Optimization 201 Transcript

24 March 2022

Cass: To kick things off, what does personalization mean to you?

Tracy: As a dev/strategist who’s done a bit of everything in this field for many years now, it’s definitely been evolving quickly. Before we focused on the idea of a consumer, or statistical consumer segments. And now we are really honing in on an individual level. I think it’s just more attention to turning consumer to a customer

Cass: And Marilyn? 

Marilyn: I think it depends on who you ask. A lot of people, when they think about personalization from the brand perspective, only think about product recommendations. Product recommendations are certainly part of it, but personalization is a strategy (not necessarily a tactic or a tool). And so it’s about being able to create an experience that’s made just for me. Building that obviously takes time, and also takes data – it requires a lot of data structuring as well. And, of course, it takes tooling, but before any of that, it’s a strategy. So you need people to think in terms of personalization, in the way: 

  • things are deployed
  • things are built 
  • campaigns are executed, etc. 

So for me, personalization is definitely a strategy

Cass: Something that has to be really deeply embedded. 

Marilyn: Yeah, throughout the organization. And you can have different degrees of it, depending on your resources, your data maturity… With today’s privacy issues,  brands have a responsibility to manage that as well. So that does add an extra layer of complexity.

Cass: Can you guys mention maybe an experience you’ve had where you were online and you experienced personalization in action? Maybe it was a good experience, maybe it was a bad experience. 

Marilyn: One common issue is when you visit a website, and actually make a purchase – but then you keep getting retargeted with it. It’s like, no, I bought it already! So, I don’t need to see that ad anymore – but then I see it everywhere for days

I’ve also had problems when I wanted to return an item but the return button didn’t even work. As somebody who has purchased something, you should be able to simply recognize who I am, because I’ve created an account. I’ve put billing information, I’ve ordered something, and now I want to return it. At the very minimum, I should be identified and be able to go through that user journey without a hassle. And you should test that user journey to make sure that you’ve achieved your goal: to get a paying customer. Now you’re losing that paying customer because you’re giving them a terrible experience returning an item. 

So I think these days a lack of personalization feels almost like a personal sleight. Even though maybe I’m just a random user out of hundreds or thounsands. Like, I had to return this and it didn’t work – it can be so annoying, you feel like it’s a terrible website, terrible experience, and so immediately you start forming these opinions about the brand even though it’s just some error or bug that somebody missed. So those are a couple of experiences that come to mind. 

There’s definitely a lot that goes on there in terms of UX. For the customer, it might not be personalization “gone wrong,” but it feels personal because I’m a user and I’m having a one-on-one experience with the brand. So it’s important you get the whole journey right!

Tracy: Yeah, we’ve all experienced that I think. Especially from from our standpoint, as an agency. And usually we’re there to fix it, so I feel you. To give one example, think about Siri. You always kind of feel like your computer is listening. You know, when you’re talking about something and you start seeing ads for it everywhere. I mean I know it’s psychological, because your mind’s on it, in more cases than it’s actually targeting. But you wonder sometimes. 

Cass: Can be good can be bad! And also, thinking a little bit about this idea of creating that really personalized experience… As we move toward the “cookieless future,” and the importance of first-party data, can you talk a little bit about contextual personalization and why has that been such a hot topic in marketing lately

Marilyn: What I see, from more of a macro view, is that companies have to want to attract users and consumers. They want to encourage them to understand that the brand is worthy of that information… that they’re gonna properly manage it, etc. And, there needs to be a value proposition in exchange. 

So, if I’m gonna create an account, there has to be something driving me to create this account… Why should I:

  • Create this account? 
  • Sign up for this newsletter?
  • Give any information to you as a brand?

And so I think brands have a responsibility to customers today to understand what is important to them. It also has to be messaging that’s adapted, right? So I think there’s a lot of testing that needs to be done around this topic of messaging. In those first touch points with a customer, you have to make sure that you are creating that sense of trust. You are communicating very clearly with the customer that if we take their information, we’re responsible for that information

Then, on the other side, what am I going to give in return? In exchange for answering a couple of questions, I’ll give you:

  • a personalized shopping cart or 
  • insider access to sales or 
  • You’ll get a more personalized shopping experience 

So it’s more about thinking about, “how can I provide value for the customer in exchange for that personal information?” I think brands that are successful in managing privacy and still getting sign ups and more interaction with customers are the ones that manage to improve their messaging, and create that trust. They do a lot of testing around these entry touch points. 

Tracy: I’ve seen it evolve starting decade ago from a standpoint of, Well somebody does this on the site, so show them A or B and just kind of send them down a logic tree… Or even with an on-site questionnaire, to try to show them what they’re looking for (or, better put, guessing what they’re looking for). And usually this kind of approach has shown very positive results. So it’s something worth pursuing, in a world where you don’t have access to those third-party cookies anymore. 

Cass: It’s kind of “back to the future” right? For the last few years, we got so specific personalizing to individuals and now we’re kind of moving back to a more linear experience.

Marilyn: Well yeah that decision tree logic – it really has to do with a combination of factors. You can have simple “if, then” logic, but you can also use AI if you have the right tools in place. These tools can add to the data you have access to – this is why the privacy part is important because if there’s only so much you can do, depending on where you are and what rules you have to follow. 

Here in Europe, and in France specifically, there are a lot of restrictions around data. You have less leeway around what you can do. If a user that visits your website and they click on one thing and you just have that “if and then” logic, then, if they click on shoes and they’ll see socks. This is where recommendation engines come in. A lot of of people right now, when they think of personalization, they think of product recommendations because of the way that you can use volume and data to feed an algorithm and be able to propose something contextual without a lot of personally identifiable information, right? 

That’s where AI comes in.

That’s where you’re using historical data or behavioral data, which will help to define or modify these algorithms. But the reality is that you can only go so far because you have different audiences, you have different segments that have different preferences. And so, if you really want to have impact with contextual personalization, you need to at least be able to segment your audiences. So you need information from those audiences. 

Contextual personalization is good, but it’s kind of logical that when someone logs in, you know who they are and, if you have a CDP then you can know what kind of audience segment they’re in. Then you can have a treatment that recognizes that user as a certain type of user. 

Maybe it’s a high lifetime value (LTV) customer that returns often that shops certain things. Maybe they’re frequently looking at black dresses. Well, the algorithm is going to tell you that this audience is always interested in what’s going to be hot next season. So instead of the black dress from last season, you’re gonna show the one from this season first. So I think the more you’re able to convince that customer to share their information, then the better your contextual personalization is going to be, of course. And so, there are a lot of different audience segments, some that are higher value and some that you don’t know yet because you’re just starting to to interact with them.

I guess the challenge for companies is, how much time do you invest in long-tail segments vs. how much do you spend on those high LTV customer segments? So, you have to have a value proposition and create that trust to get that information to be able to better personalize, so you don’t have to rely on only contextual personalization moving forward.