discount strategy is. People who buy from you without discount code, like don’t train them to expect something from you all the time. It can be something else. It can be the freebie. It can be maybe like the bundle offer or like exclusive, uh, early bird access to your new launch or something like that. It doesn’t have to be the discount Discount, discount. Yeah.

Andriy: 20:28

And we just said what, what is possible in terms of campaigns, but also those kind of segments can trigger a flow as well. Automation.

Vira: 20:39

Like, give me an example.

Andriy: 20:42

You put me on the spot.

Vira: 20:44

Yeah, I do. Give me an example of, okay, say I want to, I want to target my like biggest spenders. I want to target people who are Who are like my biggest, biggest fans? How would I like define those people? What would be the trigger for the flow?

Andriy: 21:54

So there’s a bunch of ways how to— how you can do it. And probably Klaviyo professionals will judge me what I say now. But I mean, one way how from top of my hand you can trigger by segment if they enter this monetary value 5, so the highest monetary value and the rest can be filtered how many orders they made. So if let’s say the highest number of order, like score 5, they will go further. And the last thing is recency, if they bought something in the last 30 days. So that flow will be triggered and filtered, make sure the other conditions met, and they will receive the email.

Vira: 22:39

Interesting.

Andriy: 22:40

Maybe, maybe what we said before, the automatic email from the founder, text-based email.

Vira: 22:46

Yeah, that’s a really nice personal touch.

Andriy: 22:49

Or review request.

Vira: 22:51

That’s a good one too. That’s a good one too.

Andriy: 22:53

Honestly, like if, if like if you want to invest a little bit more time, I would personally create backend flow automation which will assign those scores to customer profiles instead of creating segments.

Vira: 23:12

That’s actually what I was thinking. Because I don’t know if I’m if that issue still exists, but basically when you trigger the flow based on segment, it doesn’t happen right away. As far as like Klaviyo Pro like explained to me, like sometimes needs to pass and sometimes it takes time. But sometimes for that like post purchase sequence or something, you want that email to be like delivered right away. So yeah, that’s that’s interesting that you’ve mentioned it actually.

Andriy: 23:42

I mean, if you have a custom, let’s say you have a custom profile properties called RFM, like 3 fields and you’re assigning scores dynamically, you can go crazy in your abandoned cart flow, winback flow because you are able—

Vira: 24:01

Yeah, you can segment out those.

Andriy: 24:03

Yeah, you are able to create advanced journeys, but also it’s It’s hard to manage. So I don’t want— I do not recommend for you to go crazy and create, let’s say, 5 journeys per each flow, but know that it’s possible. Maybe focus only on those.

Vira: 24:22

Sometimes it doesn’t— sometimes it doesn’t worth it. You know, we’ve tested like different approaches and some— sometimes the simplest one works really well. So like, why bother? So it depends, of course, on like on your overall strategy, but like, what are your thoughts? And again, it’s a bit off topic, but Some brands are segmenting out people for Black Friday, Cyber Monday based on their, based on like how much money they spent, like how many like discount codes they used in the past and like stuff like that. And they give them like different discount codes. For example, they give like less of a discount code to VIP customers compared to like those like win-back people, you know? What are your thoughts on that? Are you pro that approach or do, or are you against that approach? Because like a lot of brands prefer to be consistent and just like offer the same discount throughout the Black Friday, Cyber Monday, other, or like dividing them into different groups? What are your thoughts?

Andriy: 25:17

Unfortunately, I cannot answer this question without knowing more information. It depends what their goal is. Let’s say if they don’t care much about profit, they just want to sell as much stuff as they want, they can create the biggest discount and just offer to everybody. But if they want to be more profitable, maybe 10% will be enough for loyal customers, return customers, but 20% for first-time purchasers or somebody who never purchased just to convert them. It’s hard to answer this question because each brand have different goal. Some brands, they just even if they break even on the sale, they are happy because they’re just trying to move inventory during that season.

Vira: 26:06

That’s true. Yeah, there is definitely no one-size-fits-all when it comes to email marketing or to any other marketing, to be honest. But like, if you would summarize, what are the important segments to target for Black Friday, Cyber Monday based on this RFM model? Like, what are those segments? Like, how would you sort of like structure the email marketing, say, for the Socks brand? Let’s go back to good old Socks example. What groups of customers would you target?

Andriy: 26:39

Like for me, anything that where you go beyond number 3, it’s too complicated for me. So I’ll probably create 3 scenarios, 3 groups of customers. So maybe the best customers and instead of giving them discount, maybe offer some something extra. Like one of our clients, Roosevelt, each Black Friday, Cyber Monday, they create the wrapping paper.

Vira: 27:07

Yeah.

Andriy: 27:07

But this is unique. Custom, right?

Vira: 27:09

Yeah.

Andriy: 27:10

Yeah. And they send it as a gift with any, with any purchase. So stuff like that. So this is your first group of customer. Second group of customers, somebody like very generic. They never purchase from you or they have a very low score on RFM. model, let’s say 1 or 2 in each categories. For those, you can just do generic sale, whatever the highest discount you have, 10%, 20%, 25%, and bombard them as much. I mean, during this holiday season. The middle part.

Vira: 27:50

The time when it’s appropriate.

Andriy: 27:52

Yeah. The middle part, anybody between, anybody like with a score 3 and 4, I don’t know, Vira, what would you suggest?

Vira: 28:01

Like for me, those people, they would probably be the same strategy as for the holiday shoppers or like those rare customers. I would just like send them all like generic, generic discount, but maybe I would be a bit more aggressive with reminders. You know, if I see that customers engage with my emails, maybe I would send them like more reminders. And this year we are trying to incorporate some like fun relevant information in those reminders. For example, instead of just sending a reminder saying like, hey, Black Friday, Cyber Monday will end in like 24 hours. We’re going to send reminder plus gift guide saying like, hey, this is the product that you can purchase for your mom from our website. This is what you can purchase for your dog and this is for your dad or stuff like that. Or like 5 ways to use this product in 2022 or 5 makeup trends or whatever, something like that. Just because I know that they will be bombarded with like a lot of just like generic reminders and we want to provide some sort of value as well, but only to those who engage with our emails. Obviously, we don’t want to send this type of email to a person who doesn’t care about us at all. So that’s like the biggest difference I would say would be for my clients anyways. Yeah.

Andriy: 29:16

Perfect.

Vira: 29:17

And okay, Andriy, I mean, that sounds pretty cool and pretty impressive, but what are the disadvantages of this model? If there are any, because I have at least one or two that comes to my mind right away.

Andriy: 29:35

So you start and if I have any, I’ll add.

Vira: 29:41

It’s not even like a disadvantage because I do believe this approach is really powerful. But to me, it looks like it’s based on a historical method. It looks like it bases more on a past customer’s behavior, it may or may not help you accurately indicate the future activities.

Andriy: 30:02

I agree. I mean, nobody knows the future. So the best way to predict the future is—

Vira: 30:07

Klaviyo kind of does. Klaviyo kind of does. They do have this.

Andriy: 30:13

I 100% agree with you. Klaviyo does. But how does it do it? Based on the past activities. How does Netflix know How does Netflix know what movie or series to recommend based on the past activities? How is your Facebook feed customized based on the things you like?

Vira: 30:35

And yeah, kind of. Yeah, I agree. And like the other thing, and you might or might not agree with me, that this model, it kind of doesn’t take into account like other variables that are also like important, like specific product purchased. Say like this person, they’re coming to our store and they always purchase like this particular product. They express the interest in this product. So this model does not take that into account, right? So that’s—

Andriy: 31:08

Yes, I agree. I don’t think any segmentation model is perfect. It’s, there’s pros and cons for each model. This is This is more generic. If you want to be category specific or product specific, of course it’s not the right model to use. However, you can use the combination of models. Maybe you do RFM model 555 for— again, depends how many subscribers you have. Right now we’re prospecting client and they have 2 million subscribers and I’m not sure how many clients, but But with that kind of data, you can layer different segmentation methods. You can layer the RFM model plus what product they purchased, what product they viewed, from which category, what color it was. You can like layer like 10 different segmentations and you’ll probably end up with 5,000 potential customers and you can target them.

Vira: 32:08

Yeah. Yeah, that’s true. It’s like not fun to be working with those like smaller accounts because you don’t get to segment as much just because like when you segment out the majority of people based on their interests, like who do you send it to? Sometimes it’s just like not worth it when your account is small. And actually there’s like statistic that something like big accounts that have like over 200,000 subscribers, they on average have 133 segments. I remember, I don’t know why I remembered this number, but it was like so impressive to be 133 segments. Now, when I think about my clients, we’re getting there actually. It’s bizarre. Like the bigger client you have, the more creative you can be. So you guys grow faster. It’s so much fun up there. I’m telling you.

Andriy: 32:56

But if you make a mistake, mistake also more painful, like to send one email to 2 million.

Vira: 33:01

That’s true.

Andriy: 33:02

Versus like 10,000.

Vira: 33:05

Yeah. Yeah. They can cost you a lot. But those apology emails are fun to make though. And I know that some brands are actually utilizing those apology emails on purpose just because they create a lot of excitement.

Andriy: 33:20

Sending out apology email for a mistake. I didn’t make it yet.

Vira: 33:24

No, they make mistake on purpose and then follow up with an apology email right away. Yeah. I mean, I read on some blog, I don’t know if it’s true or not, or do people really make it on purpose? But yeah.

Andriy: 33:37

I like those emails from Wine Awesomeness. I still follow them. We had Dale and Logan on the podcast in the first episode. They are 2 founders and I like how they exchange emails and they send out their exchange as a marketing email. So like, hey Logan, and they’ll like respond, Logan, I said not to send those, that discount to the customers. I mean, it’s like fun and cool and they always send as a text base and it’s actually email thread. So you don’t even see the unsubscribe because it’s like very long email.

Vira: 34:12

Oh, that’s so funny. Yeah, I don’t remember them using that example. It’s so funny. I should try it. I should try it with one of my clients.

Andriy: 34:20

Yeah, but like if you subscribe, they send crazy number of emails, I believe 5 per day.

Vira: 34:28

Oh yeah?

Andriy: 34:29

Yeah, until you buy. If you buy, they slow down.

Vira: 34:33

Okay. You’re on your own, pal. Now you’re on your own.

Andriy: 34:37

Yeah.

Vira: 34:38

Oh, that’s so cool. Well, thank you so much, Andriy. That was a very educational episode for sure. Thank you. Thank you for listening, you guys. I hope you enjoyed the podcast. I really enjoyed recording this little episode, especially about the segmentation, because I know that this is the topic that you guys love the most. That’s why we’ve prepared a few more really exciting episodes over the next few weeks. So stay tuned. We will be talking to some Klaviyo representatives. We will be talking to fellow ecommerce, email marketing specialists, and a lot of exciting stuff are in the works. In the meantime, if you like what we do, please leave us a review. This is very important for us so we can reach more customers like you. If you like what we do, please leave us a review, send us a screenshot, and we will send you your very own personalized Flowium something, Flowium surprise. Thank you so much for listening, and we hope to have you back here next Tuesday.