Checkout optimisation tips from a specialist | Gabriela Florea (Verifone)

As CRO Manager at Verifone, Gabriela and her team specialise in optimising checkout flows. They do this for e-commerce merchants who use Verifone's Hosted Checkout product. Those insights are then used internally by the Product team to constantly improve and innovate the product itself. 

In this episode, Gabriela talks about team structure, competitor research, iterating on negative A/B tests and tips for checkout optimisation.

Transcript

0:00

The checkout is the moment of truth on any e-commerce website. Regardless of what happened higher up in the funnel, this is where that decision is made that we all care about. Am I going to convert or not? Today's guest is knee-deep in that world. That's all she does with her team is checkout optimization. I don't know anyone else who's got as much and as deep experience with squeezing more out of a checkout flow as she does. Gabby, welcome. I've got loads of questions for you. But before we get there, tell us about Verifone. Verifone is a fintech leader. It provides end-to-end payment and commerce solutions to world's best known retail brands, financial institutions, and I think over 600,000 merchants. It acquired a company called to checkout approximately two years ago and watched to checkout this. It provides global payment, subscription billing, merchandising, compliance risk, all that a merchant needs to be selling successfully globally. And the to checkout product is the actual shopping cart. So we hosted the shopping cart, which makes sense for us to do anything sterile related on it. That sounds like a fairly unique operating model, I think, where you've got this product that you lease or sell, in this case, the checkout, and embedded in that product is a CRO function, where you're constantly optimizing that product for the client. To do that effectively, how are you structured? Where does CRO sit in Verifone? Is it with engineering or product or elsewhere? The CRO team here in this company is currently placed in the product organization. But a few years ago, it used to be with the sales teams. But now, I think for us and for what we're doing and how we're aligned, it makes more sense to be under product because we not only optimize the checkout, any insights that we have from our A-B tests, from our bake-offs, split tests, multivariate and so on, we take those insights and we go to product to make sure they are incorporated so that the product gets better. And while I say it also makes sense for us to sit in the product org, is because we do a lot of competitor research and we're only focusing on CRO aspects. Like we are not going through product lenses, CRO lenses only. And as I said, we take everything we find and we go to product and we say, hey, we would like to create a new shopping cart template because there are new trends on the market. We looked at other payment processors. They're doing this and this, which we might wanna do as well so that we grow the conversion rate and as of course, overall revenue for the platform. So these insights and findings from your daily CRO work with clients, you play those back to your product team internally and then use that to improve your own product, to develop, to innovate, to evolve. Do you change templates? Give us an example of how you use those insights then to improve the base product. There are a few aspects. First one, we're talking about just improving the product. It's a shopping cart in itself, like a digit separator or we're finding bugs or anything that could make the product better from a technical perspective. But then, because of all these insights that we're getting from our AB test and from this big number of merchants we're working with, we get to come up with innovating the product in itself. And by that, I mean, we can create, as I said, new shopping cart templates, but not just for the entire platform. We get to create these if let's say, hey, we wanted to focus on a specific region, let's come up with a shopping cart template just for that. And we go to different SaaS businesses and we just try to see what they're coming up with in terms of checkouts. So we get inspiration from that. We also do these exercises where we rate from one to five how well they're doing compared to us. So that helps with us innovating our carts. So everyone gathers and we give points. The highest ranked from the competitor research, we locate, let's see, what can we take from this? What can we apply with our current situation? And then we AB test it. We don't go live with it. It doesn't mean that it's gonna work for us. Also, we don't know how it's currently performing for them. Yeah, so what you're not doing is blindly copying your competitors, which I think is a trap that a lot of people out there are falling into. You've got a reasonably sophisticated process where you get together as a team, assess what others in the same space are doing, then rank them and the top ranking ones, you would draw inspiration from what they're doing, but always test those ideas. So there's perhaps a good practical tip for people to take a more formal approach to competitor research. Now, talking of tips, this is where I want to get to. Oh my God. With all that, you've got to expect this. I mean, as somebody who's knee deep in checkout flows all day long, can you think of some top tips, taking everything that you know about checkouts and all the tests that have won and the tests that haven't won and the surprises that you had? Yeah, so we have a really good success rate. I think it's 6.5. So out of 10 tests, seven are winners in our case. That's, I think that's above the average in this industry. Now, I think the reason for that is, of course, we're experts, we like to invest and put our efforts into generating good results. But I think it's just because we're super focused on the checkout. So we kind of know the ins and outs of it. Tips for that. If you want a big growth, like if you want to see double digits for the conversion rate increase, you have to make a significant impact because you only get a small percentage of the shoppers that actually land on your footprint. And by the time they land on the checkout, you know, they're kind of convinced that they want to make a purchase. So we have a big margin of success with these. But some things that you should test on the checkout only if you want to see big growth would be change the way the cart looks. And we've done this. We used to have a one column shopping cart template and we tested a three column shopping cart template above the fold, super compact, took all our insights put into it and it was two digits increases in the conversion rate, which is a lot because it directly impacts the revenue. There's no discussion about it. Like you're not going to struggle with attribution. Some other things that work for us in the checkout, social proofing and here there are nuances but social proofing usually works and it creates an increase in both conversion rate and RPV. You can try to take, you know, some reviews from some famous blogs, put it in there exactly in the checkout. You can try to take some highlights from how your company's performing. You know, you can try, you can put some logos in there with, you know, partners or people that are praising you. Another tip would be security. Icons. These work, you're not going to see a big impact but it's going to positively impact the finishing of your payment. You know, if someone is doubtful about the checkout, it's like, hey, I don't know this merchant that well. If they see the security logos, they are more open to pressing the purchase button. We also saw that having the 30 day or 60 day money pack guarantee logo works pretty well for them. Removing unnecessary fields. If, let's say, I don't know, only 5% of your orders are placed by companies, maybe don't display, you know, the company radio button or whatever kind of display you have there. Yeah, not giving away all of our secrets. Yeah, the one that I'm surprised by there was the security icons. I'm surprised that it's 2023 and that, you know, there's still this anxiety about paying online and how secure it is. Do you see a difference there in different geographic regions that shoppers in certain countries or certain regions have different anxieties or preferences compared to others? Yes, so we are looking at usually Americas, EMEA, which is Europe and Middle Eastern Africa. And then we look at APEC, so Asia, Australia. There are different tendencies and preferences. I wouldn't say they are exactly connected to the security icons. We've noticed other preferences in terms of the car display, so from our experience, APEC shoppers are not as, you know, they just like the view and the design of the car to be different than what we're seeing with Europe and Americas. Whether we are trying to make things compact or simple or, you know, just minimalistic as possible, we're seeing a different tendency in APEC. They like more imagery in the checkout. They actually like seeing pop-ups and stuff like that. So I would say that's the thing we noticed based on regions. You spoke about the minimalist design versus the design preferences. And when you mentioned that list of things that you would test in a shopping cart, do you find that as a general rule, the simpler it is, the better it performs? And so you've got to think about things to strip out from the page rather than things to put on the page. So for us, cognitive load is something super important based on our A-B tests, right? So it's not my opinion. It does work better when it's more compact, when you have less, you know, checkboxes. And definitely if you reduce the number of fields, like even put the first name and last name together in one field only, that will impact the conversion rate. It seems that even with an additional field, some people will drop. They lose interest. Whether they were not sure about purchasing the product, if you put in more fields there, they're probably not going to finish the purchase. Yeah, and this is one we often hear and we know it. And yet, you know, you still go through these forms and see all these different fields. And it's hard to kind of justify the presence of that field. Now, right now, like with PayPal, Express, and with, you know, some other APMs, we're actually testing a flow in which we're trying to provide to the shopper a super Express checkout flow. So you just land on the cart, you don't have to feel anything. Just click on PayPal, Express, and you get to make the payment. Are there certain tactics that sometimes work and other times don't? You get different results with the same kind of tactical implementation. Do you find that? Yes. So here's what's tricky about us. We have really big merchants working with us. So there are, you know, different expectations, but if we don't have a winner with what we suggested or, you know, what we tested, there are a couple of things we're doing. So first we start looking, of course, at heat maps, screen recordings, then we're trying to break things down by country devices. We're trying to find out why a hypothesis that worked for, let's say, I don't know, a couple of other regions or some other merchants, why this is not working with this one in particular. Maybe there's a bug. Maybe people in, I don't know, Poland, they're responding differently. We try to change that. And what we do is we wrap up the test and we come up with a second iteration. And this way, in the second iteration, we say, hey, let's try adding this, this, and that to see if this will change the results from the first test. And with this scenario, we usually get a winner. And I know, I know people like in the serial industry, they're like, you're not supposed to aim at getting a winner. But as I said, because we're working with big merchants and some of the AB tests that we run are of course paid services. There are certain expectations. So we're trying to meet the merchant halfway. Most merchants, most clients expect winners from the AB tests that we run. Well, I think that's true for CRO in general. You know, it's experimentation, I think, fundamentally, but with an experiment, you're trying to prove yourself wrong. You're not trying to prove yourself right. And most often, you are going to be wrong. But CRO is about growth. And it's about finding those winners. And equally, it's about finding the negative ones, which you can then weed out and make sure you don't hurt your sales. But you get a negative experiment. You don't stop there. It's a chance to learn from it. And what I've seen in experience is often your biggest winners come out of negative tests. And the reason for that is that there is no other way. Often you could have found that insight that you needed to get the win other than running a negative test. Now, I'm going to turn that question around. Are there tactics or ideas that always win or almost always win? It is a no-brainer, and you've got data to prove it. I think I need to think about these tactics that almost always win. Well, that's an answer in itself, Gabby, isn't it? If you can't think of one, then it's kind of what it is. No, no, I think you're right. Because when we start with a hypothesis, right, and it wins one time, we don't go to other merchants and be like, hey, this one, you have to implement it. No, we test it. Like with the three column template that I mentioned to you that generated such big growths, we actually tested it for two years and a half before we decided to go to product and be like, hey, this needs to be a part of our shopping cart template gallery accessible to all our merchants so that they can just swap to it whenever they want to do it. Let's take social proofing, which is one you mentioned there. And it is one I've seen it go both ways, right? So I've seen social proofing make no difference, sometimes even hurting it. In the checkout, which is such a crucial juncture, right, this is where the decision is going to be made or not. And you spoke about cognitive load and the importance of just getting out of the user's way and allow them to do their thing. Is there a risk that social proofing by adding more elements could work against you? Have you ever seen that? Right, when we first tested some social proof hypotheses that we had, they did not become wins. So we were like, why is this not winning? Let's try to come up with some other ideas. And I think the main insight that I have on social proofing is keep it super short. Take, I would say, three things that you want to brag about in the checkout, right? So we're not talking about the discovery stage. The shopper is already in the checkout. If they have certain insecurities, what would seal the deal? And for that, come up, as I said, three things. Maybe number of shoppers that are happy with you, a number of purchases that were done in the last year. And it serves to the purpose of the indecisive shopper. We only need that one to say, I don't know what I want to do. Should I purchase this? Well, a million clients are satisfied with this. So maybe I should. This is not false advertising at that point. And so you say you've got to align that with the insecurities of your shoppers. How do you know what those insecurities are? Are these effectively hypotheses? Like for us, the way we work is we have a lot of meetings with our clients and they also provide a lot of insights. They're like, hey, we notice people are not using the auto renewal option as much as they did in the past. Or hey, we're seeing that people are not finishing the orders and they're stopping just after they clicked and they filled in all the information. So strange things like that, insight that they have and they bring to us, we're like, hmm, what can we do to tackle that? A lot of the problems that we see in a checkout may be set up earlier in the funnel, right? It takes absolutely no commitment to click on add to cart. And even clicking on checkout, that's not a sign of commitment. Putting in your credit card details, that's a sign of commitment. And I've been in myself as a consumer where the checkout itself is fine. There's nothing wrong with it. But you have certain lingering questions and anxieties and fears and concerns that haven't been answered that should have perhaps been addressed higher up the funnel. To what extent do you think there is this interplay and that your checkout funnel success hinges on higher up the funnel limitation? I think you're making a fair point. They are like super interconnected, which is why, as I said, we're focusing on expanding the stereo services outside of the checkout. Cause for us, we're kind of limited, right? Merchants come to us and they're like, hey, we need, we want our revenue to grow. And we're like, but we cannot impact what's happening upstream. So we're like, our hands are tied. There's only so much of what we can do with the traffic that's being sent to the checkout. But there are some things that we're looking at. We're looking at traffic sources, especially throughout busy periods or campaigns that are being run. And we say, hey, the conversion rate is going down because you're running this, this and this. You need to stop them or adjust them. So we're touching upon some things, but we're not, we cannot influence what's happening upstream. With certain clients that we have, and they're big enough that they have their own stereo teams, we're kind of in sync. So with these clients, we get to say, hey, what are you guys testing upstream? What can happen there so that it goes even better for the shoppers that are in the checkout. And the results are better. Kavi, I'm sure you've been asked by clients about ROI. Oh my God. How do you, how do you quantify the value of your tests and CRO and optimization? The way we calculate the ROI, like specifically to the CRO team, it's a bit different, I think, than what other CRO teams are doing just because of the nature of this company, right? So we're processing payments. We're taking commissions out of these payments. I'll just give you like a super simple example of how we do it. And I hope it makes sense for people. So let's say we have a client and their monthly revenue on average is 100K. And we had an AB test at one and the conversion rate increase is of 8%, right? So what people usually do, they take 8% times 100K. Out of these $8,000, we only think about the commission. Let's say we have a commission of 5%. We take the 5% out of the 8,000 that I mentioned and we end up with, I think that's $400. And that's the actual money we make on a monthly basis with that test. That's the only number we're looking at. And the $400 that we generated as additional or extra revenue for the company, we multiply that by either times six or times 12, depending on how we're looking, if we're looking at a six month forecast or if we're looking at the entire year. There's so many more questions I have for you. It's been great to talk to you. Thank you for joining us and sharing your insights. Thank you so much, Johan.