During last year’s InboundCon, Chris Stolz (Manager of eCommerce Web Analytics & Optimization) gave some amazing CRO tips in his talk, Optimization Advice You’ll Actually Use.
Chris created the entire analytics operation from scratch at Hudon’s Bay and Lord & Taylor, so don’t miss out on his enlightening CRO talk below.
Also, make sure you don’t miss out on this year’s InboundCon on October 08. Tickets are currently 50% off with our limited quantity Super Early Bird Special, register now before they’re all gone.
So it has been a long day. We’ve learned a lot of stuff. I thought I would probably have the most jam-packed deck that you’re going to see. It’s pretty packed. I got slightly more slides than I have minutes, so we’ll rock through it. First just for fun, I’m going to do something totally random and I need you guys to do it with me and this is totally just because it’s fun. It has nothing to do with anything. Everybody thumbs up, bring your fingers, palm facing each other. Turn your hands down, so your thumbs are down, your palms are facing up, took a right hand lift at above your left hand, cross it over, link your fingers and gently squeeze. Did I get the whole room? There is one guy at the back that didn’t do. All right, we all did the same thing to get here. Let’s see if we all end the same way, on the count of three, one, two, three. Perfect. If you Google my name, you will see I’m magician two, so I have to do something or everyone was going to ask me to do car trick at the end of night.
So there you go, you’re all satisfied. This title is interesting because when I was talking to Joe, I said, “What should we do.” And he said, “Well, a lot of the things that were pitched to us we realize was optimization advice that I could Google and I would tell you right now what the first 20 slides are going to be.” I said, “All right. What if I will do 45 minutes of as little fluff as I can possibly jam-pack into a PowerPoint?” So we came up with, “All right, well that’s optimization will actually is and that’s what we called it.” It’s not trying to be arrogant, I promise. I have done a lot of stuff. I come from agency originally, now on client side with four different brands, so I’m doing all the analytics testing, products cross sell for Hudson Baylor and Taylor Saxon [SP]. I am fifth right now and I do a little bit of consulting and different stuff on the side. We don’t need to read this, let’s keep moving.
These are the kind of five areas I’m going to touch on, best practices; a few comments about best practices get us going. We’ll talk high-level strategy, a little bit of organization of optimization and testing in particular. I’ll start talking a little bit about calling conversion pitfalls. I say this carefully. It doesn’t mean don’t try them. It just means don’t launch them on your website tomorrow for 100% of your traffic necessarily. I’m going to get into a whole bunch of conversion wins and test ideas, different things I’ve done, different things I’ve tried, different things I think work worth exploring and then I’ll get into what I called the ultimate CRO strategy. So the ultimate CRO strategy, there’s a lot of stuff on here. By the way marketingleader.ca/inbound is where you can go to download this thing. They may pass it out as well but if you want for me, I’ve also got another one here called ten ways to make analytics actionable so go get it if you want it.
You don’t have to enter your real e-mail address. If you want, give me some fake, I don’t care. But the last line here really says that, “My hope is that the ultimate strategy will give you on one slide the best way to understand your customers, know which tests to run, know what to test, feel confident before you run the test, you’re going to get a good result. It is best way that I have ever found so far. I’m young guy, got a lot of ways to go to optimize, so we will get to that. Best practices, so everybody’s got a top navigation with…who’s got the top nav with a drop-down menu on it? Most of you? Who’s just got a top nav that just runs along the top of their website? Who’s got the left nav? Who’s got the right navigation on their website? Ooh, we got one. GShift is doing something different. Best practice states that you should go right. I come to split on the road, ah everybody else should go right, I should go that way.
Amazon gives free shipping, so that means people are programmed to expect free shipping from me, I should give free shipping. In a steering committee or in a meeting or a kickoff when somebody uses the phrase will best practices that we should what I find that usually means is well what I want is. I challenge you to find somebody state a best practice if it’s not what they truly want. I’m going to turn the best practice thing on his head to redesign the site. Go where your customers are, figure out what your customers are, doing what they want, what they expect, test, change things, try things, look for changes and improvements. Don’t just offer free shipping because Amazon is and that’s the best practice now and I’ll show you later that it’s not. Before we go too crazy, I come right now from a great ecom centric world; you’ll hear a lot of e-commerce things. How many of us sell something online right now? Ecom, hands up high, how many we have got?
About a third of us, okay. So we have got two different metrics, we’re talking CRO conversion rate optimization. From the conversion rate standpoint, it’s very much a dizzy, orders divided by visits, email sign ups or downloads or whatever my goal is divided by visits. Okay. As we heard earlier, there’s a lot of goals in between but it is A to Z. RPV from an ecom perspective is, I think of it as a hybrid between average order value and conversion rate. So when you’re actually taking money on the internet, it’s a nice way to measure, am I increasing my conversion rate without our losing money. I can increase the conversion rate for the bay.com like crazy tomorrow, but I’ll probably draw up a revenue way down to nothing. I can do all kinds of crazy stuff to increase conversion rates. I have to keep sales up, so that’s why we use RPV. If you use things like Adobe Test and Target, you will see that’s usually the success metric that they use for test. Here we go. Real simple.
If you got analytics’ team, if you’re looking at data, we’ve heard a little bit about this before. It’s a cycle data. I have a guy, his name is Nathan. He is my data integrity guy; a 100% of his job is to make sure that my data is good. That’s it. I have people who do reporting. Every Monday CEO gets his reporting as a dashboard. I have dedicated analysts whose job it is to dig into the data and understand what Nathan gave them, and finally I have optimization. It’s a cycle, it goes around and around to get my data and my reports, analyze, optimize, goes around. It has to be a cycle. It can’t just be different stuff all over the place. Those four guys have to meet and talk and trade ideas and communicate. So somebody asked me a little bit earlier about testing it and what should I do first. This is how I look at it. I break it in two pieces. I have user experience. I have persuasion. User experience is roadblocks along the way, what’s getting me here, what do I have to get pass, what do I have to go through.
Persuasion is is it save $50 or is it 50% off. That’s persuasive marketing copy. I’m playing with the message. I’m playing with what influences you. I usually recommend especially if you’re just starting out, do the user experience stuff first, fix your checkout, make sure that it’s not confusing where to enter your credit card number or your e-mail address, never ask for the same information twice. Get them through that process first. You will get some really good wins in their then start learning and playing with your marketing copy. So this is in your handout as well. This is what I do. It is super agile, it’s super simple; we’re actually revising our process a little bit. Number one; evaluate the actual impact of the page. A lot of people will decide to run a test on a pager, a step without actually looking at does it generate any revenue now, do I have enough traffic and conversions. I spoke to someone in the Genius Bar earlier and her concern was she didn’t have enough traffic or conversions to even run a test which is a huge challenge, so look at that.
If I’m not going to get my 200 conversions in two weeks, I mean I have to run this thing for a long time to make sure I get my significance and my data normalizes over time. Sort all of your tests based on reasonable potential impact and the effort to create it. Take your quick wins. Your quick wins our low effort, high impact. Take your stuff that needs development, let that get work on in the background and just start going nuts on your quick wins for a while. We spent eight months. I think we ran 20 or 30 tests. They were all quick wins, in the background we build some of the bigger stuff that we ran. Really, really quickly and then we’re going to get into some of the fun stuff. We’ve seen this in a couple different forms today already. This is our marketing funnel. Landing page, product page, cart review, information billing, confirmation. Moving through different levels of intent, so the typical assumption is the further they go, the more they must want this, the higher intent it is and in a lot of way that’s very, very true.
I want to offer another way to also look at this, based on levels of commitment because when we’re testing and trying to look at abandonment and abandonment rates, looking at it from this perspective totally changes it. We’ll optimize way high up in the funnel and we will throw every test we can in the lending page. There is barely any level of commitment at this point, and there’s barely level of intent, so it’s not passing either checks. So think of it term of from an e-commerce perspective it’s pretty easy. The billing pages where I want your credit card. It’s the place that I’m asking for the most precious thing that I’m going to ask for. It’s definitely a good place to start testing and playing around. This is my number one pet peeve, the UI UX guys that were talking earlier; we will hopefully agree multiple navigation methods.
Every single time a customer touches something, they have to learn how it works. If I had to encounter a different one of these for each slide that I have to present, I have to keep constantly relearning and trying to figure it out. Even though I have used these one before, this one is slightly different, so I have to kind of understand which buttons was which. Just because a customer’s seen a slider on somebody else’s website or carousel or whatever you want to call it, doesn’t mean that your customers knows how to use it, doesn’t mean that it works for your customers.
Take a look step back and look your pages and think how many different ways to reveal or explore content am I giving them. Almost any page on a lot of websites you will have accordance, sliders, and links and buttons and dropdowns and all these different things. Interactive content, if you have a whole bunch of it, a whole bunch of different styles doesn’t necessarily mean that it’s better. It may mean that it is more confusing, so a really good test to play with is just to streamline your navigation methods on a page and see what happens. This is what I’ve seen a lot of when running tests on sliders. I have run a whole bunch of these things. They almost always lose the static images.
One of the reasons is because you never know at what time, how many seconds it’s going to be before their attention moves to the slider and so when the slide changes, oh I was halfway through reading and so we set five seconds on the thing because we assume that that’s the first thing they look at and they need five seconds to read it, but really I looked over at the logo and scanned across the menu and went, “Oh that cut mine and it changed.” Over 90% of the time they click on the first slide only. Again this is just from my test, you’re going to want to try your own but a lot of the time, the reason why it either doesn’t make a difference or outperforms it is because they’re only clicking on number one anyhow so, let’s just show them the number one and then result down below. I’ve got a 98% confidence rate with a 16% left than just getting rid of the slider and just putting those two images side by side.
I can see both options. I can compare them, which is a beautiful thing for the brain to do and I can make a decision. We’ve heard a lot about this and it seems to actually becoming up on Twitter all throughout the day about being creepy with remarketing. Personalization is the same kind of idea. I hear a lot of people say, “Oh man, let’s just personal everything because Amazon does.” First of all Amazon doesn’t. They personalized key elements and they play with what those are all the time. But could you imagine a little example, I’ve got on the slide here. Going to a website say, “Welcome to our website. Did you know that you left the garage door open this morning?” Some people are getting close.
I’ve visited a website not long ago where the personalization was so heavy that after I’ve view six or seven pages, I was only seeing the same content over and over and over again because that’s what I thought is interesting. Give them the opportunity to discover something that they didn’t expect to see today as well as things that you know about them. Sorry design guys, I apologize for this one. Color definitions are not necessarily a best practice for web design. Somebody said to me recently, you can’t make the button read because red means stop, so I said send me the link where you got that from. And I am like, red also means love and sale and passion. Articles is good, shouldn’t I make it read. Test them; don’t just assume that red’s going to be the magic thing. If it’s Valentine’s Day and you go with something bright blue, you’re doing something different. Try it. The colors don’t necessarily drive it.
Focusing too much on lading pages, I have already mentioned this; they’re always going to hide the highest abandonment rates and bounce rates, it’s the entry of the funnel. It’s where you get the most traffic. So of course you’re going to have the biggest rate, people are dropping off. It’s their first piece of contact with you. Think lower down the funnel. If you’re not getting yourself deeper in the funnel, try it; get down there because you’ll start to find some really nice wins down there. Expecting huge results from tiny, tiny changes. How many of us have seen that same green button versus orange button test example every time we do A/B test ideas all over the Internet? The one is that the Gmail one, get started versus sign up now. We keep seeing it everywhere and we assume that these small changes always mean big wins. They can but they don’t always. Small changes can make huge difference if the change is meaningful.
What I do is especially early on in testing for somebody as I stand back and literally stand on my chair and go 12 feet away and look at the screen. And if I can’t see the difference between A and B, it’s less likely that I’m going to generate a very big emotional response from the change if I can’t even see it, so something to think about. Security logos are not always better. Everybody says hacker safe logos and PayPal’s secure logos and all the different stuff. A lot of time they work and they work really well in ecom but they don’t always, so we did this actually at a company I work for, I don’t know, eight, nine years ago. We put a hacker safe that was in the header of the website, was on every single page. RPV went down; call center volumes went through the roof because people saw the word hacker. When you click on it, it takes you to another website that’s talking about 128 bit encrypted and blah, blah, blah and it wasn’t our domain anymore and their site didn’t have secure logo and they freaked out and people were calling to say that we were hacked.
So we took it down and sales went back to normal. Offering free shipping automatically, this is a just kind of a fun little nugget. So let’s say free shipping on orders of $99, okay, which is what the bay does today. If you automatically just release that, the moment you hit $99, you’re going to pay for every single customer for their shipping who gets over $99, which is from a customer perspective is a great thing. We tried this on one of our other brands and we say, “Well, what if we just offer it with the promo code?” And we message it very clearly, but we ask them to take the extra step and enter it with a promo code.
What we found was A, we made a heck of a lot more money from everybody who forgot to enter the promo code, but our repeat shoppers didn’t go down. So one of the assumptions of the company was, “Well if we make people take the extra step, we’re going to lose customers.” We didn’t. Our repeat business didn’t change. What it ends up is happening is that next time they come back they remember and they always remember they have to go and look out there and see what the promo code is and answer it. So just think about that. There’s a lot of stuff that we release automatically that we give out. This was worth millions of dollars, millions of dollars. Testing just to see if this works, so when I have testing meetings, a lot of the time the ideas that come up are, we’re spending a lot of time on this banner, we’re spending a lot of time on this content and this content and this content. Can we run an A/B test to find out if it works?
Instead of finding out what my customers think and what they’re doing and what’s working, what’s not, where they’re abandoning, I’m just going to go grab some random things that I’m sick of spending two hours every Monday doing and I am going to A/B test just to find if it works. It’s a bit of priorities in the mix and it happens a lot. So how I am doing for time? What time it is? Someone shoot me the time. All right, I’m doing okay. Actually I did pretty well, so conversion wins and test ideas. So here are some things I have tried, some things you can try, some things you can take away. Clarity, don’t make them think ever. It goes back to navigation method thing. Don’t just assume, well people aren’t stupid. They will figure it out.
Assume people are stupid, they won’t figure this out. There is actually a web book out there and I believe it is actually called Your Customers Are Stupid and it’s very, very true, ad copy, headlines, call to action, clear, simple and most importantly driving emotional response. I want that. ALS drove an emotional, right? It got right into your heart. The heartstrings, the speeches people give before they dump the buckets, not everybody was as goofy as I was, a shot for shot remake of the training sequence in the Rocky movie, training with ice cubes and push-ups over plate of ice cubes but most people took it really, really, seriously. Complex call to actions almost always lose to really simple ones. There we go. I’m going to break my own rule, here’s simple small change button test. Take the readiness assessment starter and get started. Okay. The one on the left tells you exactly what to expect on the other end of the button, which is usually a rule that I set.
If someone has got to click on them, try to teach them what to expect when they click on it right so that I know what my destination is more likely to want that and click on it. I’ll try and get started, sure, what is going to happen. Well get started got a conversion rate of 16.3 for every seven leads with the old button. The new button generated 23. That’s a great little win. I didn’t think that I was getting huge win on a change like that. I was actually most wrong on that test than anybody else who ran it. But the reason why we think it works is because it’s just a lot more simple and clear, and with the readiness assessment starter, I have to think with get started, it’s kind of like, it’s a way of saying have you had enough content, great, click on this.
Show and hide major elements. I try to find a picture of my 2-year-old daughter for this one doing the same thing that I couldn’t. So play around with whether or not you know something is pushing products or pushing your content down the hole, sliders and giant home page banners, especially the really big one that’s full page, unless they got a really good message in a call to action play around with whether or not you need them, and whether or not they’re actually hurting you.
There’s a lot of people who see the full-page ones and don’t know to scroll. As much as we know that in our heart of heart as marketers we think people are used to scrolling now and that’s a new thing and we don’t have to make it 800 to 600 anymore. We can make it nice and long, turns out with a lot of testing some websites with some customers that’s not true. They want to see a little bit of the content to come so they know to scroll. So this is one where we took a banner above the category. We threw 50,000 or so visitors through it, 6,000 conversions and we said what’s going to happen when I take this banner away and they were so committed to this thing, turns out it’s flat.
It’s a 0.73% left, plus or minus 1 or 2%, I call flat. Perfect your checkout, this one goes without saying, right? We all knew this, make sure checkout is the most beautiful streamlined checkout in the world. I like checkouts with nice big fields so that it doesn’t look really busy, but it looks really simple and clean. Domino’s Pizza has a really good checkout. Fiverr.com is a good one. Does everybody know about Fiverr.com? Does anybody not know about Fiverr.com? F-I-V-E-R-R people who do things for 5 bucks. You can get SEO audits, you can get content written and they’re good, they are shockingly good at times. I will use the phrase bitchwork [SP], it’s a stuff that you need to report monkey to go make an analytics dashboard for us so here, here’s 5 bucks, go make it. They got a really nice checkout ad they are doing one click checkout here as well. Here’s a bunch of things, remove distractions, remove pages, keep showing me, this is a giant one.
Has anyone heard the phrase maintaining scent? It smells the same no matter where you are right, I’m sitting in my office and I smell coffee in the kitchen and I get halfway there and my wife has actually just made her one cup and left and I can’t smell anymore. I don’t have time for this. I’ll go back to my desk and work but if I can continue to smell, I’ll keep going because that is what’s getting me to the end goal. Huge one, most retailers don’t do is when I hit my checkout and it says, this flat screen TV is $1500 cross out $1000. Once I start entering my information on the checkout page with my billing and everything, I can’t see the deal anymore, it’s either not showing me the price I’m paying or most importantly it’s actually the deal I’m getting. Keep showing me the deal, one of the easiest ways to do this is when they have a little summary of your cart with your price that always sits there on the right hand side. Keep reinforcing it as I go along.
Auto sort by popularity, this goes back to giving people what they want, not when you want them to have. Take your menus simple exercise run AV test on this, take your menu, run a pages report, find out which pages people visit the most, sort your menu by that AV test and see what we get. A lot of the time you’ll find that if you sort by popularity that you’re going to get a better result. The other way that a lot of people are doing it or actually testing it of stocks right now is open versus collapse left navigations, so we’re going to collapse all the ones that people don’t care about as much from this page, and within women’s we’re going to open dresses and shoes and we’ll leave all the other stuff close because it’s not as popular. We’ll just show the most popular stuff. Test isn’t done yet, but it’s working like crazy. I love this. This is one of my favorite little physiological marketing tricks; you may have seen this on marketing blogs before. I love this. This is a support ticket website. I made this up. They have got a 3-package plan.
Everybody sees this everywhere right? You’re paying for my service so it’s got the three boxes with all the different features. The Lean gives me 10 tickets, five users, really basic reporting and forum report. Standard is pretty close in price, 22.95 but again 20 times more tickets and ten times more users and full everything. It’s a great package, super mega extreme, 5000 tickets, 500 users, everything you can possibly imagine under the sun and it’s twice the price as the one in the middle. Now one of the assumptions that the marketers often make and they did this whole study on this was that people pick the middle one because it’s in the middle. So they started playing around with which order they were in, turns out it didn’t matter which position it was in visually and people started to assume well they pick the middle one because it is just the middle price and it’s called Standard.
Turns out the theory that they played it seems to work. The Lean and Standard are very similar in price, but one of them is clearly better than the other, so those two are called the comparables. My brain likes to compare only two things at a time and it like it to be easy, comparing those two is really easy. Right for an extra three bucks a month, I get ten times as much stuff. The decoy is the super mega extreme which hey if anybody signs up for a score because it probably doesn’t cost me much more offer but what it does is that it takes itself out of the picture and forces me to compare the other two that I can compare and I’m going to pick the better deal of the two. I use the terminology forcefully funnel people down the path but it’s a fun one to play with too. This is another big one. Actually, this is the last one of these, automate product recommendations. If you are running an ecom site, get yourself a good product upsell tool. So I’m sitting on a product page.
You may also like or we just say we tested the winner was, you’ll also love in our case. We did this, we ran a test, we ran at a two major huge retailers, Manual, which is selected by merchandisers, marketing and buyers versus automatic. Sitting in meetings every single department in both companies fought me on this. The only reason why we got to run it is because the senior vice president had already paid for it so we better darn use it. Well, I need to be able to inject my own into that, so could you automate two of the three and then the third one can be me because Calvin Klein made this bikini top and this bikini bottom to match and that’s what people are going to want, so I should put them together. And my argument was well if that’s what people want, they will buy them together and it will recommend them together. You got a 4.14% RPV left which for a smaller site is not a huge lift. For a multimillion dollar ecom site that’s a minimum of $5 million without including Christmas and seasonality. It’s a big win and those apartments soon started to see some interesting things.
They saw two things. One, the things they desperately wanted me to overrule were happening organically anyway, and two, the stuff that got recommended together was the weirdest crap we’ve ever seen and never would’ve done ourselves. I was looking at, and this was just after we ran the test and I started to discover that cookware sets were always selling with exact same medium tan colored sports bra and that sports bra was selling with all different types of cookware sets. Women with that style in those dimensions really wanted cookware sets for some reason, right? With the cookware set, I would put the pot holders and the dishes and stuff that matched the Jamie Oliver collection or whatever it is and with the medium sports bra, I might do bottoms or something, and other stuff to go with it, right? I never would have put the two together but holy cow they just put together. You’re never going to do that if you’re fully manual.
Most of the box stuff, if you’re on WordPress and you got WPE.com or WooCommerce or VirtueMart and Joomla or any other many different ones out there, a lot of time it’s just going to give you these based on popularity. I will get that to that in a second, just a couple of things last about automation, the biggest thing here is that it’s about letting go of control, right? Everybody fought me because they wanted control because they were so scared that it was to do something wrong. No matter how badly you want to show the blue shoes with the blue shirt, you’re guessing. Regardless of whether or not Calvin Klein made them together or you think people are going to pair them together or they go well together and you just know it, you’re guessing. It’s an educated guess but it’s a guess.
When compared to a sophisticated algorithm, you’re going to lose. You’re a beginner chess player up against the master. Your customers know better than you are and that’s something that you have to let go of. Your customers are smart than you, they just are. The nice little win at the bottom here, it’s saved both organizations, a ton. They had two full time people doing cross sell with 30 people contributing to it, and then they would do the work. Now those two people are doing way more rival things with the company within the department. It took a ton of time away. We don’t have to even think and it’s smarter than us. You can’t get much better than that. As I mentioned, don’t just think about popularity or product selling together, different algorithms are going to work for different sites and different places, so if you’re looking at a tool, get one that has different combinations.
This is using Adobe recommendations. Monetate has a good one too actually as well as RichRelevance, so we tested four different affinities view purchase, so people who look at these shoes went on to buy these shoes. Maybe they were cheaper or more expensive or just different. View affinity, people who viewed this also viewed this other one. Purchase affinity, people who bought these shoes, bought this belt to go with it in the same purchase and previously viewed items, which is what Amazon tends to do a lot, not always. What else have I looked at? Personalization for me during my visit today, so I can just kind of reevaluate my choice based on other stuff I’ve looked at. In the United States, it was purchase affinity. In Canada, it was the view purchase affinity and we keep finding this thing about Canadians, we’re browsers.
Google did a study I wanted to say two years ago where they found that for every dollar Canadians spend online, they go and spend 34 in a store for that same company. I believe the United States was something like one to 10 or 12. It was a lot closer of the gap. Canadians are browsers. We like to look online, but we’re still not totally into the whole buying online things so we go to the store and we try on, right? I’m a tall skinny guy, so I have to actually go and try a suit on. I can’t buy suits on the internet, it doesn’t work. All right. Let’s finish up with this guy, the ultimate CRO strategy, so after all the stuff you’ve heard, how can you find you key weaknesses in your funnel, how can you pinpoint confusing steps, identify which elements actually did get them to buy. I got a whole page of stuff, which thing was it that made them purchases, what persuaded them? How can I predict the outcome of change, how do I know which test run that I can feel confident will get a good result?
Any guesses? It’s actually been leaked a couple of times today and I was little annoyed that everyone was saying it all afternoon. How do you know what your customers are thinking? Tada, ask them, surveys, talk to them, I love this stupid image but I don’t know where I found this. When we don’t ask, we guess, so in this image, he took a whole bunch of scientists and blindfolded them and they’ll think it’s something else, right? It’s a sphere, it’s a fan, it’s a wall, it’s a rope, it’s a tree, it’s a snake. They have got all of their senses, except they’ve removed one sense. When we use call center data, get your a call center people ask one of 10 questions for every call and log the data, each time they do a call, they just have and go through those 10 and answered the one and collect all that data. By the way what was it about that shoe that made you want to buy? So thanks for buying our service, was there any very reason when you landed on our site that just made you kind of go yeah it looks pretty good or what was it that informed you the most, where did you go?
Just ask a bunch of these good questions. Don’t lead them. Post purchase and abandonment surveys are awesome. By the way, if you’re not doing abandoned cart e-mails as a side note start because they work really well but ask all these questions. Do you use ability studies? The 5-second test is awesome, I love it. I really like it from a perspective show the image for five seconds, have it disappeared and get them to answer a question, would you notice, keyword called that thing and start playing around. When we take a look at the difference between the typical analytics model, which is data analysis, and you test it and you try again and you try some more stuff, that analysis item is an educated guess. It’s a very educated guess. We still need to do that process if we can’t talk to our customers all day and driving nuts but when you compare to customer voice model, customer feedback what I am going to call compilation and test. Compilation is sample reality. You’re compiling the truth. With analytics, you’re taking what you think out of the data and a lot of us are really good at that.
But when you’re actually just talking to people you’re compiling the truth. Just get enough of it that you got good sample data. So I will give you one big example at the end, this was a really big win, I’m part of this. This is a Canadian breast cancer foundation. I led this project with an agency. I did the design. I did the taxonomy. I told them what to do, I did the SEO. I did the whole thing, it was really cool. That was the before, this is the after. Okay. Most of this after was because I spent about a week and a half with call center people. I spoke to their call center manager and then I said give me the lowest, most angry hates her job person on the totem pole and let me spend a day with her and I did and she bitched and bitched about all the phone calls she gets and where people are confused and what goes wrong.
This is a loud, obnoxious website. It’s by no means perfect, but it’s on brand and it works. How it works? Well we got 25.54% increase in organic traffic within a month because we build more content that their customers actually wanted. Turned out they wanted a whole bunch of medical advice so we actually convinced them to hire a couple nurses who could write to write a whole shit load of content for them, the organic rankings went through the roof, huge increase in referring site traffic. We assume that’s because of all this new good content. Nice big bounce rate increase and the Golden Nugget, of course, is a 733% increase in donations in 30 days when we launch the website.
The reason why this worked, and not every single overhaul of a website is going to be this good because it was very bad first, but I think the reason why this worked is because we just listen to what people want and what they actually wanted to do. So that is it for me. You can find me on Twitter @mr_optimize. I’ve been tweeting with people today. I think the social media panel for those who were following that conversation won the curse count. You’re doing okay and then the social media guys kind of took you. Marketingleader.ca is me/inbound has a couple resources on therefore for you and thank you very much.