In this episode, you’re going to learn about how to deliver segmented experiences. It’s a terrific episode you do not want to miss. 

—

Welcome to the 2x eCommerce podcast show. The 2x eCommerce podcast is dedicated to digital commerce insights for retail and eCommerce teams. Each week on this podcast we interview other commerce experts, a founder at a digital native commerce brand, and a representative from a best-in-class commerce SaaS product. We give them a tight remit to give you ideas you can test right away on your brand so you can improve commerce growth metrics such as conversions, average order value, repeat customers, your audience, and ultimately sales. We’re here to help you sell more sustainably.

Speaking of which, I was joined by the Head of Partnerships at an eCommerce personalization app called LimeSpot, and his name is Francis Pilon. The conversation you’re about to learn about is one around the interfaces and the fundamentals of personalization. He has a unique approach to personalization, which is, the Crawl, Walk, Run approach.

According to a survey, about 80% of eCommerce merchants and leaders think they’re not doing enough about personalization. Starting the process can lead to further improving the sophistication of your segmentation, your messaging, and eventually, the segmented experiences we’re trying to deliver. What is quite synonymous with personalization is segments, which are at the base level, and the experiences you’re delivering to each segment. The title of this podcast will have the term, Segmented Experiences. I love that term when he mentioned it.

To give you a heads up, in September 2022, I would announce the specific dates but it’s going to be Thursday and Friday, we are going to be hosting the Commerce Excel Virtual Conference, which is exciting. I have an interesting lineup of speakers. I’ve gotten confirmation from close to twenty speakers for now.

In 2022, we’re going to do a lot of expert panels because that was where the fun was in 2021. The expert panel was where the real juice was. It was where a lot of data were shared but that’s not to put down the keynotes because we had some terrific keynotes in 2021. The streams are going to definitely change in 2022. The conference is going to evolve as my knowledge evolves.

In 2022, we’re going to have a panel around sustainability. We’re going to have a panel on health and beauty. We’re going to have an expert panel on not just retention but lifecycle marketing. There are going to be some unique discussion panels with world-class experts and speakers who will infuse a lot of knowledge into you.

Sign-ups are not yet ready. The website’s going to be ready fairly soon on CommerceExcel.com. That’s it. Here you go. Enjoy this conversation I had with Francis Pilon from LimeSpot. It’s a nice one. He’s very knowledgeable and there is a lot to learn if you’re considering personalization, regardless of what platform you consider using. It’s all AI these days. It’s AI-driven personalization. Cheers. Thanks.

—

Francis, welcome to the 2x eCommerce podcast.

Thanks so much kindly for having me here. I’ve been a longtime listener of your podcasts. I get a lot of value from it. I love how insightful you are. I love the softness of your question to the invited speakers. You can get more out of them and so it’s a real pleasure to be here.

Thank you so much. That helps. At the same time, I’ve been looking forward to this conversation. We were meant to speak in Q4 of 2021 or Q1 of 2022 and it never happened but here we are. We’re having this conversation sometime in June 2022. How are you at LimeSpot?

I’m great. I joined LimeSpot in March of 2021. It’s been more than fifteen months since I’ve been here. We’ve seen some great progress built from adoption in the market but also some of the products that we’ve released over that period. It’s great. The timing, overall, of the general interest and maybe how merchants are starting to value personalization is also growing. I sometimes joke and say that personalization is the new review where you must have reviews on your site the same way that you must have, for example, product recommendations. It’s a great time to be in this space.

It is. It’s relevant. There’s so much going on with personalization on the outside of websites. Whether you’re talking about what’s going on on iOS, what’s going on within the app system, Facebook, and all of that, there are so many personalization opportunities on-site in eCommerce websites, which is where LimeSpot sits. It’s how you consolidate all of that data you’re getting whether it’s from shoppers or your customers and how do you give them a better or the best experience possible? That’s your primary thesis. Let’s get into LimeSpot itself. You’re the Global Head of Partnerships at LimeSpot. How was LimeSpot ideated, if you have the story? What is the story behind LimeSpot and then we’ll jump right into tips you have around personalization if that’s okay?

You’re right. Personalization is getting pervasive in terms of how it is used at large, not only for eCommerce but also for many areas of the world. We’re at the tipping point where there are also some fears initially about how creepy AI was. Now, we’ve managed to find some balance also with privacy and making experiences, whether it’s eCommerce more.

It’s more relevant without that fear of being too restrictive and too personal in some ways. It’s natural if you see that and it segues into the why of LimeSpot. At one point, all of these tools, this ability to personalize and create those unique experiences for users were not available to smaller merchants. That was one of the geneses of or one of the key factors in the genesis of LimeSpot to be able to democratize, if you will, the ability to create personal experiences at scale. That’s one part of it.

The other part of this is we know that choice overload, especially with merchants with large catalogs, can be a problem for users. Both in terms of being able to get the friction in a journey, but also the satisfaction that they get out of their purchase. Sometimes, I experience this myself. I go out and I want to buy something and have 5,000 choices for this one thing that I want to buy. I’m looking at criteria and metrics. I come out of it and I’m like, “Have I made the best choice?”

If you can put a little bit of personalization into being able to understand the context and the user and guide them toward a subset of products that are most likely to be suitable. We know, and research shows, that it also leads to greater satisfaction with the brands but also in the actual purchase that they’ve made. These are these two as they probably are the key ingredients that got LimeSpot to get started.

I’m going to jump into lots of questions around your thresholds for personalization. You mentioned the fact that when your catalog starts to grow to a certain size, things happen. One of the points I wanted to make before we jump to that question is no one ever complains about the personalization or personalized results they get on Amazon. Never. Where you hear issues is, “This ad is following me everywhere else.” “I was on this site and this ad is following me everywhere.”

The complaint about how Facebook and other social media platforms use their data but when they’re on-site, they want relevant results because they’re concerned about their time. It’s all about time. Can you save them time? Can they get what they’re looking for as quickly as possible? It’s interesting in itself. My question has to do with the typical merchants. With eCommerce directors, how many SKUs should be the determining end-point to start to take, particularly on-site personalization, much more seriously? Even email recommendations, when should merchants start to think about personalization?

I want to add to your previous point about how ads are following you around. One thing to know is that on-site, the intent is different. If you’re on social media, you want to connect with your friends, family, or whatever. The intent is not about shopping and this is where there’s a bit of a disconnect. While if you’re on-site, if I’m looking for a pair of shoes on the site, then this is the context in which these recommendations are made much more appropriate to where I’m on a journey so to speak. That’s why maybe we’re seeing more of that dichotomy between the contexts that make a lot of difference.

To address your question, there’s always time to personalize. Let’s put it this way, the challenge arises when you have a lot of products. In the case of LimeSpot, we say about 40 to 50 are coming down the threshold and a lot of customers, because if you have fewer products, you may have millions of customers, if you have two products, it’s easy to merchandise.

When you get to that threshold of 50, the manual work that needs to go into this multiplied by the number of variants of users of types of users or where they are in their lifecycle, all of these makes it exponentially more difficult to be able to create relevant journeys for those customers. It’s a factor. Think of 40 to 50 products and then you factor in the number of customers. If you do have a catalog of 5,000 products and 5,000 customers, we’re probably looking at another great use case for personalization.

We’ll get to this but the only thing I would want to add quickly is the idea of personalization. So far, the discussion has to be around product recommendation but it goes much further beyond this and there are a lot of opportunities regardless of how many products you have to personalize. Imagine you have a catalog of five products, but you have people who have bought from you before and people who haven’t. For these two streams of customers or users, we should be able to personalize for them regardless of how many products we have because they’re expecting something different from the brand.

If we go one level below personalization, which is the outputs and the outcome, we are in this space of segmentation. From your point of view, what’s the best way to segment and how would you deliver experiences at that level? The way Walmart will segment will be totally different from a fast-fashion brand that has maybe 300,000 active customers. Do you have any pointers on segmentation?

You alluded to the first segmentation which is returning customers and first-time customers or potential first-time customers, which is a given even from a point of view of the messaging, the kind of offers you can give a first-time customer versus, returning customers. Do you want to break segments down so that eCommerce directors reading can take on to test or start to create segments with the intention of eventually running out to a much more sophisticated personalization strategy?

Sure. We seek out three basic types of segments. One of them I would call Life Cycle, those who have potential first-time buyers and we have frequent buyers. Wherever they are in their life cycle and relationship to your brand is the first one. We have Affinity Segments. Affinity could be based on brands so if you have a multi-brand store like Nike or Adidas and be able to understand this and make sure that their journey is aligned with their preferences would be one of them.

The type of affinity, a segment that we often see is based on spend. Low spender versus high spender will be unveiled to be able to understand what their expectations and limitations may be. It also can be useful in personalizing. The third one would be contextual segments. Are they seeing the site on a browser on a desktop or mobile or are they dealing with a specific campaign like a paid campaign?

Geolocation could be another one. These are the three basic segments. From there, you can layer some of those so you can have somebody who’s a high spender coming from a specific location who has never bought from you. You can get more specific from there but these are the foundational layers to build upon.

How would you approach each of these segments? I know that best practices are quite cliché these days but typically, what are the best practices? What best-in-class customers LimeSpot do from an execution perspective for each of these segments?

I’ll give you an example. The most obvious because it’s probably the clearest. We do also have an Ultimate Guide to Segmented Experience, which offers a number of scenarios.

I like the term Segment Experiences. I’ve never heard of it, but it makes sense.

Going back to the prospective buyer and those who have bought from you before, these users are expecting something entirely different from your brand. The one who hasn’t bought from you needs a little bit of brand trust building to be done there. Why should they buy from you? You need to put yourself in the shoes of your customers. I’m going to step back a second here and say that most merchants and probably most businesses understand the concept of persona. It’s been out there for many years. Most merchants and most businesses also understand the concept of the customer journey.

We’re at the point where we now have tools to be able to put these things into practice like what we would do fifteen years ago when we were talking about those personas and their journeys offer a different menu version of thinking about the journey. We were giving them all the options and then choosing their own adventure whereas now we can say, “We know what the best adventure for you is. You can always go on the sidetrack, but here’s where we want you to step in because we know that this is the most successful and most rewarding for you as a customer, and the most rewarding for us as a merchant.”

Whenever we talk about those segmented experiences, we need to step back and think about those things. Think about what the journey is. What are the expectations? What are the barriers? What are the friction points that a customer would meet in their journey? We address those. If we take the example of that potential first-time buyer, they need to trust you. Why should they buy from you rather than your competitors?

Maybe there’s an opportunity in your experience to showcase some testimonials or maybe bubble up reviews that showcase the value of your brand. We can also maybe surface some of the shipping or return policies a little bit more front and center for those users. Maybe the offer that we can put there is going to be best aligned with what they’re expecting.

How do you do this? Do you create personalization blocks like content blocks that are dynamic and the system will pick up based on the segments? Do we serve it and say, “Plug this in,” when the segment comes in?

Exactly. To wrap this up on the other side, to hear somebody who’s bought from you before, we don’t need to be that similar brand trust building. Maybe what we do there is push bundles. They’re going to be more ready with our subscriptions even right there. If they’re back, they’ve already had a positive experience so it’s time to kind of go up the ante, if you will.

That makes sense. There’ll be 2xecommerce.com/LimeSpot. We talked about segmentation. Based on segmentation, you’ve created a segmented experience, which is not synonymous with what we call personalization. Are there any other critical pillars in personalization that readers should be aware of?

There’s the experience on-site. If we want to think a little bit more holistically about the customer journey, there’s the post-purchase and anything that’s post-purchase flow so the ability, for example, to understand, even if it’s not necessarily an upsell, cross-sell, or some transactional follow up after a purchase, the ability to make sure that they have access to resources. They bought this thing for the first time, they require some setup so why not send a video that allows them to understand how to put this thing together. That’s the one step.

The other one is going a little bit beyond as well so the ability, for instance, to go outside the main site. Email is one of the channels. Most ESPs, and most email providers have some personalization so there’s segmentation, there’s often some product recommendation. That will probably be another area in what we’re seeing also emerging both as a communication channel but also an opportunity for personalization is going to be SMS. The ability to be a little bit less generic in the way that we communicate to customers via SMS.

What are the real estate limitations in the SMS for once? With an email, you can put a block with three product recommendations if you want to, but from SMS, what are you seeing? What trends are you seeing from a personalization standpoint to have been very targeted?

We do have the ability to content so it’s specifically text-based if you want to unless you want to go into the MLS route, which merchants have yet to see. I don’t have the latest number, but it seems to be lagging a little bit in terms of adoption. For instance, in the case of product recommendation, by default, most SMS clients will load a preview of the page. If you were to send, for instance, a link to a personalized product, you would show the product card.

Essentially the next serve of that product page will show a preview in the text. There is some flexibility, especially around product recommendations. It kind of works but it’s still early and there are a lot of opportunities also for building on this and hopefully text gets as it gets a more mature communication tool, which also gives us a little bit more flexibility and merchants a little bit more flexibility in terms of what can be shown to customers.

Is there any opportunity for personalization in regards to social proof? By this, let’s say I’m looking at a product. I’m looking at a laptop or an office chair for instance and I’ve been sitting on the fence. Is it possible that some personalization engines can see me and definitely put me in a segment of potential first-time customers?

Pulling out data for that product reviews for that particular product I’m looking at and updating me via whatever channel whether it’s SMS, “Jessica bought the same chair. You looked at that on our website. Here it is.” It’s using personalization as a hub that pulls in data from disparate data sources that are relevant to the product and relevant to the customer or potential customer. They are rest assured that, “I’m not alone in this. This is the experience of others. I’ll commit.” There’s a persuasion tool.

It’s a little bit tricky and if you’ve been sitting on the fence, you definitely need a chair. We can use this in a number of ways. At LimeSpot, we do support the concept of custom triggers. For example, somebody has left us a trail of data that allows us to segment into an audience. It can be a great tool that allows you to put additional context into your review. You say, “It fits large.” “I have dry skin or oily skin,” for example, that zero private data could then be leveraged to create campaigns that are more aligned with somebody who may also have the similar kind of criteria.

I don’t feel that you’re kind of one on one but this is probably falling onto the creepier side of this, but being able to bubble up relevant information based on zero party data or data that comes from other sources that align you with a group of people that share similar characteristic, and then being able to put reviews that are more relevant to reviews as possible.

If I visited a particular product page three times in the last seven days, there’s interest but I haven’t triggered or added to the cart. If I’m getting information streams, let’s say you have my email of reviews of that particular product, video reviews from influencers, and reviews from normal customers, that’s going to serve, flip or change my decision, whether good or bad. That information might be relevant. You should be careful as to getting into that creepy stage.

This is a common flow. The browser abandonment flow, for example. Most ESPs will take care of this. Personalization allows you to layer on top of this. You’ve looked at this product. The browser abandonment will try to bring people back and it’s the same for cart abandonment to that one product but sometimes we leave out the other possibilities.

Maybe that product was not a fit. Maybe you missed out on products A, B, and C and then the ability to understand the context of that customer, and being able to then present other options that may be related to one of the products that we’re looking at. It’s also what they’re the signals that they send is an inefficient way to re-engage customers into bringing them back to the site and re-engage them into their shopping experience.

It’s interesting because from an ESP standpoint, you’ll often get stuck in these abandonment flows that are transactional in nature, whereby it’s incentive-based, where it’s 15% off, but you’re not getting the context. You’re not even helping people get the context if they’re specific to that product, and then you’re sending me relevant information to that product to help me make a buying decision, which is more relevant than saying, “With this 15%, you pre-judged them.” The decision is based on price, which might not be the case.

You’re right and it’s probably building both. There are some people who would stand back like this and hope to get that 15% discount but some are generally on the fence or may not have seen the value and whatnot and then being able to kind of re-engage them with different tools and different ways. You’re right there.

Is there a use case for using personalization to minimize the return process? Are you seeing any of those merchants looking at the broadly saying, “These are objectives. We have a high return rate.” In use cases, what use cases can we use from a personalization standpoint to pre-educate the customers as much as possible so they don’t get what they thought they wanted and there’s that discrepancy. 

It’s a bit of a tricky one. I’ll give you two cases where recommendations can be useful. The one we were talking about is allowing product reviews based on maybe zero party data or some quiz that they may have taken. It’s aligning reviews with their size or type of skin or whatever characteristic might be to other customers so they can understand the context in which that product was reviewed. It could be a useful one.

Generally speaking, the other one is product recommendation especially as we have more data about given consumers so they’ve bought from you before. They’ve returned 3 items out of the 4 they’ve purchased but now we know what they were interested in. We know what doesn’t work and we know what works. In product recommendation, what we’re seeing, generally speaking, although it’s not one solution that fixes all, tends to also reduce a little bit of return and often allows the merchant to discount those products less. We see that product recommendation doesn’t typically need as much discounting as non-recommended products.

There’s a feature I picked up on Amazon for a pair of shoes whereby they looked at my previous transactional data and purchased data from shoes I bought prior. When I tried to purchase a new brand of shoes, they had an idea of my shoe size now. To let you know I’m always a size ten. With Nike and maybe 10.5 and with ASICS, I’m 10.

I decided I’m going to buy a new brand that I have never heard of. Amazon is alerting the size level like, “We think this is your size based on prior purchases.” Are you seeing more use cases in that because if I pre-selected the wrong size I’ll be frustrated especially if I needed it yesterday? I have to return it and try another site. Are seeing hints and personalization from a guidance standpoint, particularly at an SKU level?

It’s more rare and the reason for this is, this is probably for a much higher sophistication merchant. I want to bring in something quickly. We did a survey in Q4 and Q1. In that survey, there were some interesting things that came out. One of them is that about 40% of the respondents have somebody who can handle personalization. It’s the same in terms of importance or value in the idea of personalization. However, 80% of those merchants are in a Set It and Forget It flow if you will.

There’s still a lack of investment in personalization and what we’re discussing now falls into that much more advanced type of personalization. If we’re going to scale these back a little bit, this idea of Crawl, Walk, Run. Start with putting product recommendations in your cart for up-sell, and cross-sell. Put product recommendations earlier in your journey to guide that journey.

Maybe start A/B testing. We also know that about 50% of merchants don’t know what their data is and don’t know what performs well. They need the ability to be a little bit more data-driven. Do some A/B tests. Start optimizing and then building on this by adding email, SMS, and adding segmented experiences on top of this.

We’re too early for most merchants to think about those. It’s great to talk about them. It’s exciting to see the possibility but when we’re looking at this in practice it’s not even 80%. 80% of merchants think that they need to catch up. They compare themselves to their peers. That’s kind of the flip side of 80% of all drivers thinking they’re better drivers than others. We’re still kind of a little bit early to talk about these things for most merchants. Most merchants have yet to be able to use AI-driven personalized recommendations.

There are purpose-built apps out there and solutions out there, for instance, shoe sizing, especially if you’re a footwear merchant and that is so important to you. This segment-driven personalization solution applies to every merchant out there for a large part. Wrapping up, in regards to this test, learn, optimize, from an optimization standpoint, speaking to the A/B split testing you mentioned, is this happening on-platform? How easy is it on a personalization platform, such as LimeSpot to run a multivariate A/B split test?

It’s very easy on LimeSpot but not everybody uses LimeSpot so they may have to use external tools but we do have the ability to support cloning experience. You have a live experience to clone it and then use a visual editor to make some changes to your experience, assign the weight of these experiences so that the traffic is split to whatever percentage that you desire to.

Also, being able to measure KPIs from there to better understand the performance of these different experiences. It’s something that you can create on LimeSpot in a half hour. If you’re using other tools that may not have this then it becomes a little bit trickier but there are some great tools out there that you can use for split testing.

To bring it even back to what I was saying earlier, at least use product recommendations. That’s 101. Don’t use the platform’s native product recommendation. They’re dumb recommendations for the most part. Look for an AI-driven product recommendation platform, get this working, see some results, and then get into that optimization. Let’s use table stakes. Let’s go from crawl to walk. Let’s get those basic personalization blocks and then optimize on top of that.

On a final note, there’s a mention of curated collections. We talked about segmenting customers and how personalization is down to customers but what is a curated collection? It seems product-driven.

It’s a mix of product category merchandising versus segment. We were talking about affinity earlier, especially around spend. If we have a low spender and a high spender or maybe affinity in terms of you like Nike or Adidas. Collection DC if they go into best sellers for example. You know that I’m a high spender Nike lover should be different if I’m a low spender. The curated collection allows merchants to dynamically present different collections based on the segment that a specific customer would be in.

Best seller, if we take a different type of segments of the contextual segment based on geolocation. If I am in the northern hemisphere versus somebody in the southern hemisphere where winter and summer are different, bestsellers. If your product line allows you to present winter and summer products, then it should be contextual with their geolocation as well. The curate collection is a means of aligning collections with a segment expectation of what new arrivals, best sellers, and all these types of collections are for.

Do you utilize the RFM analysis or do you go more advanced?

It is up to each merchant. We’re kind of back to this idea of Walk, Crawl. A curated collection is a great first step for a first segment experience. it doesn’t require a lot of output. If you think about segmenting experience where you may change navigation, offers, images, and copy all this is intensive in terms of input. You need a team to create all this curated collection.

It typically tends to be the first kind of low-hanging fruit so what we see is typically even something simpler than this. Being able to sort the collection by high to low or being able to present a subset of collections. It’s probably even a little bit more basic than that because it’s the entry point to segmented experiences.

Francis, we could go on and on and on. I am definitely better educated on personalization with this conversation. It’s been a pleasure having you. For those who want to find out more about LimeSpot, head over to LimeSpot.com. Francis, Are you active on any social media channels? If yes, which ones?

Mostly LinkedIn is where I hang out. I’m an old white guy who hasn’t got on to the whole social media craze. Where we’re at with LimeSpot but in general, we’re across social media. We have great information for those merchants who are starting but those who are also a little bit advanced and much more mature in their personalization journey.

I would certainly encourage you to look for LimeSpot on social media. One of the things that I would like to also invite readers to do is to take the survey for themselves and be able to understand where they are in their journey to personalization when we’re comparing this to what we know of merchants in general.

Amazing. Francis, it’s been a pleasure having you on the 2x eCommerce podcast.

Thanks so much, Kunle. It’s great to be here.