428: Counterintuitive Strategies (Which Make No Sense) That Can Grow Your Sales With Kurt Elster

00:00
You’re listening to the My Wife, Quit or Job podcast, the place where I bring on successful bootstrap business owners and delve deeply into the strategies they use to grow their businesses. Today I have my long time friend, Elster on the show and Kurt is the host of the unofficial Shopify podcast and a Shopify store consultant. And today we’re gonna talk about some counterintuitive learnings from running an online store. And it’s funny, new and even seasoned entrepreneurs often just blindly copy other successful stores. But as we’re about to find out, this can get you into a lot of trouble.

00:29
Enjoy this episode. Now, before we begin, I want to thank Postscript for sponsoring this episode. Postscript is my SMS or text messaging provider that I use for e-commerce and it’s crushing it for me. I never thought that people would want marketing text messages, but it works. In fact, my tiny SMS list is performing on par with my email list, which is easily 10x bigger. Anyway, Postscript specializes in text message marketing for e-commerce and you can segment your audience just like email. It’s an inexpensive solution, converts like crazy, and you can try it for free over at postscript.io slash d.

00:59
That’s P-O-S-T-S-U-I-P-T.I-O slash Steve. I also want to thank Klaviyo for sponsoring this episode. Always excited to talk about Klaviyo because they’re the email marketing platform that I use for my eCommerce store and I depend on them for over 30 % of my revenue. Now you’re probably wondering why Klaviyo and not another provider. Well, Klaviyo is the only email platform out there that is specifically built for eCommerce. And here’s why it’s so powerful. Klaviyo can track every single customer who is shopping in your store and exactly what they bought.

01:26
So let’s say I want to send out an email to everyone who purchased a red handkerchief in the last week. Easy. Let’s say I want to set up a special autoresponder sequence to my customers depending on what they bought, piece of cake, and this full revenue tracking on every email sent. Klaviyo is the most powerful email platform that I’ve ever used, and you can try them for free over at klaviyo.com slash my wife. That’s K-L-A-V-I-Y-O dot com slash my wife. And then finally, I want to mention my other podcast that I’ve released with my partner Tony, and unlike this one,

01:53
where I interviewed successful entrepreneurs in e-commerce, the Profitable Audience podcast covers all things related to content creation and building an audience. No topic is off the table and we tell it like how it is in a raw and entertaining way. So be sure to check out the Profitable Audience podcast on your favorite podcast app. Now onto the show.

02:17
Welcome to the My Wife, Quitter, Job podcast. Today I’m thrilled to have Kirk Elster back on the show. And even though Kirk and I have kept in touch, he actually hasn’t been back on the show since episode 118, which was many years ago. So I knew I had to have him back. He runs ethercycle.com where he helps private label sellers launch their own e-commerce websites. He is a Shopify platform expert who helps Shopify users improve their sales. And he is also the host of the popular Unofficial Shopify podcast.

02:46
In this episode, we’re actually gonna talk about counterintuitive learnings that he’s made from running split tests on his client’s site. And when it comes to site design, what we expect to happen doesn’t always happen. And with that, welcome to the show, Kurt. It’s been a long time. Thank you for having me. You know, Kurt, it’s been a while. Anything new that’s up with you since the last time you were on? Like, what are you working on right now that’s exciting? That’s just kind of interesting. I’ve been doing the same thing roughly for 10 years. Helping?

03:16
We were building WordPress sites in 2009. And the only thing that’s changed is like we said, let’s just work exclusively on Shopify and then showed up on your show. We started building apps and I’ve been hosting this podcast the whole time. You know, my own show and that just staying the course, staying steady has tremendously grown things. The big difference now between then and now, a, I have better branding and B I’ve got, uh, I have a team now.

03:45
We were able to hire people full time and part time. It was very helpful. It is helpful as long as you have good people. Like we’ve had a couple of disasters over the years. All right. So I’ve been lucky. I’ve only had one disaster. So we’re talking about split testing today and I have a love hate relationship with it. So for me and I want you to share your experiences also. But for me it takes forever to set up. Then you got to wait weeks depending on how much traffic you have and then

04:15
I would say 80 % of the time the data is inconclusive. So it’s not satisfying. I know you work with a lot of companies and I’m glad you got a lot of successful tests to share with us. So I want to start actually with what your best practices are and what your kind of success rate is for a split test. Okay. Well, so you’re right. Your experience is, is accurate. A lot of people have that experience and I was resistant.

04:41
to split testing. I loved conversion rate optimization, but my experience was like yours. And so I was instead leaning on usability, like let’s do user testing, let’s look at heat maps, let’s do surveys, let’s talk to people and then let’s, look at our analytics and then try and infer update and then go back and check. And that process works. But over time, you know, I started to get a few clients where they had the order velocity,

05:10
where I could run tests and have statistically significant results in two weeks. And the other big change was Google Optimize. Like in the past, you had to mess with these tools that were tough, that were like really very enterprise tools, visual website optimizer, optimizely. Okay, I’m sorry, threw them under the bus. They’re perfectly good tools. They’re also, they are not necessarily for mere mortals. And whereas, and they’re expensive. Google Optimize is free. It is as easy to use as a Google product is.

05:39
which, you know, all right, it’s not like perfect, but it’s definitely easier than those other two. And so the, a lot of the barriers to entry that you described went away for me around two years ago. And I said, all right, I’m going to figure this out. I’m going to get good at this. I know enough to get started and be dangerous. And so I started playing with it. And then I started having the experience you did where like most of the time the test goes, who cares? You know what? That’s a perfectly acceptable answer.

06:08
In which case you go, you know what? It’s preference then. I could decide what to do with this. But what I learned was sometimes it’s because you have the wrong metric chosen for the test. And so it’s nice and optimized. can run multiple objectives. Like it doesn’t have to be conversion rate, which they call transactions. It could be revenue, which that’s going to include average order value. It could be page view. Like, are we just getting them to the next page? That’s suddenly now you can get more significant results more often.

06:38
And the other thing I learned was, you know, the classic split test that we think about is, what color should the buttons be? What’s my end to cart color? And like, of course that’s not a great test. Um, and you end up with that experience where it’s like, nothing mattered. What works better is sometimes you just go, I’m going to try an entirely different template. I’m going to change 10 things about this page, which of these two versions works better. And as long as like, there’s a hypothesis behind it as to like a goal, like

07:06
Does a cleaner version of this work better? Does a more detailed version of this work better? And they’ve got, you know, five to 10 changes between them. Ah, now I’m to start to see those bigger differences where it’s like, yeah, this one is 95 % likely to be better. All right. So let me ask you this then how much traffic and conversions do you need to, run a split test within a reasonable amount of time? Oh, that’s a really good question. You need to be getting

07:37
dozens a day. Hold on, let’s. That’s not that much actually. Well, I’m going to look up a store right now. Okay. I’m going to log into a store. Maybe we have to edit out me waiting. Yeah, that’s cool. So while you’re looking that up. So, so my store gets a lot more than that, but I still found it took a couple of weeks to before like the tool would tell me like I got a semi-conclusive result or the tool told me to just stop the test because it wasn’t going anywhere.

08:07
Sometimes it drags. Yeah. think dozens is the minimum. Okay. Once you’re at like a hundred a day, maybe 150 a day, totally fine. You’re good to go. This is, you’re to be able to run a test. Now the catch is even if you can hit statistical significance quickly, you do realistically need to be running it for 14 days to see a not skewed result. And like here an example as to why is people get paid, typically get paid every other Friday.

08:36
And so you want that like full pace cycle in there. And so two weeks really tends to be the minimum for this. We can think of this as somewhat accurate, a proper data-driven decision. Okay. That makes sense. Cause I remember the first time when I didn’t know what I was doing and arguably I still don’t know what I’m doing, but I would get like a S like a significant result within like the first couple of days and just want to just call the test right away. then the tool would tell me not. Yeah.

09:05
I’ll see that on Twitter. Like I followed a guy who claimed to be like, you know, a stats guy, like a finance econ guy who knew stats, who like had gone to Ecom and he was calling tests in two, three days. So I knew right away, I’m like, ah, this isn’t significant. This isn’t right. And this guy’s full of it. Like that’s just one of those real easy early mistakes. And I did it too. The second question I want to ask you is what like

09:32
Can you just kind of describe for the listeners how easy it is to set up now? You mentioned Google is easy to set up. Can you just kind of describe what’s involved at a high level? Okay, so you probably already run Google Analytics on your site. And because you’re already running Google Analytics on your site, it has most of the data it needs. The part it’s missing is being able to modify the site. So Google Optimize links to your Google Analytics account. Now it’s got all the data it needs, but it needs to be able to tweak the site.

10:01
It’s got a single piece of JavaScript that it’ll give you that it says, put this just on every page. So you drop it in the head of the site. Boom, you’re good to go. And that’s even like the simplified non-Google tag manager version of this. If you’re using GTM by that point, you can already figure this out. And Google has in optimized. It’ll be like, Hey, did you want us to verify the tool is there and everything’s working correctly? Like it will really try to save you from yourself. And then after it’s interested in all your site, you just need to tell it which

10:31
pages you want to test or. So then you got to set up a test and that part. All right. Until you have successfully done it the first time can feel intimidating. I watched a few YouTube screencasts to figure it out and played with it. And once I had done it a couple of times, I was like, OK, I know enough to be dangerous. I could figure this out. Or it’s like you’re going to do just a standard. It’ll give you like several options for tests. You want the first one just says AB and then you’ve got variant.

11:01
A is the original unchanged site variant B. We don’t want to add multiple variants because now we need even more traffic. Variant B is we’re going to modify that site and Google Optimize gives you a visual editor where you could just like click the elements on your site, remove them. That’s often a really like where we start with testing. It’s just like, let’s just start pulling stuff out and see if it was helping or hurting. Remove elements or like text you can rename. Or if you know enough to be dangerous, dangerous with CSS, you could just have it inject

11:30
your own written CSS code and then just restyle, however, however talented you may be. let me ask you this. Does testing on mobile throw a wrench into this? Cause you got responsive. just, are you split testing the mobile site and the desktop site? Or do you do those separately? Hi, this is another good question. Come brings into segmentation. So if I really, I wanted

11:56
make sure either I want to make sure that my test makes sense on mobile and desktop. Cause realistically on my website, I’m getting people on mobile and desktop. And it’s probably like an e-commerce is probably 90 10 split. Yeah. 80 20 occasionally happens, but that’s rare. And like, ideally the test makes sense on both. Like if you’re just changing labels, changing colors, it’ll work on both. But you can segment it. If it doesn’t, you could segment to mobile versus desktop.

12:25
Like the test will only run on mobile. Do you have a separate test just for mobile and a separate test for desktop? Because they are different even if like, because the design is so different, even changing, like removing an element will make a big difference on one and maybe not the other, right? Yeah, yes, absolutely. And with weird unexpected results, I’ll do that. And so like one test that really surprised me is I like on the category pages of your site, you’ve got your category title. Do you have like a really cool lifestyle hero image up there?

12:55
I do not actually. OK, well, a lot. It’s not uncommon. And I like those. You know, I used to be a photographer. I like that. I like the fancy stuff. I like the aesthetic. And I ran a split test where I got rid of that hero image on an apparel site with that’s like style that is driven by lifestyle and aesthetic. And the site performs better without the hero image. And that was just on everyone. And I thought, all right, maybe it’s like it’s because it’s mobile versus desktop. So I ran it as a mobile same test.

13:24
You could duplicate it and go, all right, I only want to show this to mobile and I only want to show this to desktop. Same result. It improves it on both. But I, addition to mobile versus desktop, the other segmentation we need to think about is new versus returning. And so you can do that too. And so I did, uh, not often, but a few times I found like, well, this where you had a mixed result on a test, uh, when you run it again as mobile versus desktop or new versus returning, you’ll find,

13:53
This helps new visitors or this hurts returning visitors. And so then what’s cool in Google Optimize, you can actually deploy these things rather than like go commit them to the website. You can have Google Optimize run that test as what they call a personalization, which is really just like it’s an ongoing test without reporting. Yeah. And that’s just built into it. And again, this thing, I know there’s a paid version of it. I’ve never had to use it. It has always just worked with what I want it to do in the free plan.

14:22
What’s so nice about this is it’s integrated into analytics. So they have all the data already. Whereas in the past it was a pain in the butt. Yeah, you had to go through like adding this whole extra layer onto the site just to support split testing. And it just isn’t the case anymore. Let’s just talk about that test that you just talked about. Like it’s generally best practice to just have your products right there above the fold and not have the hero image. I had pains when I was designing my

14:51
category sites because I want to include content for SEO. And I found that it was just kind of pushing things down. So what I did was I moved it underneath the products and that that was my solution. I had a little blurb at the top, but so it’s good to know that approach. That’s often what we do for exactly the same reasons. But it’s good to know that you actually tested that hypothesis. So no hero image products above the fold as high as possible is what converts the best. Yes. And so this one,

15:20
could translate into a best practice or truism, think scrolling, so it turns out the fold is real, at least as far as it affects your revenue. And so getting the products in front of the person as quickly as possible really tends to be a positive across the board. By the way, for you content creators out there, this holds true for blogs as well.

15:49
Like if your post is about like the top 17 ways to optimize your site, you want to put like the top 17 ways above the fold if you can. Some people don’t even, they look at the intro and they go, oh, this is just junk. I’m not looking for this. And then they bounce. But I never split tested that either. That’s just best practice. You could, you could split test that the way you do it. Like you, you’d publish the blog with the intro that you like.

16:16
And then you could use Google Optimize to either like hide that first paragraph or rewrite it. I like that. Actually, that’d be so easy if you have a high traffic blog. Actually, you’d probably get results right away with that. Now the question there is like, what’s the what’s the goal? And so Google Optimize, you can you can tie it to bounce rate. You could tie it to page views. You could tie it to like if you have subscriptions set up as a goal, you could do that. I love it. I love it.

16:44
There was one thing that we were talking about or that you sent me earlier about category pages removing prices of the products. I’m very curious about that experiment because I think I saw it on the e-commerce fuel forums. There’s a big discussion about that. And I think the consensus on the forum was remove prices on the category pages to get people to browse products more. What did you find with your experiment? All right, this one is a must test.

17:12
In some stores, this will absolutely help you. In other stores, it will absolutely hurt you. And then I think the other thing to consider here is like, does the store have quick view? And that I have not gotten into yet, but I know like this one, if you could support it, absolutely try this and test it. I suspect what’s going on here is pricing psychology. When I see that product in the collection grid, I see the product, I see the title, I’ve got context in my head, whether I think about it not, I have attached a potential price to it.

17:42
And as soon as I click through and I end up, whether that’s like a quick view pop up or I’m on the product page, now I see the price. I suspect what’s going on here is if that price is consistently above what the person expected. That’s going to hurt conversion. If it’s below what they expect, it’ll help conversion. So so much of this is about quality of traffic and presentation. And I think that’s why I’d say a must test. Interesting. So in your case, when putting the prices on was better, what was the price point of that product?

18:12
I, all of these tests we did for apparel stores. So our AOV was always under a hundred. Under a hundred. So not high end then. Okay. This was not, not premium. And so this is where you, I think this is so dependent on audience and product. So what was the, sorry, what was the result of that test? In this case, the test ended up, uh, improving conversions and average order value.

18:41
So we gained revenue by not showing collection or not showing prices on collection or category pages.

18:49
If you’re looking to start an online store, you’re probably deciding which e-commerce platform to go with. And the problem is that there’s hundreds of choices out there and they all start to blend together after a while. So I want to save you time today and tell you about BigCommerce, which is one of the platforms that I recommend. And here’s why. BigCommerce does not nickel and dime you with apps like other platforms. Once you sign up, you get a fully featured shopping cart with the features you need built right in. If you want to run a WordPress blog and an online store on the same exact domain for search engine optimization,

19:19
BigCommerce has a nice WordPress integration that allows you to do exactly that to maximize your SEO. And then finally, you have the right to use whatever payment process that you want without paying any transaction fees, unlike other shopping carts out there. Now for geeks like myself, BigCommerce offers a powerful API which allows you to scale your e-commerce store to seven, eight, and even nine figures and beyond. In fact, BigCommerce runs huge stores like Sony, Casio, and Ben & Jerry’s. So if you’re interested in looking for an e-commerce platform,

19:47
go to mywifequitterjob.com slash BC and check it out for free for 30 days. That’s mywifequitterjob.com slash BC. Now back to the show.

19:59
You know, so I was buying a gift from my wife on I think it was the Tiffany’s website They hide all prices. I was just listening your hypothesis now Because if it’s if it comes in higher like every time I clicked on one of those Tiffany’s jewelry pieces it came out higher than I bet it is They have a strong brand So maybe maybe that’s it and they just want to hide the prices to drive people off the site right away because I would have been probably driven away

20:28
Oh, yeah, that’s yeah, because there’s that click that like you have clicked through. You’re more committed. I think that correlates like page views and conversion. I believe they correlate like when you can really get people to spend time on the site. I think they spend more money. Supposedly, Ikea subscribes to this theory. That’s why those places are like labor. And I actually have a question for you. I don’t know if you tested this, but.

20:55
It’s generally best practice if you display the price to list the products from expensive to least expensive, as opposed to the other way around. I’d like the theory here. The theory being like you set the peak price first. That’s the first thing they see. Now everything else seems cheaper as a result, as opposed to if I flipped it the other way. I have not tested it, but testing collection sort order is a really good idea. I’m writing that down.

21:25
Okay, I’ll just tell you the theory for everyone listening. The theory is they see the most expensive item first that anchors the price. they psychologically are willing to spend more or expect higher prices. And then when they scroll down and see everything else is cheaper, they’re like, oh, okay, this isn’t that bad. That first product was expensive. So, these are much cheaper. That’s a theory at least. I don’t know what it is in practice. Well, I’m going to test it on an apparel store. We’ll see what happens. Yeah. Okay. So what else you got, Kurt?

21:55
That’s counterintuitive. Okay. Uh, let’s see what we got here. Okay. Font size. How important is font size? All right. I’ll just give you an experience from my store. It’s huge for our store because we catered a lot of people over the age of 55 who are farsighted and we didn’t do a split test for this, but we had lots of people complain via email that the writing was too small.

22:25
and they would just call their order in and we didn’t want people calling us because that takes a lot of time. So we literally made the fonts pretty big mobile on mobile and desktop kind of actually mostly mobile, should say mostly mobile across the board. And then those calls stopped. Hmm. All right. I like this. So I know the truism here, the best practices and easier to read site performs better. People do not are not going to tolerate your shenanigans with your like light gray font on a dark gray background.

22:55
They’re just not into it. They’ll tiny font. You’re like, looks so sophisticated and subtle. No, it looks hard to read. And so they’re just going to go to another site that is easy to read. And so I knew anytime we make the site easy to read, it performs better. But I said, let’s, you know, let’s, let’s, let’s test this. Let’s get the data. Let’s prove it. And so I tried split testing, just font size and results are like totally inconclusive all over the place. And what it turned out is

23:22
It’s not just font size. It’s more than that. It is about like total readability. so Baymard put together Baymard usability Institute has readability guidelines that they figured out that they recommend. And they’re like, really the thing that matters is line length. And if you’ve ever read a newspaper or magazine, they have those real narrow columns. You never think twice about it because it’s easy to read and it makes sense. And so that’s actually the thing you’re looking for. And so if you have a clever developer, they can

23:50
set this up for you. like the site’s responsive and those columns are always 60 to 80 characters. And then you want to make sure you have a nice line height between them. And that really will make the site much easier to read. When you implement like those quality readability guidelines versus a site that maybe like had full width and tiny fonts and low contrast. Of course, absolutely. It’s going to perform better for reasons you described was it’s easier to read. People are going to quit complaining about it.

24:18
Interesting. So when you did your test, the results were inconclusive. And then you, this, this other study wasn’t your study, right? This was Baymard. The Baymard. Okay. once we, when we, when I did the test again as our, our, just our previous font, which was like, you know, whatever, it was okay. Um, versus like the full Baymard readability guidelines, the thing that changed was bounce rate. Bounce rate went down.

24:47
That makes sense. That makes sense. You know what’s funny, Kurt, I know you have a couple more tests to talk about here, but what are some of the best things to test and what are some of the worst things to test since you’ve been doing this for a while? think the worst things to test are like subjective styling. And that’s where a lot of people start. I think the best things to test really are content. Like every element on the page should ideally be serving a purpose.

25:17
And if it’s not, why is it there? It’s just getting in the way and split testing often reveals that in stark data. So, uh, I oftentimes like phase one is use it as a way to clean up the site without being emotional about it. Like a website is our baby, right? And there are reasons that we did it the way we did it. But when you have the data going, well, here’s what it costs you to have these trust badges jammed in right here.

25:46
Suddenly the trust badges seem less attractive. It’s interesting. I had a friend who is split testing his landing page, which gets like hundreds of signups per day. And he had this landing page where he’s explaining it was a lead magnet landing page where he explained exactly what people got with that lead magnet. And he was testing it against his one where he’s literally a sentence. I will teach you this email form. And that one can perform.

26:14
way better than the one where you actually explain what you get, which kind of blew my mind. And it was ugly. It was ugly, So he’s he’s to this day, he still uses the ugly landing page for him. It doesn’t make sense. Ugly converts is what I’ve learned is some of the ugliest sites I’ve ever seen where we’re like, this thing seems straight up broken are the ones that print money to this day. The most the highest revenue side I ever worked on was

26:44
Also the ugliest. Really? Yeah. I don’t know what the deal is. And highest converting by what metric like compared to just the other sets of users. Not necessarily highest converting. was highest revenue. It remains. is the highest revenue. And I’ve since then I’ve seen other beautiful sites that are high revenue, but that one stuck with me and that I know good design is not a requirement of high revenue. Sure.

27:13
I mean, I got a story to tell here. A long time ago, I had my site over Bumblebee Linens have left side navigation because it kind of made sense, right? You look from left to right and all my product categories on the left. And I was in this mastermind group where literally everyone there said your site is heinous. You need to be redesigned. And I ended up redesigning the entire site. Problem is, I changed like 20 different things, but it ended up converting like almost 50 % better.

27:44
So when you’re talking to someone about split testing, do you recommend changing one thing at a time or just making like a big change in testing at all at once? Like how have you been conducting your experiments? When you test one thing at a time, you it’s way more often that you end up with the inconclusive results. Test, doing like, we changed this entire site. That’s not that useful. Focusing on like, I think the, the middle ground.

28:13
The thing where like you’ll see the fastest, most consistent results, where it’s easy to wrap your head around it, where it’s just more practical is redesign a template at a time. Like this is, going to, we know the cart page is this bottleneck, you know, cause we could see like how many people viewed a product, how many added to cart, how many reached checkout. Maybe we know the cart pages are bottleneck. All right. So let’s come up with a theory and make three different versions of that cart template. And now we’re going to test.

28:43
systematically test those against the original and see, all right, which one performs the best art? Let’s implement that. That seems to be a better approach. And when you have multiple versions of entirely different pages, yeah, you don’t know what the individual thing is that perform better. But as you’re iterating through tests, you start a picture starts to emerge of this is what my audience likes. And that’s, think one of the other like real learnings I’ve discovered here is it really comes down to

29:12
Well, what does this audience expect? What do they want? What do they prefer? And it often does not have anything to do with what you as the brand owner, what like design trends, like what everybody else likes. It’s the people who you’re asking to open their wallet is their opinion is what matters. You know, it’s funny is I used to be really into split testing or not, not doing them, but actually just reading up on them. And I remember there was some statistic like

29:42
I can’t remember the exact percentage, so I don’t wanna screw it up. But the majority of the successful split tests are when you remove stuff from your page, not add to it. Because ultimately every page should just have one goal where you want people to go and the more things you have people look at, the less likely they’re gonna do the one thing that you want. At least that’s the theory. I agree with that. Yeah, so I’m curious, what other split tests do you have to share? I think the other one that’s worth looking at again with new eyes

30:11
is free shipping. I’m no longer convinced that we have to have free shipping. Really? Okay. There’s always, there’s been this idea that the number one unexpected, there are unexpected shipping expenses, the number one conversion killer in checkouts. I don’t think it’s true anymore. So you could, depending on the platform, there are ways to split test shipping results or shipping rates. for us, it’s like, is you have to be on Shopify plus.

30:41
And there’s a couple of apps, IntelliGems or Ship Scout. So that’s what we’re using specifically. I don’t know how to do it with Shopify and Google Optimize. And with these apps, we can offer different shipping rates and they’ll have an announcement bar on the site itself. like before they hit checkout, they’ve seen what the free shipping threshold is. And it’ll tell you like, this is the conversion rate with these different checkout thresholds. And if you give it the, like your average cost of goods sold,

31:09
and your average shipping expense per package, it’ll also figure out profit per visitor. And once you start doing that, suddenly you discover, oftentimes offering no free shipping at all ends up being way more profitable overall. And the impact on checkout conversion with different thresholds is significantly smaller than you’d think. All right, give me some numbers here. I’m fascinated by this. All right. So we tried this in a store.

31:39
Um, and I’ve run this test now and in four separate stores and we’ve gotten similar results every time. Let’s say no free shipping. This is check our conversion rate. No free shipping will be, uh, about 70 % free shipping at $25 at checkout. They’re going to convert closer to 80%. And then at $75 free shipping threshold, they convert at 75%. Okay. So that’s in line with what you would expect, right?

32:10
Yeah. It’s a 75 is like the ideal balance, but really what matters is profit. so at profit, like for this particular store, their fulfillment cost was 750 in order. I don’t know where that falls in with yours. know, fulfillment costs are all over the place, depending on how heavy your goods are. Yep. But like for this particular store, uh, the $25 free shipping threshold profit per visitor, 13 bucks.

32:37
the profit per visitor at $75, $16.50 profit per visitor with no free shipping, 20. So the chances are, I probably want to not offer free shipping at all is going to result in the highest revenue and profit. Interesting. Yeah. Most people don’t think about profit. They think about conversion rate. I will add a wrinkle to that though, because once you get someone to open up their wallet,

33:05
then you can probably sell them a whole bunch of other stuff. So maybe it’s a little more complicated than that, right? Now, yeah, I could do cross sells and use free shipping as a nudge. What I like about not offering free shipping all the time is that that can then be one of your promotions instead of having to rely on discounting or new product development. You know, yeah, you know, I agree with you. I hate discounting. And whenever I think about discounts, I think about Bed Bath and Beyond.

33:35
Oh, right. Are they still in business? The local one closed like the two Bed Bath and Beyonds by my house closed because I would never shop there without that 20 % coupon. I suspect other people were too. But yeah, profit wise, I can definitely see why that’s the case. We do a shipping threshold and we set it at just a threshold where we try to get someone to add in one more item. So what’s the best practice then, Kurt? What do you recommend here? Well, so the issue here is

34:04
Without this, like these very specific tools, it’s really hard to test. And even within that, like, depending on how your, your fulfillment is set up, testing these can be really hard. So I would still say, Hey, you probably want to offer free shipping. I would make that free shipping threshold higher than your average order value would be my, my safe starting point, but don’t.

34:31
don’t sell yourself short and have like an unnecessarily artificially low free shipping. Cause it’s just a tremendous cost center for most people, especially shipping is getting expensive. It is getting really expensive. In fact, we had to jack up our shipping rates cause yeah, it’s like a loss leader. Yeah, it’s bad now. I mean, it’s going up like all the time before like USPS used to raise their price once a year. I feel like it just creeps up like every time I check now. we did, uh, was working in a store that had

35:02
heavy items. ended up being like six pounds shipped and it was glass. So you had to like really pack it carefully. And it was, it was like, Oh, it’s shipping, you know, from the like Northwest to Wisconsin to the Midwest. And it was like, think it was 30 bucks and like the item itself, you know, with like 40, 50 bucks, like the shipping was over 50 % of the cost. was crazy. It was, what do you do? I don’t know. Do you really, once the item hits a certain weight, like there’s only so much you could do.

35:31
to minimize shipping costs. You know what we’ve struggled with? When people need it overnighted, it’s often really difficult. Like we have table rate shipping. It’s often really difficult to predict because it depends on where they are geographically also. Oh, you got to get that carrier calculated shipping. That makes life so much easier. don’t like that? Because sometimes it goes down and then you don’t get any shipping costs. It actually happens more often than you think. When anything is tied to another service, I don’t like that.

36:01
a buffer to it to make up for that. You know how often it’s low and by how much. And so add either a percentage or a fixed dollar amount to the carrier calculated rate to try and balance it out. Yeah. Yeah. I mean, there’s technical ways to solve this. I just haven’t expended much brain power on it, but occasionally it’s a problem and I get pissed when it happens. Let’s just say the worst is people pay for the overnight. It’s expensive. And then the item does not arrive overnight.

36:28
That’s another story because it’s gotten much less reliable to like FedEx. That’s my resistance. And then you have to be like, all right, well, we’ll refund you the cost of the shipping. And so now, like all of the profit on that order is gone. Well, actually, the carriers started reimbursing again for late. Oh, all right. They stopped doing that during the pandemic, which really was annoying. But they it’s come back, although they make you jump through more hoops now to get it. So anyway, that was just an aside. Sorry about that.

36:58
guys are listening. Do you have any more split tests to share with us? These are really fast. By the way, yeah, one one last quick one. If you you know, on Amazon, it’ll be like your recently viewed products. Yes. I think those are those are neat. This is a good personalization one. We ran this on a for new visitors, the your recently viewed items decreased conversions by nine percent.

37:27
For returning visitors increased it by 33. And so this is an example of like, wow, testing the different audiences and displays can be hugely beneficial because you definitely want that recently viewed items widget. And I like to put it on the cart page. You definitely want that on the cart page, but if you’re there a new visitor, you actually don’t want it.

37:53
If it’s a new visitor though, they presumably don’t even have anything populated there, right? They won’t. And then they’ll have like, you know, as soon as they view something, it’ll show them like the one item is usually how those are set up. Can you explain why you think that’s the case? No, this baffled me, but it’s all like, all right, we’re going to try this again. And so I ran it again and I also did it mobile versus desktop and the results with this were consistent. Really? I guess it’s just more things to click on then, right?

38:21
I assume it’s it goes it like when you’re new it starts as a distraction when you visit the site leave come back a few days later then it’s helpful. That makes sense I guess because if you leave and you come back you probably already know what you’re probably ready to buy or more ready to buy. Yeah I think I’d stop for most stores on subsequent visits like if you’re returning there you’re much more likely to buy than when you’re first poking around. Whereas I guess when you’re first there you’re looking at something and all of sudden you got like 10 choices you’re like forget it I’m not.

38:51
I’m just not buying here anymore. Oh, choice paralysis is totally real. I have not worked out a test for this yet. Yeah, I guess I could hide like total number of items displayed, but that would be an interesting test. It’s like, let’s figure out is paralysis analysis real? And I’m convinced it is.

39:08
I just want to let you know that tickets for the 2023 Seller Summit are now on sale over at Sellersummit.com. Now, what is the Seller Summit? It is the conference that I hold every single year that specifically targets e-commerce entrepreneurs selling physical products online. And unlike other events that focus on inspirational stories and high-level BS, mine is a curriculum-based conference where you will leave with practical and actionable strategies specifically for an e-commerce business. And in fact, every speaker I invite is deep in the trenches of their e-commerce business.

39:37
entrepreneurs who are importing large quantities of physical goods, and not some high-level guys who are overseeing their companies at 50,000 feet. The other thing I can assure you is that the seller summit will be small and intimate. Every year, we cut off ticket sales at around 200 people, so tickets always sell out fast, and in fact, we sell out every single year many months in advance. Now, if you’re an e-commerce entrepreneur making over $250,000 or $1 million per year, we are also offering an exclusive mastermind experience with other top sellers.

40:06
Now the Seller Summit is going to be held in Fort Lauderdale, Florida from May 23rd to May 25th. For more information, go to sellersummit.com. Once again, that’s S-E-L-L-E-R-S-S-U-M-M-I-T.com, or just Google it. Now back to the show.

40:22
So one of the things on my list to implement on my site is frequently bought together. I’m very curious how that performs. Cause I see it on Amazon. Generally I get these ideas when I shop on Amazon, I go, Hey, that’s a nice feature. Let me try that. Which is absolutely the wrong thing to do by the way, if anyone’s listening, like Amazon is not your site, obviously, but that frequently bought together. I always think of ways to just try to increase the average order value. Have you ever tested that with any of your sites? Frequently bought together, well, could

40:51
Assuming like the merchandising makes sense, it does consistently help to boost average order value. And this is another thing with split testing. Like you really don’t necessarily, we’ve implied it here, but you don’t necessarily want to just strictly have conversion rate or transactions as your goal. You can look at revenue. So that includes average order value that could be helpful or.

41:16
a lot of these other items like, all right, does this decrease bounce rate? Does this increase page views? Can I get people on the site longer with this? So once you have like pages instrumented, you can literally test all these variables all at once, right? Without any extra work. Yes. I don’t know what the there’s a limit in Google optimized, but you say like, this is my primary goal. And here are like, I could choose secondary goals. And I think it’s limited at like two or three, I forget which. So that’s really helpful.

41:46
And then you could run multiple tests at once, assuming you have the order volume to support it. But as you add them, it really starts slowing things down because you’re dividing up your audience. Right, right. You know what I have always wanted to know? You know that certain sites have that slide up? Like when someone buys something, it says, this person just purchased that. Oh, the FOMO pop-ups. The FOMO pop-ups.

42:15
always been curious whether that increases the conversion rate or not. I think that’s another one where it depends on the brand. At this point, when they happen, I’m so blind to it. I don’t even see them anymore. And I’m just like, well, at least it’s not the spin to slide out. You know what’s funny about all this is I almost feel like you have to split test often because things tend to work when they’re new. But then, like you said, once people become jaded to something,

42:44
It probably doesn’t convert as well anymore. A hundred percent. Yeah. What works now may not work in 12 months. So what is your strategy then on keeping this up with your clients? Are you constantly running a split test of some sort? Yes. I wish I had some like formal approach to like, all right, you know, we implement this and then we’re going to set a reminder. We’re going to revisit this. It’s more like, um, for the people we’re doing split testing with, we have regular recurring meetings.

43:12
And it’s like, we’re just always going through KPIs. So guess it’s data driven and then trying to identify like, these are our opportunities. And if it comes in, if we get to a point where it’s like, all right, well, suddenly the cart page is our bottleneck again. Often I’ll revisit tests just to like, Hey, let’s just, let’s verify that it is what our belief is still the case. Yeah, absolutely. Nothing wrong with like, let’s run that test again. See, you know, maybe that was a weird outlier.

43:40
So we’ve presented a number of tests to the listeners. Let’s say you’re just starting out and you want to try this. What is like some low hanging fruit that you should test? For sure. I would. I would. You’re going to look at like, what are the prime pages here? It’s going to be homepage category product cart. Look at me. One of those four templates. And I like to like cart is a really good place to start because you get yourself in less trouble. And.

44:10
Cause like cart, we know the very next thing we want people to do is proceed to checkout and make a purchase. So you set that as your goal, uh, go on the cart template and then within it, a lot of times over time, like the cart develops a lot of cruft. And so I think the ideal Google optimized split tests, and this is, this has come up, this has been kind of a minor theme here is what elements. Aren’t aren’t doing us any good. What’s cruft at this point, you know, can we just shave off some of this stuff?

44:39
And so just like Google split tests where we go, this is the original version and this is the cleaned up version where we pulled all the garbage off. And that’s like such an easy thing to do and optimize, select the element, hit remove. It’s gone like magic. Give me an example of cruft, just hypothetically speaking. All right. On a cart page, I often see like, um, a shipping calculator. And I always think this is so weird. Like you’re going to get your, your shipping rates in the cart just to go to checkout and then get them again. I get you like, well.

45:09
We doubled up on steps here. So if I see one of those, like I’m immediately testing it by just like original versus we got rid of this thing. I would actually, now that you think about it, there is a lot of cruft. There’s like some people have coupon codes on there too. Yep. And apply the coupon code. And I bet that always hurts because if they don’t find a coupon that works, they might leave, right? Yes. I actually hid my coupon code field for a long time.

45:38
From people because they I found that people were just going off and going to like whatever those coupon sites were So I made them like it was just a link they had to click on it to expose the field and I think that mitigated that problem a little bit it still happens, but But having on the cart page, I personally think and again, I don’t have any data for this I personally think that’s a no-no along with the calculator. Oh, yeah, calculator drives me nuts Yeah, okay. So start with a shopping cart and then what would be your like your next page? I would then do I’d work backwards

46:08
Then I go product and like the product form specifically tends to pick up a lot of junk. And that’s a thing that’s like usually above the fold or near it. And so that’s a good place to test. That’s also one where I’ll test appearance, like on the product form, like, all right, what happens if I put like a light background and a border on this to call attention to it? Sometimes that helps or it makes no difference in which case, like choose which one you like better aesthetically. And then, all right, collection page, like hide show price.

46:37
is a common one. Hero image collection description, filter sidebar, and then work back to homepage. And then homepage is just like, there’s so many weird random sections on a homepage. Yeah. All right. Which of these help or hurt? If you really want an exotic test that makes everybody nervous, get rid of there. I guarantee the first thing on your homepage is a hero image or a slider video. Hide it. You don’t even replace it with anything else. Just hide it. What happens? That one often the site performs better without it.

47:07
Really? Okay. Yeah, that’s fascinating. Okay. Here’s why I I’m skeptical, but I mean, obviously you have data and I don’t oftentimes if it’s someone brand new and again, I guess you have to test returning versus new if someone’s brand new landing on your site and it’s not clear what you sell or what you’re good at. I feel like they’re to leave because you have to kind of prove to a new a new customer that I agree with you, but I think often brand owners wildly overestimate

47:37
how good their hero slides, et cetera, are at explaining that. That I can believe. That I can believe. Okay. That is very interesting. That test would scare me, actually. But I’d be very curious. It scares everybody. Okay, yeah, that’s good. So I guess there is a distinction between you and returning visitors that you have to worry about when you’re… I’m running a test on a site right now. It’s four days in, so it’s not statistically significant, technically.

48:05
And we’re doing it’s their homepage. And I got three versions, the original, has like hero content and shopping a version that’s anything that’s non-shopping is gone. So this thing is just categories and products. And then a version that does not have the hero slider. That’s the only thing I took out. The shopping only site has a probability to beat the original of 96 % and the version with no hero image, 90%.

48:35
Sorry, you lost me there. So what’s in the lead right now? The shopping only site. is purely, we took out any about, any blog posts, hero sliders, all the content, gone. It is just, these are the departments, these are the products. That’s for the board. That one converts at 2.3%. The version of the page that is just the original page minus the carousel slider converts at 2%.

49:03
And then the original with all the junk in it is one and a quarter. So based on that, it sounds like it’s the hero image and it’s because it’s similar to what we talked about with the hero image on a collection page. You’re pushing so much stuff down the page. Right. Yeah. And I agree with, uh, it makes intuitive sense. Also removing all navigation outside of like finding the product that you actually want to buy. I move all that stuff in the footer usually a captive.

49:32
Captive page. Right. Dude, Kurt, this is awesome. I’m actually very curious about all the split tests that you’re running and especially the one about like the reversed price order. I’ve always been very curious about that. That one I’m going to run in my, I figured out how to do it in my head. It’s going to be a redirect test using query strings. Yes. Yes. That would work. Yeah. Cool. Kurt. Hey man. I should have you back more often. This is, this is interesting stuff.

49:59
If anyone needs help with their Shopify store, if they have any questions for you, where can they find you? On Twitter at Kurt Inc, K-U-R-T-I-N-C, or grab my newsletter at curtlster.com. Awesome. Well, thanks for coming back. My pleasure. Thank you for having me.

50:16
Hope you enjoyed that episode. Now this just goes to show that before you make any assumptions about your business or your customers, you have to test everything. For more information about this episode, go to mywifecourterjob.com slash episode 428. And once again, I want to thank Clavio, which is my email marketing platform of choice for e-commerce merchants. You can easily put together automated flows like an abandoned cart sequence, a post purchase flow, a win back campaign, basically all these sequences that will make you money on autopilot. So head on over to mywifecourterjob.com slash KLAVIO.

50:47
Once again, that’s mywifecoderjob.com slash KLA VIO. I also want to thank Postscript, which is my SMS marketing platform, a choice for e-commerce. With a few clicks of a button, you can easily segment and send targeted text messages to your client base. SMS is the next big own marketing platform and you can sign up for free over at postscript.io slash dv. That’s P-O-S-T-S-E-R-I-P-T dot I-O slash dv. Now I talk about how I use these tools on my blog and if you are interested in starting your own e-commerce store,

51:15
Head on over to mywifecoderjob.com and sign up for my free six day mini course. Just type in your email and it’ll send you the course right away. Thanks for listening.