#80 A chat about $PodcastSubject with $GuestName (with Rasmus Houlind @ Agillic)
This week Dan and Dara are joined by Rasmus Houlind – author of the book Hello $FirstName and CXO at Agillic. In this discussion, they chat about what personalisation is, the differences between implicit and explicit personalisation, how the customer journey belongs to the customer, and how to adapt a CRO mindset for personalisation.
The kind folks at Agillic are giving away 10 copies of Rasmus’ book Hello $FirstName when you sign up on Agillic’s website here.
You can buy Rasmus’ book Hello $FirstName: Profiting from Personalization over on Amazon here (UK link, but you can find it internationally too).
Check out the Martech Tribe for independent tech research, benchmarks and selections here.
Enrolment is now open for June’s cohort of the GA4 Immersion 6-week cohort training with early bird pricing for 25% off!
In other news, Rasmus gets cold swimming!
Follow Measurelab on LinkedIn and on Twitter.
Intro music composed by Confidential – check out their lo-fi beats on Spotify.
If you’re like what we’re doing here, please show some support and leave a rating on Apple, Spotify, or wherever really.
Let us know what you think and fill out the Feedback Form, or email podcast@measurelab.co.uk to drop Dan and Dara a message directly.
Quotes of the episode from Rasmus:
“…personalisation is a very, very broad umbrella term for all these sort of strategies and tactics within making a personalised customer experience regardless of the channel. That also means that the definition of personalisation, the one definition to sort of rule them all is so vague and so broad that you can hardly use it for anything.”
“…if we manage to not insult anyone along the way, and not come across as creepy then I think we’re doing good job…”
Transcript
The full transcript is below, or you can view the Google Doc.
Intro
[00:00:15] Dara: On today’s episode, Dan and I are joined by Rasmus Houlind, who talks about his new book Hello $FirstName, which is all about profiting from personalisation. So we have a really good chat about personalisation, some of the dos and don’ts and what organisations need to know about if they want to go on the personalization journey.
[00:00:32] Daniel: We’ll link off to his book and where you can find that in retailers and including Amazon and what not. Also, stick through to the end and you’ll find out how you can be in a chance to win one of the 10 books that he is giving away. So check that out, listen to the end, and check out the show notes for a link to get in that draw. More details in the show notes of course, check in whatever app you’re listening to. One last thing to mention, this is the end of our little mini-series. We generally have a little break after 10 episodes, so we, me and Dara and others will be back in a couple of weeks on episode number 81. Can you believe it, Dara, going into the eighties. Anyway, we’ll see you in a couple of weeks and I hope you enjoy this one.
[00:01:06] Dara: Hello and welcome back to The Measure Pod, a podcast for analytics and data enthusiasts. I’m Dara, I’m CEO at Measurelab.
[00:01:13] Daniel: And I’m Dan, I’m a consultant and trainer at Measurelab.
[00:01:15] Dara: And we’re also joined on today’s show by Rasmus Houlind, who is a published author of multiple books, also a keynote speaker, and is also CXO at a company called Agilic. So Rasmus, firstly, welcome to The Measure Pod, thanks for agreeing to come and talk to Dan and I.
[00:01:33] Rasmus: Thank you very much, pleasure to be here.
[00:01:35] Dara: Great to have you. So what we always do is rather than doing a really bad job of introducing people and then making them very annoyed with us, we get them to introduce themselves. So could you tell our listeners in as much or as little detail as you like really just about your journey? So what got you to where you are today? A bit of a whistle stop tour maybe?
[00:01:53] Rasmus: Well, yeah, I’ll see if I can keep it short. So currently I’m, just recently published a book about the topic of personalisation, which is really what we’re here to discuss. It’s called Hello $FirstName and before that I’ve been, my career actually started with digital agencies working with building websites and systems integrations and loyalty programs and marketing automation and then I did my first book in 2015, which was about omnichannel marketing. I did another one back in 2019 but before that I joined the marketing company, software company, Agilic, where I’ve been for seven years now. Currently in the role of 100% evangelists, so basically I get to do what I like best, like deep dive into subject matter, about marketing and marketing automation, personalisation, all those kinds of topics. How do you make money from this. So I’ve been in countless discussions about value creation and analytics and AI and personalisation over the last almost 20 years. That’s probably where I am in my story to keep it short.
Topic
[00:02:55] Dara: Brilliant, and I think you’re right when you said that, you know, the main focus of our chat might be on the current book, but actually I think we could end up venturing down a few side tangents maybe as well. I think the previous book and the mention of AI could be something that we asked you about. I was just going to say firstly, I think the name of the book is great by the way. It’s catchy and it’s clever and it made me think of all the times that I actually genuinely received an email saying, hello first name. I mean it might seem obvious to some of our listeners, but maybe just expand on that a little bit. What was the motivation behind the name of the book?
[00:03:24] Rasmus: So the name of the book, Hello $FirstName which is sort of mimicking the syntex that you use for parameterized personalization if you’re doing email marketing, for instance. So most of have received an email where it says hello and then whatever first name or somewhere where unintentionally it actually did say, hello, first name, or hi, first name or dear first name and it just feels so dumb, right? And we’ve, we’ve all seen this, or even worse when the syntax goes wrong. And if it’s like, hello segment name. And it also brings up, I think when I’m speaking to people who have been working with direct marketing, I think one of the reason you see them frown or smile or like somewhere in between there because I think this brings up a lot of shame. Basically shame is a really strong thing, a strong motivator because people have tried themselves sending something out where they didn’t get the syntex right? So like, oh yeah, it’s this eerie feeling, a bit creeped out about past failures.
[00:04:19] Dara: I’m going to ask you another, this is probably going to be a very annoyingly open question but it kind of mentions in the description of the book that obviously the personalization is about so much more than that, it’s about so much more than just getting somebody’s name right, in an email campaign. But how would you define if somebody had never heard of it before, where does it start and where does it end I guess, because it can be such an expansive, it can really touch every marketing channel and every touchpoint a customer might have with a brand, could involve some element of personalisation. So what would be your succinct definition of what personalization is in a kind of business context?
[00:04:52] Rasmus: So first of all, before the definition, I believe you’re absolutely right. Personalization can and does apply to many different sort of branches of marketing. That’s both the beauty of it, because it’s like very versatile and it’s, I mean, everything that goes into the customer experience, which is essentially any conceivable touchpoint can be personalised. Not saying that it should be, and thus personalization creeps in everywhere. But because it does so, because it does creep into different ways of working with marketing. That also means that the marketers within different disciplines will tend to look at each other and be under the perception that the other ones are getting personalization wrong or they’re working with the correct type of personalization. And they’ll be misunderstanding each other using the same term, but meaning slightly different things using different terms, but meaning the same thing. And business models with personalization vary greatly between marketing automation, between campaigns and performance market and product recommendations and so on.
[00:05:52] Rasmus: And so it’s really hard to argue that the other marketers, the ones not belonging to my particular tribe, whatever that be, that they got it wrong because they are indeed getting paid for doing this. They’re getting their paycheck, they’re sending invoices where it says personalization, consulting or whatever. And the paycheck never lies, right? So if you’re getting your money from working with personalization, I mean, why would you care that someone is sort of claiming that you got the wrong end of personalization, it doesn’t make sense. So I’m under the impression that personalization is a very, very broad umbrella term for all these sort of strategies and tactics within making a personalised customer experience regardless of the channel. That also means that the definition of personalization, the one definition to sort of rule them all is so vague and so broad that you can hardly use it for anything. But it is about changing or hiding or showing content to customers or prospects or users or recipients, where you have an idea that this will be closer to what they’d be expecting or wanting to engage with or what they would like to see based on explicitly stated preferences from them or inherent preferences from coming from analytics or simple data points or whatever information or data that you have on these customers, regardless of the channel, regardless of the discipline, regardless of where this is taking place.
[00:07:16] Daniel: You mentioned there Rasmus around all of these touchpoints could be personalised, but whether they should be or not, that’s a separate question. That’s interesting to me, so what are those kind of situations that you wouldn’t want to personalize it, even if it’s technically possible and other people might be doing it? Why would you make the call not to?
[00:07:31] Rasmus: Good question. So I envision the customer journey basically as a ribbon. If you are buying a present for someone, you have a ribbon and you tie around, you make a beautiful bow on top of the, of the gift. Imagine this ribbon being the customer journey like from when the customer first hears about you and your brand until they don’t want hear about you again. So from start to end customer journey and imagine that there are sort of, and you want to get them from A to C right. And imagine that the customer journey can unintentionally break at a certain point in the customer journey, and obviously you don’t want that to happen. So for instance, you come to the checkout page of a website on an ecommerce site and they don’t have the, the right payment provider, for instance, the ones that is particularly useful for your territory or geography or whatever.
[00:08:16] Rasmus: So that tends to be a very sort of, geography or location based thing, payment providers. So in these cases, you don’t necessarily want to personalise which payment providers, you actually localise obviously, but basically want to just remove friction because this is not, this not an interesting place. This is not a place where people make up their minds or where they’re particularly in doubt or whatever. This is about removing friction from whatever the customer is trying to achieve. So I wouldn’t personalise necessarily, I’d start other places where there was a bit more at stake, where the customer, I mean, I defined these places in the customer journey as moments of truth. So the places in the customer journey where either the customer is trying to make up his or her mind about something that is related to your brand or they should be making up their mind about something that’s related to your brand.
[00:09:03] Rasmus: And these particular moments of truth in the customer journey, let me give you a few examples. So say that you, your health club, PureGym for instance, right? What are the most important points in the customer journey for a health club operator. So that would be in the beginning, so in the onboarding phase when you just signed up for a new fitness subscription and you really want to get fit. So you are in the mood of changing your habit. You’re open for a new mindset, you’re open for advice and whatever. So really winning those first hundred days of the customer relationship. So you as a health club member, PureGym will help you establish a sound habit of working out. That’s a very, very important part of their customer journey.
[00:09:43] Rasmus: Another part of their customer journey, which is extremely important. I’m using fitness as we can all relate to it. So mainly this part where people become inactive, talking about all the couch potatoes. So very dynamic insight that is unique for each individual health club member. When does he or she lose motivation for going to the gym? That’s totally individual. And in these cases you don’t only want to remove friction because it’s not about friction, it’s about really using everything you have to not make a more frictionless experience for the general customer, but really giving a really personalised and relevant and eye-opening and motivating personalised experience for the individual customer. Using all you have to make sure that you find out who they are and what is driving them and you match that with content which is basically what do you think will make them come back to the gym in this case. If you’re a grocery store, this could be like getting the third purchase, establishing a habit or whatever.
[00:10:45] Rasmus: So in these cases, you want to not only sew or, or sort of remove friction by sewing the customer journey ribbon back together. You want to tie a beautiful bow, which is memorable, which is adorable, which is pleasurable, which is something that stands out. Makes you give a genuinely unique personalised experience.
[00:11:02] Dara: So that’s probably a good point for us to tie this a little bit back to our favourite subject, which is around measurement. And maybe actually testing as well, or experimentation as well. So that experience, we’ve probably all had that. I mean, I get those repeat subscriptions and you get the email and it does say, would you like to also buy this? It’s quite clever. I know what they’re doing and I still fall for it so I guess I’ve got maybe, maybe two, two interrelated questions here. So, how much experimentation, and I know this isn’t a one size fits all, but take a typical, let’s say it’s subscription business like you mentioned. How much should they be experimenting with those different types of, of offers or cross-selling? And then the second part to the question is what would be the advice in terms of actually measuring that and understanding what’s going to drive the better, because It could be a conflict between maybe a short term upsell, but that might lead to them less likely to repeat further down the line. So I could imagine it could become quite complicated where you’re thinking, what is it that we’re actually driving here? Are we trying to get more repeat customers or are we trying to increase the average order value or do a little bit of both. What’s your kind of take on that?
[00:12:03] Rasmus: I mean you talk about value creation as I hear it. And when it comes to the three different ways of three different branches of marketers working in personalisation, I’d normally say you have the marketers working with campaigns, whether that be on owned media or paid media. So either like performance marketers, whether you also could also include like email campaigns, whatever. That would be like sort of one branch and you’d have the marketing automation branch, which could be another one. The third one would be the ones working with personalization on site or in app. So on inbound platforms and really measuring the success, it relates back to sort of which time perspective should you, that’s one dimension of measuring value. Which time dimension are we looking at here? Are you looking for short-term gains and the upsize which would probably be the case if you’re doing campaigns. So lifespan, time span of a campaign is normally fairly short, like a week. Or I think Christmas campaigns probably what the longest campaign that we do have that we normally do like for a month.
[00:13:00] Rasmus: But campaigns tend to be fairly short. So you’d be looking at conversions and sales, all those short term numbers and topics and metrics. If you are in a, in a business where you’re struggling to meet the targets for this, with quarterly targets, you tend to focus more on the short term and the money that you can create right here then and now. But if you’re looking into, for instance, and this, this is also where the, there’s a distinction between conversion rate optimization (CRO), which is very much about getting the conversion rate right here and now so you can count the money and you can reap the benefit, and you can see whether it was a good idea or not. However, some of the examples that I used before, you should also be looking at different ways of creating value. Because one thing is chasing the short term conversion, I’m not saying that you shouldn’t, but you can chase it to such an extent that you chase the customer away. So you may succeed with the ones where you do have the conversion, but the other ones, all the ones who don’t convert, who choose not to, and you’ve been applying a lot of pressure, really optimising going for that commercial. You’ll alienate them and you’ll basically chase them away.
[00:14:03] Rasmus: So what are other ways of measuring value and I think what you should be delivering in the way that, I mean the messages that you are personalising. So the way that I see it, the presentation consists of two things. So you have your insights on one end, in one hand, and you have your content in the other. And really, personalization is about matching those two and tieing that into a beautiful bow. So matching the customer journey was cut in half, insights one end, content in the other. You want to not only sew it together again and amend it, you want to make a beautiful bow. And the insights consist of sort of the, the segments that you do have will consist of the moment of truth that we talked about before, and the content will be either your creative messages or it’ll be your content feed as in your product feeds or your news article feeds or your show feeds in the case of Netflix, or song feeds in the case of Spotify.
[00:14:51] Rasmus: And you really want all the four to come into one personalised, hybrid experience. Not for any moment of truth in the customer journey, not for any moment, for all customers all the time, but for the ones where you know that it matters the most for your business. And then you should make up your mind whether you’re chasing the short term conversion, or you are also at least I strongly advise that you don’t ignore the long-term benefits from this. So you’ll be looking not only at conversions, but you’ll also be looking at what does this mean in terms of like emotional loyalty and long-term loyalty and customer lifetime value and such. So balancing the short term and the long-term result that you can be getting from this and also keeping in mind you’re messaging that you’re not always asking for the, like, the conversion here and then, but sometimes you just, you take the chance to convey a message that is genuinely helpful, don’t always be selling.
[00:15:39] Rasmus: So you may have like, say they have a hundred sort of trigger flows that you could potentially set up, a hundred signals about where the customer is, the customer journey. And 50% of these you would be linking a conversion or inciting a purchase or whatever. But the other ones, the other 50 would be places where you could help the customers, but it wouldn’t necessarily benefit you short term. Of course if you are, if you are about to succumb from like lacking revenue or whatever, chase the conversions by all means. But don’t forget about creating the long term value that will also keep the customers coming back next time and next time and so on.
[00:16:15] Rasmus: That’s one part of the answer, I think the other one is control groups basically. So this is more like into a measurement tactic. This is often forgotten by people who work with personalization because they genuinely believe that working with personalization is the right thing to do. So they don’t necessarily measure it. They’re saying, well we need this message for this particular segment, this message for that segment, this message for that segment because we believe it’s the right thing. And if we get to measure this against the control group, which is sort of randomly mix of all these segments, which we’ll be getting a sort of a generic message and seeing what does that create in terms of difference, in terms of short-term uplifts, conversions, customer lifetime value (CLTV) in the long term net promoters score and such. I strongly suggest that you use control groups for measuring this, not necessarily on long-term basis. So you keep 10% of your customers from getting into the onboarding program at PureGym forever, but at least once you have sort of a fairly strong idea about which kind of difference does this make in terms of churn prevention or uplifts or whatever you’re chasing then you have this report and at least you hide it in the drawer so when someone comes knocking, asking for, so how does this work? Basically, I want to turn it off, our CIO (chief information officer) says that we’re paying a lot of money for this tool and we’d really like that invoice to be lower, you can say, yeah, but it’s generating of this kind of value here, so strongly suggest we don’t turn it off.
[00:17:30] Daniel: Obviously keeping the control group that gives you the kind of overall uplift or I suppose downlift if it’s bad and experiment or personalization. But I think is there like layers then of testing the messaging within that as well? Because with the whole A/B testing is that you would actually test your ideas before making that permanent change of rolling it out. And I can imagine the same applies for personalization because just because it’s personalised doesn’t make it better or more profitable or more interesting. You might end up finding that you’re giving money away unnecessarily. Like you might be incentivizing people with a 20% discount when they wouldn’t of you know, they would’ve been happy with the 10. And I’m just wondering then, so how does that kind of tie together when you’re running personalizations, you’ve got your control group, which is the kind of, is personalization working. But then is there within that, like personalization options and testing assets and creatives within that, how does that all tie together.
[00:18:16] Rasmus: At a theoretical level that definitely is, I think in practice, I think depends also which kind of marketer you are. So say that you are, you’re working with an ad in advertising. I mean, you can’t run an ad without thinking about whom you want to show it to. So basically you’ll be working with some degree of personalization, at least segmentation, which I define as being under the umbrella term of personalization. So people working with advertising are working with personalization from the get-go and the point of testing that towards generic ad normally wouldn’t make sense because they, I mean, the ad may be generic, but at least they’ve made up their mind whom not to show it to. That is split testing that against showing the same ad to everybody just doesn’t make sense.
[00:18:57] Rasmus: But working on the, on websites and ecommerce stores, I think you can spend much more time optimising the general experience for the average customer and you can make a lot of money and returns, and that’s like classic CRO. I think you generally should be spending quite a lot of time optimising the general experience before you switch into personalising the specific experience for customers that you do know and where you have insights. That said, once you get there, I think that you should apply the same principles. So trying out the generic version as one of the versions in a multivariate test, for instance, seeing how does that apply and how does that perform compared to a specific piece of content, specific message for a specific audience, and so on and so on.
[00:19:43] Rasmus: So, applying the same principles of measurements and multivariate testing for the personalization tactics as you would for the CRO tactics. Often this doesn’t happen for marketing automation or for personalised campaigns on owned media. You’d rather just test the whole full personalised thing against the generic version. If you take it to the max, you should be applying the same principles there as well.
[00:20:05] Daniel: It’s really interesting and I like this idea of that kind of hierarchical view of it. And it’s almost like a maturity thing. Like the more mature of a marketer you are or a website owner you are, you kind of get to that point and it’s not running before you can walk. What I’m really curious about is the, is the how to handle and the kind of management over time, especially as this gets bigger of the kind of anonymous users and how when we are, for example, in our Google Marketing Platform, we’ve got an idea of our audiences we’re targeting, and then they go over to our CRM and they’re a different person and then they go to our website or app and they’re a different person. So the one thing is how do you communicate consistency there? Because I can imagine if they are expecting in getting a personalised experience on one of those, head to the other, and it’s kind of disappeared. I can imagine that to be almost maybe a negative experience for them because you’re saying you know, maybe more personalised stuff and then it kind of disappears over here.
[00:20:52] Daniel: Not everyone can be identified, let alone across all of these networks. And I’m just wondering, like how, do you have to then have a different strategy for the anonymous users? And does this mostly focus on those users that have identified themselves in maybe existing customers? Like how does that all, how does that work or how do you account for those anonymous people?
[00:21:07] Rasmus: I distinguish between what you would call implicit personalization and explicit personalization. Implicit personalization would be showing different content to different people, but not hundred percent sure who they are. I mean, you’ve seen them before, they’re unknown users. You have these, for instance, data points from their browsing history or their browsing session or maybe previous sessions as well. You don’t know who they are necessarily, but you have these ideas that this person could be in the browsing phase or could be ready to check out something soon or whatever. No, don’t know who they are and and they know that you don’t know who they are because they haven’t logged in.
[00:21:42] Rasmus: So if you decide to, say that even that you have the data shared and you know perfectly well, who they are, you wouldn’t show it to them explicitly because that would be totally creepy. So you’d be more subtle, more implicit in the way that you’d sort of make it almost seem as a coincidence than you were suggesting these products to them, even though you’re quite sure that this will fit like a charm with their full history because you already got figured out who they’re. But you decide to sort of keep that knowledge to yourself that I know who you are, I’m not showing it to you unless it’s completely out in the open that you logged in and you’ve made yourself aware and you’ve dropped your mask or whatever, and you’ve logged in, and then you can move from implicit to explicit personalization where you make it clear to the customer that I am showing you this because of this and it doesn’t matter whether it’s online or whether it’s in an email or whatever, but generally owned media tend to be a lot better when you’re doing explicit personalization.
[00:22:33] Rasmus: I mean, you haven’t really been called out by naming an in an advert anytime I hope, and you’d probably be super freaked out if you did. I once saw an abandoned basket ad on Facebook. I’ve only seen it ever once and I don’t know if they stopped running it or whatever, probably worked, but it was also a bit creepy because they’re not used to sort of being explicitly called out on a, on paid media. But once you do have that, and I think it comes back to like an emotional consent more than a legal one, and it’s not a technical exercise as such, of course, that’s underlying, we’re really figuring out who am I speaking to now and do they accept that we are having this conversation and that we, we are having a relationship here where it’s okay for me to be explicitly personalising the experience for this particular customer. Another thing that can affect this thing about explicit and implicit personalisation is if you’re dealing with sort of delicate topics.
[00:23:24] Rasmus: So say that an insurance company have you figured out, they know that you are in a divorce, for instance, because they can see on your browsing pattern that the propensity score of you being in a divorce of say 90% or whatever. They probably shouldn’t let you know that or be explicit about that. So they would keep that to themselves, but they may be sort of hinting at it or showing more articles about how to handle your insurance claims or your pension funds or whatever when you’re going through a divorce, but they wouldn’t say why they’d be showing that to you. On the other hand, if you do know who the customer is and they know that you know who they are and you have this history, transactional history, engagement history, click history or whatever, that enables you to give very, very good recommendations, then it actually drives up to 30% better performance if you tell them why you’re showing them what you’re showing them. It has to be like appropriate and not inappropriate, I don’t know if it makes sense.
[00:24:17] Dara: It makes complete sense and in a kind of theoretical world, it feels very kind of black and white that you as the user would know what you have previously shared. But I would imagine there must be a ton of grey area where you may have, even as a customer, given this information away before, but you could forget that you’ve done that or you could have done it in a slightly different context. So I’m just thinking, as the kind of brand it must be, it’s a tightrope to walk, isn’t it? Where you’re trying to keep it implicit until it’s absolutely undeniable that you’ve recognised the person and that they have consented to give you this information and have you used it to make recommendations to them? It’s not an easy line to walk, is it? To kind of keep that safe and keep it implicit only until. Because I could imagine even when it’s explicit, even if the person’s identified themselves, they could still think, do you know what I might have told you this, but I don’t really want you to now be marketing me these recommended products. I guess it’s a long-winded way of saying it’s not really black and white, is it, it’s quite subjective to how that user is you know, how they are at that point in time, what the context was when they shared that information and it could be that, that information maybe has come from a separate part of that business or even this third party data. It’s a bit of a minefield, isn’t it?
[00:25:29] Rasmus: And some would find it okay and some wouldn’t. So i’d normally say that the, the customer journey belongs to the customer. Basically, we don’t know what’s going on in customers’ lives, but with the data points that we do have, we can try and be less wrong, so to speak. We can guess with lower fail rates basically, but we don’t know what’s going on in people’s lives. We can try our best to propose some kind of content or call to actions or knowledge or information to them that we believe will make them make the right decision faster. But we basically don’t know, and if we manage to not insult anyone along the way, and not come across as creepy then I think we’re doing good job basically.
[00:26:08] Daniel: Actually just on that, that’s quite another nice sort of segway and this is around not insulting anyone or not stepping on toes. I mean this is, this is all going talking about the customer and the personalization to the end user at the moment. What about internally? So I’ve worked with lots of organisations over the years, even just mentioning around A/B testing and having that kind of testing culture or mindset of always be, you know, testing out a new creative and holding back a, you know, a sample audience that we can measure against and look at uplift studies. But like when we’re talking about personalization, that feels like another kind of cultural change for an organisation to go through. And if you know, the PPC marketer never speaks to the SEO marketer, that never speaks to the CRM manager, like, you know, it’s going to be hard to get it. They’re going to find it harder to understand this idea of personalization. So I mean, what has to be in place for personalisation to be a valid option? Like how would you know if you or your organisation or your marketing team, if you are the CMO (chief marketing officer) for example, how would you know this is something that would be useful for you?
[00:26:59] Rasmus: Yeah, that’s a good question. So first of all, I think there are business challenges and there are go-to-market models that make personalisation seem like the wrong choice of tactics. So say that you are struggling to get first-party data, either because you’re selling indirect through your go-to-market model. Or if you are having a product that is generally a really low involvement, take KitKat. Really fast moving consumer but again, really hard to come up and with some ideas for interaction that will generate you a sound body or big database of first party data, it’s almost impossible. So if you’re within that category, don’t bother too much, there are probably other fish that you should fry before going to customisation.
[00:27:41] Rasmus: Then there’s also the matter of whether you already do have a lot of, I mean, even though you have the chances and the opportunities to get first party data, you should also look at what is the current size of the customer database. Both in terms of custom records, people that you know and that bought from you and your sort of memory of these people. I mean, your transaction history, your click history, or whatever. Once you have a sound body of first party data though, then it should be most likely a good idea to embark on this. Yet you would still be looking at how you should optimise the general customer experience before optimising the personalised experience for the known audience, for the known customers.
[00:28:16] Rasmus: So, I think generally personalization is a advanced form of optimization. So applying much of the same mindset that you would in a CRO way of optimising that will be a good place to look for when you’re looking to organise most effectively, efficiently for personalization. In the book where me and Frans Riemersma, who is with Martech Tribe. This this European company based in Holland, working closely together with Scott Brinker on marketing technology side. So he originally came up with idea that these three sort of maturity levels for how your organisation is mature in working with marketing and personalization that you have these hack/pack/stack levels.
[00:28:56] Rasmus: So at the hack level, you’re basically just trying out all kinds of stuff, which insights work, which content work. You somehow implement that on an ongoing project base into MarTech and you see what works and what doesn’t work. And if you’re able to learn from your mistakes, I’m not saying that everybody is, because sometimes there’s a change in team and whatever you should be able to ascend or progress from the hack to the pack level where you know what you’re doing, you have a description for it and you’ve found those insights and that content, those content pieces that you know work, and then you can start optimising because once you’ve describes what you’re doing, you can reproduce it, and then you start optimising. Once you have that dialled in, you can basically turn that into it as almost as if it were a product.
[00:29:36] Rasmus: So you’re productifying your marketing basically because you’re also automating it. And once you get to that point in time, you can almost look at parts of your customer journey as if it were a product. I mean, most marketers don’t have any control over the product as such, but you can look at the onboarding part of the customer journey as a product. And once you have that automated and relevant communication set up and it’s running and it’s automated, then you can start looking at it as product and you can optimise it much the same way as you would if it were a product in fact. And in that case, you’d be looking towards organising in more cross-functional teams. So you’d have the PPC specialist, you would have the email marketer, you would have the website manager, or the ecommerce manager or whatever.
[00:30:17] Rasmus: They’ve been working together in looking at how can we make sure that the onboarding part of the customer journey becomes so much better, or that you have another team working at the winback part of the customer journey. How can we make that product better by continuously optimising what is already there and how customers are meeting your brand. This particular part, customer journey.
[00:30:38] Dara: So Rasmus, you kind of briefly touched on one of the concepts from the book earlier, which was this kind of bow tie concept. Could you maybe go over that a little bit more and then also talk to us about the second concept in the book as well.
[00:30:49] Rasmus: Yeah, sure. So long time ago I learned that people don’t actually read business books, they mainly flip them through and they look for useful and colourful models that they can put into a PowerPoint presentation tomorrow, and immediately sound wise or more clever than their peers. And I took this very seriously, and so since then I’ve always looked for ways that I could do models that would help convey the topics that I’m writing. And you are completely right, the first model and perhaps the most important model of Hello $FirstName is what I refer to as the bow tie personalization, which is well, it comes from this metaphor. I don’t know if you ever tried tying a bow tie of your own, it’s incredibly hard. It’s incredibly easy once you get the hang of it and I would argue that personalization is, it’s just as easy as tying your own bow tie. Some will find it hard obviously, and this is because you have in one end, you have your insight part on the other end, you have the content part, and you need to sort of tie them together into a beautiful bow, a beautiful bow tie.
[00:31:42] Rasmus: And the insights come in two parts they come in the, like the customer types, that would be the segments. So which people are we talking about here in the customer database? That could either be like demographic segments, it could be value based segments. High spenders, low, or the other part of the insights would be the moment of truth. So moment of truth where in the particular, where the customer journey is for each particular customer right now, mostly built on behavioural data or advanced analytics as well. Content part would be either the creative messaging, what is it that you want the customer to do? What is it that you want the customer to feel? What is it that you want them to know? So basically that would be especially important in outbound communication, that this is where the marketer pours in their creativity and where they make sure that they insert parameterized personalization and such. This is the content that they’d be crafting.
[00:32:27] Rasmus: And then last part of the boat, our last corner of the boat, I will be the content feeds. Which would be content coming from like other departments, other data sources where the marketer doesn’t necessarily change the content as such, but they sort it and they rank it and they divide it and they filter it, which part of this particular product category to show to a particular audience or whatever. And the hyper-personalization is then when it all comes together so that’s one of the main models in the book. This ideal for understanding what is personalization and getting a common view, common understanding of what is this and how can you easily, more easily onboard new team members to your team? How can you easily get to a page where everybody understands what is it that we are working on here. Which parts of personalization should we be looking at.
[00:33:09] Rasmus: Then there’s the second model in the book, which is the pyramid of personalization. And the pyramid consists of a front end and a back end. Front end would be three levels of scope that you’d normally build, you normally personalise when you’re doing your personalization, either in campaigns or three maturity levels for how you personalise within campaigns, three maturity levels for how you personalise within market automation, three maturity levels for how you personalise when you working on inbound platforms. And the back end of the pyramid would thus be, what does it take for an organisation to support this kind of work with personalization? That would be, which technology are you applying and how mature is your way of working and deploying technology to make it work for you? How can you automate more stuff and such? Which platforms do you normally deploy in which order and it will be your people and skills.
[00:33:54] Rasmus: So we talked a bit about this earlier, how do you structure your work with personalization in a way that you’re gradually more productifying your marketing, working more and more automated, ultimately with a customer journey focus as opposed to a conversion focus or project focus, with a cross-functional team that’ll be focusing on one part of the customer journey as if were a product. And then the last part is the governance and process part like a six-sided pyramid basically. So this will enable you to sort of gauge as a marketer or as a CMO, which level of maturity is our organisation currently at, and what should we be doing to sort of break the glass ceiling to the next level?
[00:34:32] Daniel: A lot of this Rasmus and I appreciate that kind of hierarchical view of it. The kind of the pyramid of that is always about kind of like pushing through and it’s like, okay, you don’t need to have a solid foundation before you move on to the next layer and there’s no point running before you can walk because then the whole thing will crumble down because you’re not quite ready in a mindset culturally, organizationally or even just technologically. There is a point, and you mentioned something earlier around like different organisations and different verticals that might not be, it might not be relevant for, so like is there an argument there to be like, okay, it’s not always about chasing the next thing. I mean, we work with lots of clients on their analytics setups and often we kind of talk about analytics maturity or marketing maturity and you know, sometimes they’ve just found a happy equilibrium or something that, you know, either a third party or someone in the analytics world can’t actually help them culturally change and progress into.
[00:35:14] Daniel: Is it always about kind of like getting towards that personalization or is there a quite a happy medium where if you get, for example, a very, you know, a 200 year organisation that still has a digital team that covers everything from CRM to email to web and stuff like that. If they even start thinking about testing a creative in an ad copy, that’s probably as far as they’re going to get because it’s not about us being able to help them progress it’s actually more of a systemic, kind of cultural thing maybe that is a bit out of control or out of our power, I suppose.
[00:35:41] Rasmus: I did mention before that, there are some business challenges that may not necessarily be aware, personalisation may not be the right tool, and there are definitely other ways of making good business than through personalization. So say that you have an extremely strong product that sells itself. Well, maybe you are better off optimising somewhere else in the business. Maybe you should be working on how you are distribution system is and how well you are getting the products to your customers because really, it sells itself. Or if you are, we did talk about KitKat earlier. I mean, it’s much more important to make sure that you make deals with retailers where you make absolutely sure that KitKat is placed just in front of the till where children can reach them, much more important than doing personalised marketing. So definitely I would say that there are cases and businesses and industries, maybe even where you are better off, not personalising because that’s not what’s going to move the needle.
[00:36:37] Daniel: I appreciate it was a badly worded, non-question, but I have one for you because it can’t be a podcast in the year 2023 if we don’t mention about large language models (LLMs), AI, generative AI and things like ChatGPT, so I mean that’s big enough in itself for another hours conversation. But I mean, where does this fit in and where’s the industry moving towards incorporating, rejecting or kind of working alongside these kind of models?
[00:37:01] Rasmus: It’s an extremely interesting question that I also did spend some time trying to answer for myself. So coming back to the main model of the book, you can say that we’ve actually been using, not generative AI, but advanced analytics and AI in that sense for figuring out the insights of our customers. So which customers are most likely going to leave us, and for which reasons? So we’ve been using AI to try and crack that nut for years. But at least AI has been helping us with analytics, churn propensity scores or category propensity scores or expected add to basket models or check out whatever. So expected customer lifetime value, stuff like that, helping us to prioritise which customers to give which treatment.
[00:37:44] Rasmus: AI has also been helping us to sort on rank product feeds or news article feeds or your Facebook feed or whatever for years now. So trying to make sure that whatever was most relevant in the feed of content would be ranked on top for each individual customer. So really what is new and the last part of the bow tie that hasn’t been susceptible or hasn’t been affected by AI up until now, that would be the content part, the creative content part where marketers have been putting in their creative efforts in order to move people, in order to make people make different decision based on lines of argumentation, based on creativity, based on humour, based on the right yeah, call to action you could say. And this is where I see that generative AI can help the marketers quite a bit.
[00:38:37] Rasmus: I’m not sure I would entirely leave it up to the generative AI too. I mean, I still have a look at it like, from a human perspective, would this really work for this particular segment here, or this particular audience, or this particular moment of truth, I’ll still have a look at it, but I think it’s pretty safe to say now that generative AI can definitely get you to 80% of the creative message that you want to put out there very fast, and that opens up for the idea of having so many more variants of your content. So whereas in the past you may have only worked with like three to four to five large segments of your customer database and really haven’t bothered to make more creative variance of your messaging than for these high segments. Now you can take this much further and you can make content for these four or five segments very fast because you can have generative AI propose to you this content within a very short timeframe. Or you can use that extra time to make more variants for the smaller segments as well.
[00:39:34] Rasmus: So I think generative AI here in terms of coming up with text, doing language versioning, making it more engaging, more humorous. Making the images that you use, producing them dynamically. I think we’re going to see a huge explosion of generative AI being also native parts of the personalization platform that people are using.
Wind down
[00:39:53] Dara: So just two more questions and thank you again for coming on the show to talk to us about this. This has been a really, really interesting conversation. So the second last question is where can people find out more about you or get in touch with you if they wish to do so?
[00:40:06] Rasmus: So I’m open for networking on LinkedIn. I think that’s probably, I’m fairly active on LinkedIn, so I should be fairly easy to find there. I realise that spelling Houlind may not be the easiest task for native English speakers. I’m sure my name will be in the show notes, so that should be doable. Find my book Hello $FirstName, you can find that book on your local Amazon store. So do check that out, I’ll also be happy to send you an abstract of the book if you care to check it out. And my employer, Agilic, one of the best marketing automation systems in the world, have to say that obviously, is giving away 10 copies of the book. We’ll put the link in the show notes if you follow that link, we’ll do a prize draw of 10 copies of the book, so you can check that out as well. I think that’s probably where you should be able to find me. If you can’t find me, I have failed.
[00:40:50] Dara: That’s brilliant. So thanks to you and to Agilic for the kind offer of the 10 copies of the book. Final question what do you do outside of work to wind down?
[00:40:58] Rasmus: So to wind down, I was just about to say, I spend a lot of time with my kids, but I don’t really wind down much with three kids. So when I do manage to escape them as well I normally go kite surfing or mountain biking or running or cold water swimming has become really big in Scandinavia. I swim all year round, not for lane after lane, but at least until I get cold. Learning a lot about longevity, I’m doing a bit of angel investments as well, which is hardly free time, but still super fun.
[00:41:28] Dara: You’re a much braver man than me. I’m still stuck on the cold water swimming, to be honest, couldn’t, couldn’t do it.
[00:41:33] Rasmus: I occasionally force colleagues to try it out. But yeah, it’s fantastic once you get into it.
Outro
Dara: That’s it for this week, to hear more from me and Dan on GA4 and other analytics related topics, all our previous episodes are available in our archive at measurelab.co.uk/podcast. Or you can simply use whatever app you’re using right now to listen to this, to go back and listen to previous episode.
Daniel: And if you want to suggest a topic for something me and Dara should be talking about, or if you want to suggest a guest who we should be talking to, there’s a Google Form in the show notes that you can fill out and leave us a note. Or alternatively, you can just email us at podcast@measurelab.co.uk to get in touch with us both directly.
Dara: Our theme is from Confidential, you can find a link to their music in the show notes. So on behalf of Dan and I, thanks for listening. See you next time.
Daniel Perry-Reed
Subscribe to our newsletter:
Further reading
A step-by-step guide to migrating scheduled queries to Dataform
How to set up a Dataform repository with GitHub & Google Cloud integration
