This is the transcript for Episode 1 of The Conversions Podcast with Michael Aagaard.
Francis: Welcome to the podcast, Michael.
Michael: Thank you very much. Thanks for having me.
Francis: So, could you tell us a bit more about how you got into the conversion optimization space?
Michael: Yes, sure!
Well, many years ago while I was still studying at the Copenhagen Business School, I got offered a job at an online agency. I was a translator back then. I had a little translation company going while I was studying and, yeah. So, I got a freelance job at an online optimization agency, mostly working with SEO, and I didn’t really have any prior experience with anything online, really, and I was completely blown away by it because, all of a sudden, I could see how words have a direct and measurable effect. It was mostly SEO I was working with back then but it was very interesting to me because, all of a sudden, I could see that it was more than just communicating — it was also actually optimizing your business and optimizing how the customers find you.
Ando so, I got more and more into the copywriting bit and then I started finding out that there’s got to be a way to write online copy in such a way that it makes people do what you want to do. And, from there, I kind of got more and more interested in sales and marketing in general and I read all these books and I got really fired up. Once I started using all the techniques that these different so-called gurus were teaching me, I found out that they didn’t really work, you know? And so, that kind of led me into the whole split testing thing because then I thought, “Well, if this doesn’t automatically work — what all the gurus are saying — then I need to find out what really does work,” and that’s when I really got this huge “Aha!” experience because I found out how easy it is to test things online, and I actually find out whether all your hypothesis and assumptions actually work in the real world.
And, from there, it just kind of escalated and then, I mean, I quit the agency job and I became self-employed and I’ve been self-employed for the last four years and it’s just been a lot of testing and hands-on optimization work since then. So, that was kind of the semi-long version of the story.
Francis: Great. So, you’ve been doing this for about four years?
Michael: Well, I’ve been doing it for longer. I’ve been doing it for a bit longer because I was at the agency for a couple of years also so I think it’s about, yeah, I guess, six years in all with my freelance period. But I’ve been doing it solo for, yeah, well, almost four years now.
Francis: Well, four years is quite an eternity in the conversion optimization space. I think this is quite a new thing to do, I guess.
Michael: Definitely, definitely. And, I mean, we’re all learning, and we’re all getting better, and, you know, I think the more you test and actually have scientific research and evidence-based conversion optimization, well then, we’re all kind of, you know, helping each other out and helping everybody who has a website — helping them out with finding out what really works in online marketing.
And I think that’s one of the main things why I really like working with the online medium, that is that you can actually test and measure pretty much everything which is very difficult in the offline world, and I think that, in the offline world, still, I mean, there’s a lot of assumptions about what works and what doesn’t work, but it’s very hard to actually find out specifically if it does or does not, and that’s the beautiful thing about working online as is that you can pretty much test everything and actually find out whether what you’re doing actually works because the word “optimization” gets used very liberally. “We optimize this,” “We optimize that,” “We got this guy, yeah, he optimized our website,” but very few people actually take the time and effort to find out whether it was an optimization, and the only way to know whether you’re actually optimizing your conversion rate not hurting is by getting data — it’s by testing it and finding out whether it really works.
Francis: By testing, you mean split testing?
Michael: Yes, split testing.
Francis: Okay. I would just like to follow, well, your train of thought on the whole having an online presence then you can do some split testing and stuff like that. What I found — this is my personal experience — is today, in 2013, we have people with multiple devices, you know, a person might have a phone, might have a smartphone, might have a tablet, might have a tablet, might have a laptop, might have a desktop at home, and it’s becoming increasingly problematic to be able to track conversions all the way through all these devices. So, do you have any thoughts on that?
Michael: Well, I mean, when you set up a test, I mean, you can segment on channels, for example. So, you only run a particular test on, for example, mobile devices or on iPads or you only run it on people that actually come in from a PC. So, that would be one way of doing it. But, another thing, I think — just to take it one step back — is that conversion optimization, a conversion can be many different things and we have a tendency to see a conversion as a sale, as a hard conversion, but, you know, a conversion is pretty much every time you get someone to do something you want them to do online.
So, a conversion could be to get someone to sign up for a newsletter, or it could be to get someone just to click on a button, or go to a particular page, or download a PDF. So, conversion optimization can be many different things and there could be a lot of, like, micro-conversions and you can have a macro-conversion.
Tracking the macro-conversion all the way through can be difficult but, in many cases, you might want to start out with something else. I mean, just for most companies, the newsletter represents a huge value. So, the more you can get to sign up on your newsletter, the better.
So, there’s a lot of different areas you can attack with a conversion rate optimization mindset.
Francis: So, do you have an overall strategy or process you use for increasing conversions on websites?
Michael: I do, but I think there’s two things that are important kind of to talk about generally in relation to conversion rate optimization and CRO strategies, I think.
Well, first of all, I think the main thing I’ve learned is that, you know, all businesses and products and websites and landing pages are different and, you know, the motivations of the potential customers are also going to be different. So, you know, there really is no one-size-fits-all solution that works every time and you really need to focus on the kind of the individual website or the individual project and find the best solution for that particular case.
As humans, we have a tendency, you know, we like the cookie-cutter approach. We like the idea of learning to play guitar in three days and we like rules, you know, “Your copy should never be more than 200 characters,” for example. Or, “You should always have a green button.” But, in my experience, it’s very difficult to try and develop rules like that because just because a green button worked on one website doesn’t mean it’s going to work on another one. I mean, if your entire website is green, a green button will probably not do very well; a red button will probably do well. So, you really need to understand that all cases are different and that you have to attack it as an individual case.
And then, the next thing, again, touching on what I talked about before is that you really need to find out if what you’re doing actually works because one thing I’ve learned from having performed several hundred tests is that sometimes you’ll really be surprised. I mean, I’ve had tons of tests where all my experience, all logic had told me that my new treatment is going to kick ass but actually it converts, you know, worse — it actually hurts conversions compared to the control version. And that, I mean, in some cases, I’m like, “Wow! I really didn’t expect that!” and then you kind of have to go back to the drawing board.
So, just assuming that you know what’s right every time is dangerous and also, you know, pretty arrogant if you ask me. And, the only way to know that you’re actually optimizing is to test, and that’s the only way to get away from guesswork and gut instinct.
Third thing that I think is very important is actually to have a clear idea of what you’re doing, you know? Actually having a strategy. So, you need to identify a problem and then you need to find out how to solve that problem. I think a lot of people who are just getting into split testing, they get excited and they go, “Okay! Okay! Cool! Let’s change the logo,” or “Let’s put up a yellow button instead of a green button,” or “Let’s whatever. Let’s change the global navigation,” and then they set up a test with two different variants but they don’t really know why, you know? There’s no kind of clear strategy. There’s really no idea of why they’re testing it and what they want to achieve, and those kinds of tests usually are not going to be very valuable because, I mean, even if you ramble by coincidence into something that gives you a lift, you really have to no clear idea of why you got the lift or what really made it happen.
So, you need to be very clear about what you want to achieve and how to achieve it. And, having said that — that kind of leads me into the strategy I use. So, conversion rate optimization is, you know, it’s really about optimizing decision-making processes. You want to optimize the decisions and the actions of your potential customers. You want more of them to do whatever it is you want them to do. So, it’s important to focus on the goal because the page or the copy — whatever you’re trying to optimize — that’s not the end; that’s a means to an end. It’s a means to optimizing the decision-making process.
So, I like to start out by identifying a problem that needs to be solved, and a problem that needs to be solved could be getting more newsletter signups, it could be getting more leads, it could be to sell more products, reduce the bounce rate of the home page, it could be a lot of different problems. But, once you identify a problem, for example, let’s take a newsletter sign-up, for example. On the website, we have a newsletter signup. It’s important for us to get as many signups as possible because we do a lot of selling directly through the newsletter. Right now, we have problem and that is that we only have 0.03 percent signing up for the newsletter, okay? So, that’s a clear problem. You’ve identified that. Then, you need to come up with a hypothesis for what’s causing the problem — why aren’t there more people signing up to our newsletter? And, I generally work with three areas which are clarity, relevance, and friction. So, clarity and relevance are positive. We need to increase the clarity and relevance. And then, we need to reduce friction because friction has a negative effect. So, you could work with, say for example, “Well, my hypothesis is that we have an issue with clarity because the copy on the form — the newsletter sign-up form — is very generic. It just says ‘Sign up’ so it doesn’t really talk about what you get.” So, the hypothesis could be that, by adding clarity, we can get more people to sign up.
And then, you have to come up with a clear hypothesis for how to solve the problem. So, that would be, for example, “What is the main benefit of signing up for the newsletter?” Well, the main benefit is that you can get 20 percent discount, for example. Well, then, that might be a way to clarity the value of signing up for the newsletter.
So, once you have those two hypothesis, you put those together for a main test hypothesis. So, that could, for example, be the lack of value and relevance conveyed by the form copy is causing friction that keeps potential customers from signing up. By adding a value proposition and relevant button copy, we can reduce friction and increase conversion, okay? So, then, you have a totally clear hypothesis and this is what the whole test is revolving about.
The next step is to come up with a treatment based on what you hypothesized and then it’s time to set up the test and find out whether your whole kind of hypothesis was working. Then, afterwards, you want to record all your data regardless of whether it was a successful test or not, whether it had a positive effect or not, and you need to, of course, validate your data. You need to be sure that you can actually trust the data you have so you don’t stop the test before you have a test result, and that’s a problem for a lot of testers out there. That’s quite a large area in itself so I’ll provide you with some links for that afterwards so people could check it out.
I’m doing a lot of talking but that’s actually the strategy, I think.
And then, afterwards, one thing that I think is very important is to make sure that you have, you know, screen dumps of the different versions, the different treatments you’ve tested of the control version, you need to have your test hypothesis written down, and then you have to, you know, provide some answers afterwards — “What did we learn? Did the hypothesis hold water or did it not?” and then you can say, “Well, if it didn’t, and the treatment performed worse than the control, well, then, you can develop a new hypothesis for why that happened and then you can set up a follow-up experiment.” And, as long as you record all your data and as long as you have a clear idea about where you’re going and how you’re going there and what happened along the way, well, then, every test you’re going to perform is, in fact, going to be a successful test because, even if it didn’t give you a lift, it gave you an important learning and that’s the most important thing because a lot of people have a tendency to say that, “If the test gave a negative result, then it wasn’t a successful test.” But, as long as you learned from it, as long as you know why, then that’s a value in itself because you need to say, “Okay. Well, that solution didn’t work. We can cross it off the list and then we can continue with something constructive that will actually work.”
I think the most important thing is to actually have a strategy and the strategy is to identify a problem, create a hypothesis about what is causing the problem, create a hypothesis about how to solve the problem, based on those hypotheses come up with a treatment, test the treatment, and then look at the data afterwards.
Francis: The point you made about every single test and every single website and every single market and every single client, everyone is different and your customers are different, and I got into conversion optimization by testing the buttons as well. Like, you know, I’d test the buttons, it’s orange, it’s big. Don’t need to increase the CTR. It’s good. But, like, as I got into it, I realized that, “Well, every single website, every single product is different,” and there’s really no cookie-cutter approach. I wish there was, but there isn’t. You know, it’s what makes it exciting as well, I guess. And, I mean, the point you brought up — like, you focus on clarity, relevance, and friction — that’s really interesting because, when I do optimization, it turns out that a lot of my clients, they have, like, commoditized products. So, the thing I tend to focus more on is, like, credibility. “How do I market this product and show you that this is a credible product?” and that has really helped my clients. So, it’s really interesting that you take a different approach but it’s still part of the same overall strategy.
Michael: Well, I’d say that lack of credibility would be a cause for friction so I’d probably — in my approach anyway — I would probably just have that as a point under friction. For example, I would say, “So, if I look at friction on this particular landing page for this particular product, well, there’s a serious lack of credibility and that is causing friction,” because, I mean, friction can be identified as, I guess you could say, something that slows down the decision-making process — that makes it harder for the potential customer to make the right decision and the right decision is saying yes to whatever it is you are offering them.
So, something that exists in the mind is psychological, but it happens due to something in the marketing process. So, something that would slow down the decision-making process would be a lack of credibility, for example, and then the hypothesis would be that, well, due to the lack of credibility, there’s a certain amount of friction going on on the page which keeps people from signing up or buying the product. And then, you know, the hypothesis from there would be, by adding testimonials and trust seals and case studies, we can increase the credibility of the company and thereby reduce friction and get more people to sign up and buy the product,, for example.
Francis: So, when I run tests for clients, I always feel that they don’t know that there are certain validity threats, for example, some products, and they will insist that, like, “Oh, no. My product is not seasonal,” but I run the test on a public holiday or over Christmas or over the New Year and there’s just crazy conversion spikes that totally disappear right after the holidays end. After, like, a couple of times I experienced this and I became really very careful about this and I always advise the client, you know, “There’s a possible threat here. We must be careful about when we run the test and as well as, like, if you have any, like, PR coverage, this needs to be communicated to me because that might affect the amount and the quality of the traffic that’s coming to the site,” and all that and I become quite paranoid about that but I’ve found that a lot of people, when they run the test, they don’t care.
Michael: Well, I mean, I think it really depends on what you’re testing and it sounds like you’re running tests on products that, you know, for example, the level of motivation of the customers will kind of go up and down and vary a lot or where the products kind of have a seasonal quality to them. So, I think it depends on what you’re testing. But, of course, it’s very important to be aware of. But, I think, I mean, one of the main things about A/B testing, the main thing is that you can cut out a lot of bias. I mean, if you were comparing periods then you would have a lot of validity threats and you can kind of reduce those a lot because you’re running, it’s the same period of time and it’s going to be the same quality of traffic on both variants. So, that helps you a lot.
But I have run tests, for example, where I was helping out this one American online marketing guy and he’s a popular speaker and he has, for example, he has an ebook that he wants people to download because then, you know, they become part of his newsletter and his loyalty program. So, I was helping him. I did a few changes on the sign-up form for the free ebook. And so, what happened was that, you know, I started the test, he has a fair amount of traffic so it was going pretty fast and every time I almost had a conclusion, the difference kind of disappeared between the two variants. It flattened out completely for a day or two and then it started developing again and I saw that a couple of times where you get the diamond shape and that was so weird because every time I was almost ready with a conclusion then, you know, the test tanked. And so, I called him up and I said, “This is weird. I haven’t seen this before,” and then, we were looking at the dates and it turned out that those were the exact days that he had been out speaking at an event, and at the events, people love him, and he’s a very popular speaker, and he pitches the book. So, that means that the motivation level of the entire audience — maybe he was speaking to 800 people and you get a storm of visitors and they don’t really care about the sign-up form because he’s already convinced them that they need to download the book. So, that would be, you know, a validity threat because the motivation of the potential clients is so high.
So, yeah, definitely. You need to be aware of that. But, again, it depends on what you’re testing. If you’re testing clicks from one page to another or newsletter sign-up forms, well, that, you know, of course, it’s important that you know what you’re testing actually works, but I’d say, as compared to, you know, critical steps in the last part of the checkout flow, for example, that would be critical, critical, critical that you get that right. So, I would run that test for a very long time and with a very large sample size in order to be 100 percent sure, or as sure as possible, that, you know, you can actually trust the data you’re getting in.
So, yeah, it really depends on the case. But, of course, it’s always important to be aware of fluctuations.
Francis: How do you get aware of the fluctuations? Do you just tell them beforehand, like, “You know, if there are fluctuations…” or do you wait for the problem to come up and you let the client know?
Michael: If I do know beforehand that there’s going to be a serious validity threat of some sort then, of course, I’d be aware of that. But, I mean, usually you don’t know before you get the test started. So, you have an idea of how much traffic you have and you have an idea of what’s expected and you get the tested then you can kind of see the development from there. And, I mean, what you’re always looking for is just a nice, clean development in the test. I mean, the more spikes you get, the more diamond shapes where the two aversions kind of cross over, or at tanks, I mean, the more fluctuations would be a better word, the more you have to be aware that there’s something going on out there that’s affecting conversions, and then you’ll have to delve into whatever that could be. But, if you have a nice, clean development from beginning to start where you can see that one variation is doing better than the other one all the way through, that’s a nice sign that you have a clear test where you don’t have too many threats coming in.
Francis: So, before I get too many complaints about this podcast, I know what a diamond shape means but maybe you could explain for everyone else what the diamond shape means?
Michael: Oh, yeah. Well, if you have a graph and the two different variants that you see on the graph, if they start crossing over and crossing over again, then that’s actually going to paint kind of what looks like a diamond because that would be the two versions — one taking the lead over the other and then the other one going back again — and, the more diamond shapes you have, kind of the less I would be inclined to trust the data because then you have large fluctuations and that would be an indication that you don’t really have a clear tendency — you don’t have one version doing better generally than the other one the whole way through.
On the other hand, if you have a nice, clear graph where one is just pretty much permanently outperforming the other one the whole way through, well, that’s a very good sign that you can trust your data.
But I think the main thing about validity is that, one of the main problems when people start testing, when they start using test tools, is that they trust their tool blindly and that they only focus on statistical significance. They are only looking at how much chance there is that one variant is going to beat the other. So, a lot of people, they might stop the test the first time it hits 90 percent, 95 percent of statistical validity, without paying attention to how large the sample size is, for example — how many conversions have been through, what has the devolvement been in the test if you look at the graph, and consider factors like standard error. There’s a lot of different things that go into determining the validity of a test and you need to look at several different factors and not just the confidence level.
It depends on what you’re testing and how much sample you have. But, I mean, as you know from experience I’m sure, during the first couple of days, for example, depending on how much traffic you have but, if you have a small sample size, it takes a lot less to get fluctuations because the sample size is so small. So, maybe one conversion or two conversions on one variant is enough to actually make the test flip over totally. So, I mean, within the first couple of 100 hundred visits, you’re probably going to see a lot of fluctuations. But, after that, you know, you should hope that the test is going to kind of have a clear development all the way through. But it’s natural in the initial phase of the test that you’ll see larger fluctuations.
As you also know, but just to the listeners out there, it’s a lot about experience also. You know, getting a lot of testing done and getting it under your belt and kind of getting a feeling for how a test develops, you know, in order for it to be valid and in order for it to be trustworthy.
Francis: So, what are some of the conversion problems you see on the websites you’ve done optimizations for?
Michael: Generally, what I see is there’s a lot of kind of lack of clarity and relevance, and that can have to do with, I mean, a lot of different aspects — for example, the design and so on — but, often, it has to do with copywriting and I see a tendency that either people go for the kind of very usability angle on it where you go, “Click here. Buy now,” I mean, it might as well just say, “This is a button. This is a sign-up form. Use it.” And the other end of the spectrum is, you know, people going, like, becoming overly creative with creative messaging that’s super sexy but is really, really hard to understand.
So, what I’ve found from, yeah, several hundred tests on copy is that the more value you can convey via your copy, the more relevant it is to the target audience, the more conversions you’re going to get. And, I mean, I do a lot of tests where I go, for example, running a test against an advertising agency version, for example, that’s often going to be super creative and really, really sexy. But, the problem is that it usually doesn’t really entail a specific benefit so I find that super boring copy can outperform super creative and sexy copy by several hundred percent as long as it actually is clear and conveys a specific value, say, 50 percent or whatever. So, I think that’s a very big problem, actually. People getting clear about what they’re offering and what value, what specific value it will represent to their potential customers. I think that’s the most important thing.
That also crosses over to design. I mean, for example, I see a lot of websites using a lot of imagery because they’re learned that using images is important but then they just, you know, buy stock photos that have maybe a completely different meaning or that really don’t have much to do with the product at hand. I mean, recently I saw a landing page that was selling I think it was like a financial audit and there’s a picture of a girl with a snowboard, and you’re like, “Well, pfft!” Often, it’s good to have people on your web pages but I don’t see the connection between a girl with a snowboard and the product here so, you know, a way to make the offer more relevant would be to use a different image.
That’s the main problem I see and I think the second one has to do with information architecture and usability — I mean, the way you present the information — because, in a lot of cases, it’s really not very logical, you know, going from A to B to C to D on the website can be really, really challenging and difficult sometimes. So, that’s one of the problems also I see. And I think it all pretty much comes down to the fact that I think a lot of marketers, I wouldn’t use the word lazy but it seems like a lot of marketers are kind of expecting that the target audience or their potential users are just dying to do whatever it is they want them to do. It’s almost that they’re assuming that people are just, you know, they can hardly sleep at night because the only thing they want to do is get up and click banners, or buy stuff, or sign up for newsletters, because it kind of seems that way because, you know, they’re like, “Oh, we’ll just put it on the website and people will do whatever we want them to do. We don’t even have to explain why they have to do it or make it easy for them to find.” And, I mean, I try to go the other way and I try to assume that nobody — nobody in the world — wants to do what I tell them to do so I really have to make it as clear and simple and obvious as possible. “This is the reason why you will benefit from doing whatever it is. Here is the next step, and from there, this is the next step.” Kind of an oversimplification but I hope it drives the point home.
Francis: Did you say you tested against a creative agency?
Michael: Yeah. I mean, a couple of my clients, they have advertising agencies, you know, doing their kind of overall identity and stuff and I’ve, for several different clients, I’ve just performed, for example, landing page tests where I’ve, you know, tested, for example, my headline against their, or different banners, or different overall variants of landing pages. And, generally, what I see is that, online anyway — I don’t have much experience with offline advertising so perhaps, I mean, perhaps their creative messages work really, really well offline but, you know, and I had a lot of experience optimizing websites and making them sell and all my experience and all my tests tell me that, you know, the direct clear message will beat the creative message, or the blatant order, nine times out of ten.
And I don’t want to generalize because, I mean, creative messages can be really, really good. But it shouldn’t be the default. You should go for a creative message if that’s what your hypothesis tells you, if that’s what data and experience tells you will be the right thing in this particular case. But I see a tendency that people forget everything kind of about how you would normally approach a selling situation or talk to a human being when they have to write marketing copy and, all of a sudden, it becomes this crazy stuff that you would never hear in the real world, you know?
Okay. Let me give you an example. The other day — this is not something I’ve been involved with — but I just saw an ad the other day, this is translated from Danish, but it says, “Tomato or broccoli? We love beef.” That’s an ad for a steakhouse and I’m like, “I have no idea what that means.” I mean, would you ever say that if a person came into your steakhouse to buy a steak? Instead of saying hello, you go, “Hey! Broccoli or tomato? We love beef!” I mean, those people would just look at you like you’re an idiot, you know?
Francis: Well, it’s a lot of what I call marketing by hypnosis. Basically, you try to hypnotize the visitor or the prospect into buying your stuff.
Michael: Yeah. And, you know, that probably has its justification here, I’ve just never seen it work online and, I mean, I would personally wonder if it ever really works because, I mean, it might. It’s just kind of this maddening approaching to marketing where it’s like, “Oh, this is voodoo and people don’t know what they want. We know what they want and you don’t know how to do marketing because we have the creative superpowers to make people listen,” and I think that, probably in the sixties when there wasn’t that much advertising going on, just the fact that you had advertisement was enough to boost your brand. But today, I mean, especially online, there’s an abundance — there’s a ridiculous amount — of different offers, all the time! And, I mean, very, very few products out there, or companies, don’t have competitors and, you know, I mean, you probably know from your own behavior on the internet when you’re looking for a solution to your problem, you know, I mean, I bet you that you do a lot of research and then choose the one that seems the most relevant that has the most credible solution to your problems, and not the one that has the loudest or most creative messaging.
Francis: Yeah, I find it quite fascinating as well. Well, I realize that there are — well, in my experience — there are, like, two sides. So, like, there are the creative agency people and they’re, like, the fancy slogans, the fancy graphics. And, on the other side, where most of the optimizers are working in because we know it works is we subscribe to a more direct marketing approach. So, I’ve actually spoken to some creative agency people. I mean, this is a generalization — they are not all like that and a lot of them are very, very talented people.
Michael: Oh, yeah, yeah, of course. No, I would never say that they weren’t.
Francis: But, like, there are some people who have this opinion that direct marketing is dead, “You’re not going to sell anything,” and I’m like, “Well, but that’s not what my tests say,” you know? Every day, day in, day out, I create more sales and revenue for my clients.
Michael: Yeah, but I think sometimes it depends on how you’re interpreting the definition “direct marketing” because I think a lot of people think of direct marketing as these long, long, long, long sales letters that you send physically by mail and, I mean, that is the classical form of direct marketing but direct marketing is a lot of different things also, I think, nowadays. And, I think that, of course, online marketing has a lot to do with direct marketing, but I think that there’s a lot of different ways of doing it — there’s a lot of different forms. But I agree that the direct approach is very interesting, but also, I think that a lot of people take it to the extreme and, instead of respecting the decision-making process that people have to go through in order to make the decision, for example, to buy, they jump that process completely and go straight to the “Buy now!” kind of solution. And, if that’s kind of the take on direct marketing then I agree that direct marketing doesn’t work.
Francis: I think they just don’t see the effort that took the process, that took the user all the way from the start of the funnel to the “Buy now!” button.
Michael: Yeah, exactly. And, again, I would say it depends 100 percent on what you’re offering. I mean, if you’re trying to sell someone a jet plane online, well, you’re probably not going to sell that jet plane just by having a one-page landing page with 200-words and a call to action above the fold. You know, you’re probably going to have to have a pretty extensive content marketing strategy in order to make that sale. But, I mean, if you’re selling an ingenious iPhone app that costs $1.00, well, then the right approach would probably be more of the “Buy now!” with 200 words on the page and a call to action above the fold. So, I think it’s really, really important to have a clear idea of what you’re selling and who you want to buy it and then find out what are the mechanics that work, especially or in this particular situation.
I think that’s one of the main problems because you see a lot of, for example, you have a lot of copywriters that would either say, “You should always write long-form,” or, “You should always write short-form,” or, “You should always use emotions,” or, “You should never use emotions,” or, you know, “You should cut out all adjectives. Never use adjectives.” What I’m saying is, well, you can’t generalize that way. I mean, you know, you need to find how it works in every particular case and that varies a lot. And, in some cases, the creative messaging might be exactly the right thing to do. And, in some cases, the direct approach is better.
Francis: You know, the more I do my own tests and the more I talk to other people who do optimization as well, the golden or the right answer seems to be, “It depends.” Like, that’s the answer to everything.
Michael: Yeah, exactly.
Francis: And it’s so true!
Michael: Yeah, it’s tested. It depends. But I think you can, I mean, as I mentioned before, I don’t like working with rules because then you have a tendency to take the specific case and try to impose rules on their case and then, usually, that backfires. But you can use principles.
You can, for example, have a principle that, if you reduce friction, you will increase conversions, and then you can, within that area called friction, there can be a lot of different things that cause that friction, and then you will always be able to identify those by using that principle. But you can’t have a rule that always says, “Credibility, we must turn up or crank up the credibility,” because, in some cases, you might have a huge credibility level just because of your brand because it’s very well-known.
You could also have a principle that says, “The clearer your value proposition is, the clearer you can convey your value through your copy, the more conversions you will get.” You can always use that principle. But you can’t always use a principle that says, “A clear headline is only 65 characters long,” because, I mean, it might take you three times that amount to actually convey value through your headline.
Francis: Well, the rule I’d like to use for the headline is it should be as short as possible but just enough to convey your message.
Michael: Yeah, definitely, definitely. But I think, again, everything copy related, I mean, I think there’s kind of a tendency where there’s kind of text-copy anorexia and I think it’s interesting because there’s this whole thing about people don’t read online — they scan, which is an observation that holds water. I mean, I don’t read everything online. But, again, it depends on what it is and how it’s presented to me because I do a lot of reading online but it depends on the commitment level I have to put in there. If I have to spend $20,000 on something, you bet I’m going to read the copy, you know?
Francis: And the fine print.
Michael: And the fine print, exactly. So, I mean, your copy, in all cases, it’s better to have copy that’s a little bit longer but conveys a value than copy that’s too short to convey a value because then there’s no purpose in it anyway. So, it’s all about finding balance.
And there are also ways that you can build up websites and landing pages in such a fashion that you, for example, that you summarize pretty much everything above the fold in the first seven inches of the screen so you have a headline that conveys clear value, you have a sub-header that picks up and kind of validates whatever it is you’re saying in the headline, then you have some bullet points or a clear image and a little bit of extra information and then maybe a call to action. Then, you have everything summarized up there for people who are ready to make a quick decision and then you can, below that, you can have a lot more information for the types who need to read a lot more in order to kind of feel secure that they’re making the right situation. So, you can build up, you know, landing pages and websites so that you kind of consider different types of users also.
Francis: What you just mentioned, I just have one version of a web page where we cater to all kinds of users. Maybe there are highly motivated users who are perhaps repeat visitors to the website and this is their third time visiting and they definitely compare products and all the competitors compared to a first-time visitor who knows nothing about a product. Do you do any kind of behavioral targeting for your clients?
Michael: Yeah, but I think probably the main thing is that what I’m probably focused more on is that the channels, just channel segmentation actually, that you try to make sure that repeat visitors aren’t going to go through your AdWords ads, for example, and land on landing pages that are actually designed for first-time visitors. You want your repeat visitors, you want to get them on your newsletter, you want to get them into your loyalty program so you can segment it on channels. So, I think I spend more time on that actually making sure that you have kind of an individual landing page for each campaign and for each visitor type that’s going to go through the campaigns. So, I usually kind of take it one step back and focus on that. But, if you can segment all the channels, I’ll try to have different versions for different types of users and it depends on the software you’re using. But, I mean, with cookies and everything else, it’s not that difficult to have different versions shown to different visitor types.
Francis: What are some of the tools you’re using to help with this conversion optimization process?
Michael: Well, I mean, like I mentioned earlier, I think that testing is essential. I mean, I can’t see you doing conversion rate optimization without testing because the whole optimization part comes in because you can actually confirm that what you’re doing is an optimization.
So, a split testing tool is essential. I use Visual Website Optimizer — it’s probably my favorite tool for a bunch of different reasons. It’s very, very easy to use, especially for smaller tests and it has a lot of advanced functions also. What I’m trying to say is that you can use it at any level you’re at because it doesn’t involve very much coding. It’s reasonably priced and they have great support so it’s very easy to get in contact with the guys on the back-end.
Another tool I actually use is something called Unbounce which makes it possible to create any test landing pages without involving IT which is a dream for most marketers and optimizers. So, Unbounce is really, really cool, especially for campaign landing pages. Awesome tool, you should check it out. It makes it very, very easy to kind of just build landing pages by dragging and dropping elements without having to tweak the code so that’s a great tool that makes everything a lot easier.
A third tool, I think it’s interesting to conduct simple surveys in order to get a better impression of what the users, you know, what their needs are and what kind of information they’re lacking or which specific product features they particularly like. But my experience with the surveys is you have to be very careful because surveys are a hassle — nobody really wants to do them and, the more questions you ask, the less credible the answers are going to be because people, they don’t feel like answering a lot of questions. So, I like to use, it was KISSInsights before, they changed the name to something I can’t pronounce — Quaraloo, Quarangaloo, Quoraloo, I can’t remember.
Francis: Quarango or something.
Michael: Yeah, exactly. So, yeah.
Francis: I’ll put it up in the links somewhere.
Michael: But I love using that because it’s a very simple little survey tool and it’s actually just, you can design it any which way you want, but I like to have as a little pop-up at the bottom of the page. You can use it on your website. Generally, you can use in on landing pages and I like to just have one question and then two answer options. What I usually do is I try to find out, for example, if the three different aspects of a product and, in my mind, they seem that they could be equally weighted then I might ask, “Which of these features do you prefer?” just to get a clear idea of it, or maybe test different versions, for example, inspiration for a headline. So, I’d say, “What’s the most important to you?” Just randomly, for example, it might be to save money, or it might be to — I don’t know — feel more secure, or whatever. That way you can kind of nail down what’s important to users by creating super, super simple little surveys and collecting the data.
The cool thing about this little tool also is that you have kind of a little “Thank you” part in the form after they send it and you can link to whatever you want. So, you could, for example, link to five more questions. So, you can say, “Thank you for providing these answers! Would you be interested in answering five more questions?” so it’s not very intrusive; it’s a very polite way of gathering information.
Francis: Do you use any user testing tools?
Michael: Yeah, I do. I have a lot of Danish clients so I’ve used what’s called Brugertest.nu in Danish — their international version is Usertribe.com. I’ve used that quite a few times. It’s pretty cool. I mean, I use it strictly for, well, it’s kind of in the name — you just test this for qualitative, not quantitative, testing. But it’s quite interesting, just with five users, there’s usually a lot of small things that kind of emerge where you’re like, “Wow! I really didn’t think about that.” Just small points of friction or things that don’t seem logical or just anything that kind of slows down the process so you can get a lot of valuable insights. And the cool thing about Usertribe is that you just order the user test online, it’s reasonably cheap, and then, afterwards, you get a full report — you get videos and everything. It almost feels like you’re looking over the shoulder of your potential customers so you get a lot of insights that way.
Also, I need other tools, I mean, heat maps are super valuable. I mean, there’s also a built-in function in Visual Website Optimizer for heat maps. Heat maps are click maps that can really give you a lot of important information that you wouldn’t be kind of able to get otherwise. And, of course, basic stuff like Analytics, of course, is very important, and Excel spreadsheets.
Francis: The user testing tool you mentioned is called Usertribe, is that right? Is that only for the Danish?
Michael: That’s their international website, Usertribe.com.
Francis: T-R-I-B-E, is that right?
Michael: Yeah.
Francis: Okay. That’s interesting. I’ve got to check it out.
Michael: Yeah, do. I mean, traditionally, a user test can cost a lot and lot of money because you kind of invite people to your office and you given them cake and coffee and stuff and you have them go through your website and whatever. And it takes a lot of time, also, it’s a large investment — both money and time-wise — and, I mean, a tool like Usertribe is pretty cheap and, you know, it takes very little time because it’ll pretty much handle everything for you.
Francis: I think that, on tools, there’s this great explosion within the last six months to a year of all the tools that can really help with optimizing websites. I use Optimizely for most of my campaigns, my optimization campaigns, and I’ve seen the tool grown from like a really simple tool to, now, it can really run like full funnel campaigns without really hacking the code all that much and it really decreases the time that you take to implement a test and that’s really good for everyone, basically.
Michael: Optimizely and Visual Website Optimizer are kind of built on the same principles and I’d say yeah, that’s very important. That’s one of the main reasons why I’d rather use those tools than, for example, like Google Experiments because the other stuff, it’s just that you need pretty much either to have the develop skills yourself or you need to hire a developer to set everything up, that makes it immensely more expensive and it takes a lot more time to test.
Francis: I think beside that, I don’t think Google provides support. I think it’s called Google Experiments right now. I have no idea what it’s called. They keep renaming the thing. But I use Optimizely so I can share a bit more about that. Like, when I run a test, sometimes it’s a very complex test because I do a lot of stuff on eCommerce sites and I basically email the support team and ask a really, really specific question, “How do I get this implemented?” and they can answer and I think it might be similar… What can you say about Visual Website Optimizer? Because I don’t use that tool.
Michael: I mean, they’re just really nice guys and support is really quick. I mean, they’re in India so, I mean, if you write them while they’re sleeping, it’s going to take a few hours, but usually, you know, you can get your answer really, really quick and they’re always so much polite and they can really help you with the right insights. You know, I don’t have a lot of insight into code so, I mean, I think sometimes I ask them stupid questions but I never get the feeling that they never answer me in such a way that I get the feeling that they think I’m being a pain. So, that’s nice.
Francis: Maybe it’s because you’re famous.
Michael: Oh, I’m not that famous. I definitely wasn’t famous a year ago and they were still nice to me. So, that was pretty cool.
Francis: So, yeah, I mean, it’s really incredible the tools that have been coming out. I’ve looked at EyeQuant recently and I found it really, really quite useful.
Michael: Okay. What is that?
Francis: EyeQuant. E-Y-E-Q-U-A-N-T. So, they are like some, let me type it out. So, there are some neuroscience concepts which are used in optimization. The thing is like you see what kind of elements grab the attention of the visitor and this tool has somehow made that into an algorithm and is able to, basically, it’s a heat map but you don’t have to run the heat map test and it just uses the algorithm to determine which parts of the page will call out to the user more strongly. And I did their trial on a website and I found it, like, “This could be really useful.” So, stuff like that which wasn’t available before.
Michael: So, is the algorithm, is that before you launch it that it kind of starts to calculate what will be viewed most or interacted with most? Or is it based on data so you launch the page?
Francis: It’s not based on data. It’s based on, basically, they have a couple of neuroscientists and they made their knowledge into an algorithm, into a model and into an algorithm, and they use that to analyze your web page.
Michael: Okay.
Francis: And it’s shockingly, well, useful.
Michael: Well, interesting. I’ll check that out.
Francis: For a lack of a better word, yeah.
Michael: Yeah, yeah, I have to check that out. I mean, my only concern there would be that it’s based on kind of best practice and assumptions and not data from the specific individual case study or project that you’re working on, and it’s not like an interesting way of kind of like a user test, you know, getting some basic ideas down and hypothesis.
Francis: I have no idea how it works. If you get to talk to them and I should do that because I ran a test on a page that I already optimized for a client and, well, the tool basically shows that certain elements that I wanted to show to draw attention to, the tool showed that this is what a normal human should be drawn to based on their model. So, I mean, stuff like that. This is cool, you know? Stuff like this, a year ago or maybe two years ago, wasn’t available.
Michael: Yeah.
Francis: So, the future is really bright for this optimization thing.
Michael: Yeah, yeah, yeah, definitely, definitely. And, I mean, it seems like an obvious development because, when I first started getting into the online thing, I mean, even though it wasn’t that long ago — and this was in Denmark, of course — I mean, we were at the point where we were just trying to teach businesses that there’s actually money to be made online. And then, the next step was kind of that businesses found out that, “Wow! This traffic thing is incredible! We need SEO, traffic traffic traffic! PPC, traffic traffic!” and then, you know, now people are coming to realize, saying, “Hmm. Maybe it makes sense that we can actually convert this traffic into paying customers.” So, it seems like a logical next step to work on.
Francis: I think it’s also because traffic’s so expensive these days. Like, you know, in the day of five-cent AdWords clicks, it doesn’t really matter whether you’re copy converts at 0.05 percent, it’s okay, cheap.
Michael: That’s a very, very good point.
Francis: Now, each click costs,, like, $0.50, $1.00, or some clicks cost even, like, $10.00 or $15.00 per click. You know, so, every cent counts right now.
Michael: Exactly. You want those clicks to convert, definitely.
Francis: What is your top actionable tip that our listeners can take away for improving conversions on a website?
Michael: Probably, I think the most important one — and this might not be quite as hands-on you want it to be, this is more like in the right approach but I think it’s so important — I was touching on it earlier but I think the main thing is that you focus on the decision that you’re trying to optimize because, like I was touching on before, conversion rate optimization is not really about optimizing web pages; it’s about optimizing decision and actions. We have a tendency to sit down and look at a page and say, “How can I make this a better page? How can I make this copy sound better?” But the point is never to have to make the page look better or have the copy sound better. The point is to get more people to make the right decision. So, you need to focus on the decision and you need to focus on how you can improve or how you can make the web page help you optimize those decisions and actions. “How can I make the copy, how can I write the copy in such a way that I get as many people as possible to do what I want them to do?”
So, if you’re, for example, sitting down and you are going to write copy for a sign-up form then you need to focus on, “How do I get as many signups as possible?” and not, “How do I make this copy sound as good as possible?” I think that’s probably the number one piece of advice I would try to instill upon a young frail mind.
Francis: So, if people want to get in touch with you, will you be at any conferences or anything like that?
Michael: Yeah! I’ll be speaking at a couple of different summits in Denmark and I’m going to one in Estonia and another one in Latvia, I think. But I’ll also be speaking at the MarketingSherpa Marketing Experiments optimization summit in Boston in May on the 21st of May. I mean, I’ve spoken at several MarketingSherpa events and I’ve attended a lot of them just as a normal attendee, and I recommend it to anyone who wants to get into optimization. It’s really, really a learning experience, some valuable couple of days, there’s a great networking, and they have one-to-one clinics where you can get in and you can have some of their analysts help you out with your whatever problems you’re having in relation to conversion rate optimization. There’s tons of sessions. I’m going to be talking about how to optimize copy for conversions and there’s round-table sessions and there’s live optimization. So, yeah, really, really great events and I’d recommend it to anyone who wants to learn more about conversion optimization.
Francis: So, where can people find out more about you or get in touch with you?
Michael: I invite people to check out my blog, ContentVerve.com — that’s kind of the main spot. I have case studies and articles and videos and podcasts going on there and I update it regularly. I’ve been very busy lately so I haven’t been updating it quite as much as I would like to but I also, my style is to write kind of, not as frequently but then have, like, very, very deep, delving articles where maybe it’s 3,000 words where I really go deep, deep into one particular issue and just show a lot of case studies in order to make it as tangible as possible.
You could also check out, I mean, I regularly write for different blogs. For example, Unbounce.com, I usually have a monthly blog post on their blog and sometimes I write for KISSmetrics and different other channels.
Francis: So, thank you once again for joining our podcast and we’ll talk again soon.
Michael: Well, thank you for having me!
You Can Help Us Reach More Listeners
Your reviews and ratings on iTunes help us reach more listeners and get more quality guests to give you, our listeners, even more valuable content on conversion optimization.
Click here to leave us a review on iTunes now.