Arlen: Welcome to the eCommerce marketing podcast, everyone, I am your host Arlen Robinson, and today we have a very special guest, Shahram adver, who is the founder of an exciting e-commerce personalization platform, DataCue. Previously, he was a data scientist at TripAdvisor and a consultant at IBM.
Shahram: And they had to be on it. And thank you for having me.
Arlen: Not a problem. I’m super excited to talk to you today via a particular area that we haven’t really discussed much on the podcast. E-commerce, personalization and how you can accomplish that. So I’m super excited to dive into that, which I know is a subject that is dear to your heart and you’re digging in the trenches with e-commerce personalization every day.
So super excited to see what you can bring to the table with regards to that. But, uh, you know, before we get into all of that, why don’t you tell us a little bit about your background and specifically how you got into what you’re doing today.
Shahram: I used to work at TripAdvisor as a data scientist. And one of the things I noticed was just how much, you know, well, always talking about this inequality in all kinds of spaces.
And I found that there’s massive inequality in how people understand how to use data. So you know, you’ve got a really big companies like the one I worked at, and you know, you’ve got Google and YouTube and all the rest, and they’re just so good at using machine learning, AI and all kinds of new buzz words really on a, an improving their business.
But. Oh, you know, if you look at the rest of us, so to speak, not so much. I think that usually just buzzwords. And I would be very surprised if one out of 10 people can really tell you what machine learning does and how it can actually help you. And that’s a problem, right? Because we’re kind of in this stage where you’ve got these massive giants and they’re just getting bigger and bigger.
And then like, what about everybody else? And you’re kind of seeing people kind of getting the stuff together a little bit. But what I, for me starting with Derek Q was I wanted to build a software that could. Be like a black box, which would do the machine learning and all that fancy stuff for you, but you don’t really have to understand what’s going on.
You should just be able to enjoy the benefits. And the thing that, these kinds of topics is that you can apply it in so many different ways. So the one thing that really struck out for me was this idea of personalization. And we’ll get into that a little bit more as we go on. But you know, just very simply, it’s like when you go to YouTube or Netflix or any of these really, you know, high end websites.
You probably don’t even think about it, but you know, that’s how the website is so intelligent. It looks at it, knows you, it knows what you’ve been doing and it’s trying to be more helpful. It shows you like Netflix would show you more videos that it thinks that you are probably going to like, same thing with YouTube and you know, if you look at the average website, you know, I’ve got some stats, like more than 70% of the websites don’t even do basic things like that when everyone on marketing stocking about being customer centric.
But are we really doing that? I don’t think so.
Arlen: Yeah, definitely. And um, you’re totally right. Like all of these companies are really going that route of trying to understand their customer and then just provide them what they think is going to be products or services or content that they think that their customer will want to see or consume.
And that’s really the trend and I think really comes down to today. There really is no more kind of a one size fits all for everyone. You can’t just kind of blanket your efforts and it just expect everyone to be able to consume it the same way and for it to be appropriate for everyone. So I think it is definitely required.
For at least a any e-commerce business to at least assess personalization. And so these days, you know, we’re at 2019 we’re just about at the end of the year. Now we’re getting ready to go into 2020 which is, which is hard to believe. You know, when you, whenever you say 2020 you’re like, man, you know, I remember back in
Shahram: the day.
Arlen: Yeah, exactly. When I remember back in the day, I was watching all these scifi movies about how things would be, you know, you just couldn’t imagine that date. But you know, here we are where a, you know, I’m going to say 20 Twain. So, um, for all of the, you know, eCommerce business owners listening and e-commerce marketers that are listening, why is personalization so important these days?
And what do you think.
Shahram: So let’s get a few definitions out of the way. I think. Let’s first talk about when personalization is not important and that I think finding out a little bit sure. Personalization is the idea that when someone comes to your website, you show them different content based on their interest.
Right? Right. And so think about that for a second. And if you imagine that you’re an eCommerce owner and you’ve got five products. You don’t need personalization makes no sense. You can just put all five products on your homepage and you’re done, but it becomes more of a problem when you’ve got say, a thousand products and now suddenly you notice that you’ve got different.
Kind of customer behaviors all coming to your same site. Like a, an example I like to use is, you know, imagine a store selling baby products. So you might have quite a few products, but just think about the customers coming in. You’ve probably got expectant parents and they’ve got totally different kind of needs.
And then you’ve got maybe friends who are going to a baby shower and then they’ve got totally different needs. And then you’ve got maybe people who are already parents and then depends on the age of that child. So. You’ve got X products, but you’ve got so many different people coming for different things.
So that’s when personalization becomes really important. But if you think about what eCommerce companies are doing right now, they’re spending all that time on traffic acquisition. If you ask any e-commerce person, they’re already probably pretty good at Google ads or Facebook ads or what Instagram and what have you.
But the question I always have is like, okay, you’re spending so much money and so much time getting people to come to your site. But when they arrive, you’re kind of depending on lack a little bit. Right. And so, you know, like, shouldn’t you, you know, I was listening to a broadcast you did previously about customer retention, and you know, that’s so true.
Like, you know, it’s so much cheaper to keep the customer that you, you work so hard to get where it says keeping on getting new ones. So if you know that, let’s say you had an ad for, I don’t know, let’s say, um. You know, some baby shower gift, let’s say a stroller, I don’t know. Say you’re very rich and you’d like bank strollers for your friends.
Use slap a Google ad and then someone clicks on it. And then you kind of want to make sure that the in store experience is really geared towards people who are looking for products like that. Right? So, and it’s astounding how little that actually happens right now. So you’re kind of, what you really want to do when someone comes to your website is to be helpful, not obnoxiously, but essentially help them find that spot to the checkout process.
And that’s when you’re actually going to start converting more. And that’s basically what personalization does. We can jump into it a little bit more detail, but hopefully that gives you a brief overview of like what it does and why you should do it.
Arlen: That totally makes sense. It’s kind of the same and the same coin that Google in most search engines try to follow.
Their whole ultimate goal is to provide the users or searchers the most relevant answers to their particular questions, and so they’re trying to. Customize their response based on what’s asked and, and provide them with the best possible answer. And so, yeah, I think as a business owner and an eCommerce site owner, you have to look at the fact that, like you said, when somebody comes to the site, what is going to be the smoothest path to check out for that particular person?
And, um, you don’t want to, you have to do to make them accomplishment. So yeah, you’re, you’re, I think you definitely nailed it on that.
Shahram: One of the quotes I like a lot is that, you know, when a world of information overload, so you know when you’re running a site or any kind of. Media, the problem, the thing you should not be doing is adding to the information overload.
So the issue which you need to focus on is curation, not adding more stuff. So you kind of need to make it easier for the person who’s coming to make that decision, not just him even more information and just make the Doosan and even harder, if that makes sense.
Arlen: Yeah, yeah, definitely. Totally makes sense.
So I think one of the things that. Listeners would probably appreciate it. Since you’ve been really kind of in the trenches with the personalization, what has your research uncovered with regards to any statistics on, you know, personalization in general, sites that do personalization versus sites that don’t.
Shahram: So that’s always a tough one because honestly, it really depends on the customer. So like I said, let’s assume that you’ve got the basics right. So I’m assuming that you’ve got enough products and you’ve already got some revenue coming in. So if anyone’s listening today and you just started almost side.
And maybe you’re making less than a thousand dollars $2,000 a month. I would say don’t touch this topic. It’s not something you should look at. You should be focusing a lot more on traffic acquisition and this kind of stuff, but now let’s say you’ve built a business, people are coming, people are converting, and now you’re looking for that sort of next step.
Then the idea is to just look at. Things like your conversion rate. So you know, one of the stats which I like to talk about is things like three and four customers that we know, just they, they’ll leave your website in the first 30 seconds if they don’t see what they want. No, that’s obviously just an average statistic.
It’s, it depends a lot on different websites. But if you look at an average conversion rate of e-commerce, that’s 1% right? So why does that 1% really mean. It means that out of all a hundred people coming to your site, only one of them is going to spend money on your store. The other 99 a passive that they’re just going to come and they’re probably going to bounce.
So my point to the eCommerce owner is that right? You’ve got your conversion funnel, that’s great, but now let’s focus on the other 99 that’s where the value is, right? Like how do we make things easy? You’re never going to make all that in and convert, but if you can increase your performance there, but I don’t know, a 10 20% that’s a big jump.
So then the question is. What do you do to make them more interested? And one of the things I think the listeners here should be quite familiar with is the idea of a balance rate. I’ll just define that by everybody just in case, but the answer is essentially how many people who are coming to your store who don’t take any action and just leave.
And so if you take it as a analogy of a physical store, your first job. With people walking outside your store is to welcome them in and hope they come in. Right. So if you think about the home page, that’s kind of what you want to do as well. You want to show them something which you think would be attractive or interesting enough to get the first click, because we know statistically that once they’ve clicked and they’ve started looking at product, now they engage with your store, and now the chances of this person actually making it part of this, that much higher than someone who just came to their own page and just decided, nah, and closed it.
Hopefully that makes sense. Yeah,
Arlen: totally. And that’s the key thing is getting people right when they come to the site, because like you said, that bounce rate is something that really can, can make a break or business. You know, if people come to a particular site and they don’t, maybe they’re looking for a particular item, or maybe they just happen to stumble across your site.
In either case, if they’re not pulled or compelled to go further and they just bounce away, you know, nine times out of 10 they’re not gonna come back. So that initial. Stop at the site is, is really critical and I think that’s really where the personalization comes into play and doing things to try to compel people to get all the way through checkout.
Now I know with, you know, with what you’re doing in a data queue, you guys are really doing a lot to help eCommerce businesses come up with personalization strategies and plans and things like that. What are some typical principles that people should abide by and that that you guys utilize internally.
Shahram: So the thing with personalization, I think, you know, just following on from the previous question, but I’m assuming some listeners might be thinking, hang on, how do I personalize it? This person’s only coming for the first time. Like, how do I show them something attractive because I have no data on this person.
Right. So the answer to that is the first thing you need to be doing is to actually, even if you don’t use a tool as a principal, you should know maybe it’s per week or per month or per season, what your best sellers are. What are the products that are most likely to, you know, engage your users. So that’s, I think the number one thing, which I’d say it just, you know, make sure your own beds has things which are relevant and are most likely to move.
And you should know this data by now because you’ve. You know you’ve been in business, so that’s the first thing you should do. And that’s what I do like as this, but you automatically, so it’ll just look at the orders and all the customer behaviors. So the first time visitor, every day, it should be updating your homepage and all the other pages with recommendations, which are really tailored to the best sellers.
But then the next principle. Is okay. So once they actually take an action, you should actually act on it, right? So if you think of it like a physical store, as you know, some sales reps probably gonna walk up to you and say, okay, you know what you’re looking for today? How can I help you? And maybe you say, well, I’m kind of looking cause I’m jeans.
And then what are they going to do? They’re just going to walk you to wear the jeans and then they’re going to shield you from everything else that’s in the store and then say, . Just focus your attention on what you’ve actually been asking or maybe think about it. That’s kind of what users on your website are doing right now.
Whenever they’re typing a search on your search box and whenever they’re clicking on a product, they actually telling you the website what they are interested in right now. Right? So the next principle is actually to take that. Feedback into account and then try to dynamically change the experience based on that.
So just to be very specific, so on the homepage, what people do typically is you’d have banners at the top, right, just to showcase your best categories, and then you’d have some products. So what we do is when they click on a product, we can associate that with a battery. So you could upload, say, 20 banners or all kinds of product categories, but we would pick the most relevant three.
For instance, you can choose, it can be three or five or whatever, whatever you want. And so if you’ve got a January furniture store. And now this person flakes on a chair for the kitchen. The moment they come back, we should be showing more kitchen or chair related products on their own page, just so that you know they’re going to engage.
Then in the product page when they’re actually looking at something, now you’ve got their attention. You also need to be able to show a different sort of, let’s say related items or similar items or things like that. Just because people don’t really make a decision immediately. Like they’ll look up, you know, they are always like to compare.
So if they are looking at these blue jeans, they probably want to see it in different color or like, uh, some, some other styles. So that’s really important. So the idea is that. And your eCommerce site. Each page has a very specific function, so the homepage is supposed to really introduce what you do and bring them into, say, a category page and a category page.
Hopefully they’ll bring, come to a product page and then eventually they’ll add something to the cart, so they’re on the cart page. So in each of these pages, there are very specific things that you can do, just like what I mentioned for the home page of the product page that would update the content based on what they’d been doing previously.
So that sort of very gently nudging them towards, I mean, by being helpful, not in any like creepy, obnoxious way with a popup. You’re just really literally showing them things, which you think might be interesting. Just like a, you know, a real life sales rep would do.
Arlen: All right. Does Sharon? Yeah, that totally makes sense.
Um, everything that you said in one thing actually resonated with me was the example that you’ve mentioned as far as somebody going into a store and saying they actually approach a salesperson. A salesperson says, what are you looking for? And then they say, I’m looking for jeans. And then they lead them to guide them guidance.
That Jean section, which is. Which makes total sense cause that’s what they’re looking for. And you’re guiding that person to what they’re looking for. And the only thing that I, I don’t know if this is the downside, but you can kind of enlighten me on it when I’m thinking of recommendations, not only let’s say an in store recommendation, but maybe a recommendation from, let’s say, one of the online streaming services such as Netflix.
You know, they do a great job of personalizing their con, the content based on your viewing preferences. And I think that’s excellent. But I just wonder, does personalization go too far? Because it assumes that someone just based on the things that they’ve watched, they, it’s assuming that they are just liking that particular genre, but maybe they’ve been into a, let’s say, a whole Saifai kick for the past
Shahram: couple months.
Arlen: you know, there may, maybe they’re the type of person that just likes to switch it up, but if Netflix is going to continually keep feeding them Saifai recommendations. That may not necessarily be the best experience. So it’s almost like you have to, there’s like a fine line between personalization and being too targeted, if that
Shahram: makes sense.
No, absolutely. And this is a actually a very relevant discussion when you do recommendation algorithm. So no data scientist, but test all wood over in mathematics that’s got a local minimize. It’s basically saying, okay, just because this person watched a little bit of that Saifai. As a mean, that you should be recommending completely OSI stuff.
Right. So, you know, just to give you a concrete example, basically I’ve got a set of five recommendations and you in watching PSI fi, I’ll probably throw two scifi titles in there, and then I might throw maybe, I don’t know, romantic comedy and like two other things just to test the water a little bit. And that’s really a algorithms work.
So. You should never do five side five items. So you know, if they click on genes, showing five genes is generally not a good idea. But one thing that’s very interesting, there’s something actually is on recess. It’s Spotify. Did they tell you that recommendations work very well when you tell the user why you’re showing them what you’re showing.
So, for instance, if I show you a recommendation for a song and I say, Oh, I made this playlist for you because you listen to this, this, and this, you notice that on Netflix or Spotify or what have you. They’re never really very strong on this one thing, but they say you should do. They just give you a whole bunch of options, right.
And the idea is always to keep those options. Like you keep it wide, but then you try to make those options as relevant as possible. It’s the same reason why Google would never. Just take you to the first result. They are constantly learning because on the first page there’s literally 15 results, all slightly different.
And that’s really like where people like us come in because that’s really my job. So, you know, if we are showing terrible recommendations, then no one’s going to be clicking on them. And then you know, the tool is best for you. Use this. Like you won’t be able to see any improvement. So, you know, that’s kind of on people like us to actually make that recognition is better than the possible.
Arlen: It totally makes sense. I see how you have to mix it up. You can’t be so personalized where you’re just assuming that you’re keeping people down one particular path, but if you’ve got other products, other content, other services, you still have to kind of squeeze in those other things that could possibly be a venture.
Shahram: I think one really important point to add. Is that this idea? And if you ever hear people saying this, you can dismiss it outright. You cannot predict what people want to do. It is impossible. You just can’t. So, you know, like for instance, I’m 33 years old, I’ve got a little daughter. If I went to a website and I’ve been looking for work clothes for me, the algorithms just going to assume certain things, but tomorrow I could be looking for something for my daughter and then my behavior is going to be completely different.
So that’s why for us, we put a lot of emphasis on the idea of being real time. And that’s the idea. Like you know, time to really mimic a sales person because you could have just go to the physical store experience. Like tomorrow when I walk in and I say, I know Ashley, I don’t want jeans to them. Really looking for something for my daughter.
Right. Then that’s where you know, we kick in. So if you just bring it back to the eCommerce side, it doesn’t mean that like this recommendation system replaces everything you have. So if I show them something which isn’t as good, they click on something else, or they go to your navigation. I, they do a search.
I’m already learning banking that, okay, so this person right now is looking for product. Type X, so now I’ve got that signal. So beyond anything else, I think the idea of just looking at what the person’s doing right now is a far more button signal than anything else that you can predict.
Arlen: Totally makes sense.
Now, you know, I always like to kind of leave with a
Shahram: good takeaway for our
Arlen: listeners and for anybody that really hasn’t jumped into personalization and wants to get started. What are some ways into some examples of how they can get started either. Using a solution like beta Q or not even using a solution or any other type of platform.
What are some ways to really just get started into personalization for the eCommerce
Shahram: site? So if you thinking about it and you do, and you don’t want to think about doing it with the solution, you can just go ahead and think about the kind of customers who come to your site and try to group them into segments.
And you can actually make different landing pages for each segment. So if you think about what I talked about with the baby store right. Maybe store.com/baby shower, baby store.com/toddler baby. You know, and you can already, I mean, as a store on a, if you don’t have too many products, you can already just imagine what kind of things each of these segments, one, right?
And then you can just really align that with the traffic acquisition that you’re doing. So instead of sending your Google ad traffic or Facebook ad traffic to your.com you could send them to that sub page, which is very relevant to this kind of. This audience that you’re really going for, but when you start doing this more and you realize that you’ve actually got a lot of products and you’re spending a lot of time updating these things because some of them become sprinter and there’s lots of changes that are happening, that’s when I think when you’re actually feeling the pain, then you realize that, okay, I sedation.
That’s the my panko sense because for us the idea is that the build this sedation, that’s something that’s plug and play that you actually, we shouldn’t be adding one more thing for you to worry about because we know you’re already spending so much time. Thinking about Facebook ads, Google ads, and traffic acquisition.
The idea is that when they come to the site, that should just work automatically for each president, right? So there’s a bunch of personalization solutions out there. You can have a look at them, and I think they all have very different views on what they should be doing. For us, we want to be like, for us, the focus is really on simplicity.
So. When you install the app. So on Shopify, uh, any major platform, we’ve already got an integration. Uh, you, so, you know, we pride us as you can get started in under 15 minutes. Okay. But once you’re there, what I really want to focus on is, okay, so the first thing is we changed content for the visitors who are coming in so that they have a better experience.
Well, one thing which we’ve been hearing a lot from customers is, okay, but you’ve got such great data on what people are doing on site. Can you give us advice on what we should be doing more? That’s an area which we’re really investing a lot of time on. So already what we do is we’ll tell you like what product images, it could be the most successful.
We can tell you if you upload 20 banners, when you look at something that Google ads, it just tells you numbers, like how many page views and things like that. You can’t really do much information with that. We can’t really do much with the information, whereas for us, we can actually tell you, look, this Panasonic’s, you should take it down.
You should replace this. So barely working a lot more now on like adding more recommendations to the stoner just to prove that conversion. Okay.
Arlen: Well great. Yeah, that’s awesome. That’s some really good practical advice and appreciate. Appreciate that. Shall Roman definitely appreciate you. Have you being on the podcast.
Cause like I said earlier, we hadn’t really talked too much about e-commerce personalization before and you definitely gave us some concrete ways that we can be successful with that if we are working with any commerce business. And so, yeah, that’s awesome. And so, yeah, just lastly, you know, if our listeners want to pick your brain anymore regarding email personalization, what’s the best way for them to get in touch with?
Arlen: All right. That sounds awesome. Well, thanks again, Shahram, for being on the eCommerce marketing podcast.
Shahram: Sure. Thanks, Arlen.
Co-Founder of Datacue