Marcus: Hey Warriors! Welcome to another epic episode of the warrior podcast. We have an exciting episode ahead as we venture to the majestic world of split testing and all things to do with conversion rate optimization, data as we know is one of the foundations of almost every business online whether its custom data, analytics on the site, performance, or user experience every modern digital marketer and business relies on data. Now with me today is a dear old friend of mine and hes one of the leading experts in split testing all the way from Omniconvert is Valentin Radu, I met Valentin last year when we joined forces on partnership with Warrior forum and we build up one of our shared interest in conversion rate optimization, web design, automation, we kind a have a similar love for Adele as well, some pretty fun cold emails that we share, so Im happy to give a very warm welcome to our guest, Valentine, Welcome man!

Valentin: Thanks a lot man! Yeah, we all love Adele, we both love Adele, I dont know if everyone loves it. But Im really happy to be here and with the Warriors, I dont know if Adele is good for all the Warriors but a thing is clear, All Warriors should listen to Adele when she says 'Hello from the other side' Actually

Marcus:[laughs] actually I remember using that on the cold email when we first met, totally fun I should probably screenshot for the warriors to check out too, its pretty funny. Alright man, before we get started did you want to give our listeners a little bit information about yourself and what Omniconvert actually is?

Valentin: Yes! So Im a serial entrepreneur, Im coming from Romania, Bucharest. Ive started doing building companies 16 years ago, when I was playing StarCraft and Counterstrike and I wanted to have a broadband, because of that I ended up building an internet service provider. The first one we were Broadband connection and after that things got crazy, I build an agency, I failed it completely, and then I made a loan at the bank and then I got up in the Entrepreneurship Arena and I build the largest toll car insurance player in Romania ended up in more than 120,000 customers in 2012

Marcus: Wow! Okay,

Valentin: And then Ive become obsessed about the conversion rate optimization because we were burning a lot of money thru google ad words, we were actually having about 80 to 90% and I was so pissed off because I was paying so much money on that. And then Ive asked the developers 'guys! Help me out I want to test it' 'oh [laughs] If you want to test theres adobe testing target for you man, it only cost 12,000 bucks' 'what!' and then we started to change things with them. I was pissing them off all the time and eventually they made me a dashboard so that I can test it out. And that triggered some over layers, asking people from various cities, what stops them from converting, after that Ive increased our own conversion rate by 60% in a single year. And then I said 'thats my next business.' And here we go Omniconvert now is one of the most important conversion rate optimization platforms. Now we have around 20,000 users playing with it and I have the privilege of seeing a lot of data, so more than 1.5 billion request every month so that I can see the trends. You know, Im obsessed with data, so appealing to see what works and whats not.

Marcus: Yeah I think, its funny that you mention that you have access to 1.5 billion data points for you to analyze cause I think a lot of us, really just be drooling over the amount of data you crunch to on a regular basis. I noticed that you mentioned that when you first got your 20,000 leads with your first digital agency that you went thru was a failure. I think for us, we can relate really well, like it worries in general, particularly when we get thru the hardships of trying to build up traffic, build up leads, get people to actually come to us as the first problem and try to solve it with the conversion site. Was the ads is actually the 1st method of online marketing you used to pull in leads as well?

Valentin: Yeah, Google AdWords were the first

Marcus: Okay, what where you doing straight after that, if you dont mind me asking as well?

Valentine: Well after that we build a quiz on Facebook, so we post on the site on Facebook. We build a quiz to give you a rating about how good a driver is you. So this was actually from the test that youll do when you, in order to get your permit. And people were posting on Facebook that theyve got an A degree or B and that lead us to 200,000 fans in 2 weeks. So that was rather an important channel and after that, we made another massive thing, we realize that we cant make it no more with Google because everyone was searching for car insurance but we were selling cheap car insurance, so it was like 'cheapcarinsurance.com' but that ruined me. And we were looking at the [Inaudible] on the street then we want to make them change their minds instead of searching for car insurance to search for cheap car insurance that we were there to search. Weve made those small stickers, got 4 students, printed like 15,000 stickers then send them everywhere in the city so they can place small stickers on the garbage bins, everywhere where the eyes of the drivers could look. And in this manner, we changed the query so made from 6,000 queries per month; we got 60,000 queries per month, in just a single month. And then we became the most important online car insurance because we were the first there, no other website were doing what we were doing. That was back in 2007 I think.

Marcus: Oh wow, so that is actually 10 years ago

Valentin: That lead us to an organic, no, fantastic organic group.

Marcus: Oh okay, Well I think, I think for we just wanna jump on the split testing and everything as well. I mean, with what youre doing on Omniconvert now, youve got access to the data, youve got all the clients, youve got big name clients like Samsung and T-Mobile. So from there every growth hacker, every marketer, every marketing philanthropist will say that some ultimate split test and ideas experiment, etc. basically perform , among them, in all this claims, starting with the ugly what are the most common assumptions, mistakes people make when it comes to split testing?

Valentin: Well talking about the functions, I dont know if you that in 1950 some doctors thought that smoking cigarettes were the best way to relieve a pregnancy constipation. Actually women which were pregnant, theyve got this prescription right? Smokes on cigarettes. That was crazy but that was a bad assumptions, so I think bad assumptions are the 1st problem because people are [inaudible] instead of being data drivenand then they split test without actually understanding what they are testing and why they are testing, whats the impact? Why they are doing this test and there are lots of fabulous experts which are doing the same mistake, so for instance, I dont know if thats another funny thing, Albert Einstein, for the nuclear energy wasnt possible. Only 10 years after that, Enrico Fermi, he successfully demonstrated that self-sustaining nuclear chain reaction is possible, right? The 1st problem is they assume, they dont use the data instead they transform into a hypotheses and they are not testing based on data driven. So that is one of the common mistakes and another one is they think split testing is an easy thing so its a quick fix, they fix the conversion, they think that 'Okay! Lets do conversion rate optimization or lets do the testing' What test? So you need to have a lot of test, its like thinking that youre gonna finish with, split testing is like saying youre gonna finish with making showers. No! You got to do shower every day and thats the thing with split testing its on going and long term process so you find the best performing version of your website that can always be improve. You should decide when its not, when you reach the 9%, 4%. 5% conversion rate. Maybe they are, the costumers, which are ready to buy but you are not having the right language for them. So thats another problem, they assume this is an easy win. Another thing, another mistake is that to work, so once they start, they look for the conversion rate to grow up in a test and they dont have enough data so the statistical significance is not there, so they start the test, they see that the variation busy control 2 days after and they stop the test to find the winner. So thats another problem. And I must say, the average win rate is around 20% , so that means 4/5 test are crappy but you learn from them, sometimes you win sometimes you learn.

Marcus: I think, its interesting to see when you sort a pull out up the statistic of 20% win and 80% you know, 4/5 basically fail. From that, I think split testing conversion optimization is basically inbred by interface designers, user and experience, enthusiast, and specialist sort to say, and from that, anything above the fold is a content on the website, basically a digital play pen. Yeah, thats much better. {laughs] Well what I think when it comes to split testing a lot of interface designers and user experience, specialist and enthusiast basically theyre the one to sort of breeding the digital play pen of conversion rate optimization with anything thats above the fold. So from there, Its one giant will that keeps spinning and people are trying to improve copy, they are trying to improve bonds, layout of the website and the actual responsiveness as well has some formal plate. From there, the big question always being asked is whats the superior color of a CTA, does it exist?

Valentin: Oh that, [laughs] oh thats really simple, there is! There is only one, but is the one that wins after the test, so first assumption could be to make it distinct for the rest of the content and thats right but its not specific like green or yellow but you can make it like set gradients like purple cow, you can win, actually maybe we can test now right? So to make a button with purple cow, put the CTA on, we can test it on our own website.

Marcus: I mean I wouldnt mind testing Seth code on purple cow, I think that would actually resonate with Warriors actually [laughs] I think we could all appreciate a little bit Seth code and love. Alright, I think the other thing thats really interesting in split testing is when it comes to…I guess you could call it the ugly conversion test, so weve seen some really shocking landing pages or sales pages that looked like they walk of the 90s. How do you determine the balance between what looks pretty and what converts the better than the normal? Lets say we are well established species that has a knack on being attracted to all the pretty things in life but sort to sum that up pretty for us is ugly, can the 2 styles coexist?

Valentin: well yeah, but not on the same website.

[Both laughs]

Valentin: depending on the audience there are still some people that are buying kich clothes right? So they appreciate the sheeny music so maybe theres an audience for any prop cause I dont know if you know the linkscar.com, that website is a totally crappy website thats very performance so, they are selling the cheapest leasing in the UK, they want to be ugly because they want to be relevant to their audience, thats why they look like a second hand website. Your website has to be easy to use in order to sell but it also needs to reflect your audience. The meaning 'ugly' could be more efficient than pretty, so like craigslist right? So they still havent changed their website so it still looks like 1998 but it works! Maybe you need to find the right balance between user experience, the prettiness and the efficiency. Maybe marketers want to have pretty websites but not all the pretty websites are actually selling. Were in this economic game, right?

Marcus: Yeah, I think its pretty interesting that youve mentioned how they cant exactly exist on the same page basically, its almost like having a really really bad ad on a site that was built literally last week [laugh] it looks like an ad that walked out of 1998, was trying to sell some MLM scheme and what not, so its interesting when you put it in that perspective because I think for some platforms weve noticed, I think it was on buy/sell ads that had really really high open rating and click through rate so to say, but Im not too sure about the conversions cause its just 9 day between them, with that being said, for anyone who has decent traffic on their site, what would you say like the 3 top things that they should test first, cause theyve never tested anything in their life, in their website, what are the 3 tests that they should run and from that, what should they actually analyze and try to improve.

Valentine: Yeah, so 1 important thing that youve mentioned Marcus is the having enough traffic, so weve actually made a calculator for ecommerce website to realize when is a good time to start testing because some of them are starting too late and some of them too early, so my rough estimate is that you can start testing after you have around 20-40 thousand business per month but you can do beforehand you can do, surfing your audience, find out things about your buyers persona, their barriers, their motivations, thats important to do before testing so you could understand who you are selling to, so that you can craftthose 3 elementsthat youre testing according to know hows or according to this data insight that Ive talked to you beforehand. Focusing on the main 3 element that I think everyone should take is the 1st and the most important test is your uni value preposition. Does it amplify or treat the pain of your ideal costumer and to test it make a lot of variations around. And another thing is the layout, too much noise above the fold can be very bad but you can think that okay, Ive squeezed all the things that are important because I know people are 'bouncy' but you need to take to account that people, Human attention span is 8 seconds and we are looking at the space of 200 pit holes to see it clear,so this are the limits of attention thats why you should change what kind of elements you want to put in the layout and 1st screen they are seeing and another thing which is important is the bottom of the funnel, any funnel has a forth, even if its an incomers or if its an lead generation website or whatever it is, you have a form and you have a contraction, so you should test which is the best form that you can get. For instance, a nice test that we somehow found is instant gratification on the form, right? So when people are typing, you should give them some green check mark 'okay, go further!' 'Continue the process' because the mind, its being tricked by addressing the subconscious that they can continue a thing which is giving slight dose of dopamine to the brain. So thats a thing that you could test.

Marcus: I think thats interesting, how you, like youve been so long in the Omniconverts, sign up model where you go, you know, 'your account is created' 'this is a great name' 'thats a great password' all this great little sayings to motivate someone to continue the process through the account set up process. And that sorts of pull to my next question actually, so you know, Split testing, whats in your actual funnel? I think its interesting cause I remember seeing Facebook for instance, right? So its not the initial sign up that were talking about instead we are talking about the existing user experience, particularly on the mobile and the market place side, now I remember about..2 or 3 months ago? I think it was, where they 1st launch the actual market place, some of the volume that was coming through to Australia had pricings on the actual listings where I had some of my friends in New Zealand which is, I guess you call that the physical play pen for a lot of businesses because of the first native English speaking to the timeline. They didnt actually had pricing on the listing and that enforced the user to actually click and open the listing to then be more engage and focused with 1 product and that provided very little resistance to them to go on and actually convert to communicate to buy from a seller. So from that example I guess, lets say for instance youre a SAS company, oh actually you kind a are [laugh] a SAS company, What test would you take down to sales funnel? Lets say for instance someone signed up and they are already a paying costumer, you know you split testing to anything to improve or raise the happiness levels of the user experience side of things. What test would you run thats purely focused on the actual user experience?

Valentin: One important thing to step here is that Sufter is a service gimmicks, un-wielding and is very hard to [inaudible] . its not like ecommerce which has 10 to 15 years , more . so theres more knowledge, more leaders and Sufter is a service company has this huge funnel from attention to referrals. The first thing after you finish the acquisition phase and you have a new sign up is to bring the costumer to the wow moments to whats the how moments. Tests which are important here to make the costumer have this another dose of dopamine right? That 'Wah! I made it!' once the things that the product or the Sufter delivers and how can you deliver positive emotions because we are human beings, we are running from pain towards pleasure. Thats it, thats the brief. So If your user can get instant value because weve seen that, the on boarding rate is crappy, so we imagined that at the beginning we got 1 out of 20 signups actually making an A/B test, so that means 5% of them only. After 2 years weve made it to 8% we were actually 'what the heck are we doing here' and then we realized that we need to deliver instant value but also to allow the costumer to be in control so they can select to what they want to do, theres a very good book from Charles Duhigg, the one who that wrote 'The Power of Habit', his latest book. Book is about what motivates users, what motivates everyone and one of the most important things is the power to decide, to be in control so you need to let the users to be in controland you to guide him towards his how moments. Things that are important to test here to have activation and then [inaudible] is to amply the pain instantly and to come up with a solution, to give him the control, to do this stuff that we bring him to the amok. One of the most important thing is the welcome email instantly, So what are you say is telling him the email and u can do A/B testing were Im using Intercom to do that. And weve made a lot of tests with this, we are still unhappy to it so we want to craft it even further. So another test that is important is to test a blank state, nothing on the, just do your stuff, do select your options, 2-3 for buttons or funneling them, so funneling your audience towards a how moment. So thats one of the most important things that you need to test in your software service so that you can head for activation and then for retention, of coursethere are all this tricks on giving him status about his progress, like LinkedIn used to do, your LinkedIn profile is 20% complete, thats what you can do with software

Marcus: Yeah, Ive never noticed that, its funny that you mentioned LinkedIn as well, cause I do recall them going through the different stages of having a profile, newly team member then you had expert and then you go to pro and then it was all star, but the funny thing is and it really ticks my OCD was they always leaves this tiny 3% and you could never complete the 3% profile [laugh] and that was the most annoying thing cause I go back and Id always try to work out 'how can I get the 3%, whats gonna make me, like this little tiny 3%, what am I gonna do with it? How am I gonna fix this?' I think thats interesting to trigger that happiness but its also allowing them to feel the motivation to come back, I think that solves a lot of peoples retention issues and it stops them from getting into the point where they to be resurrected as well.

Valentin: Yeah, because we are playing Marcus, we need to be challenged, we have tasks, we have feelings, we have progress, so one next step for me. So if you dont guide your audience towards the next step and if you dont challenge them to finish youre not gonna make them speak with your problem

Marcus: Yeah, its the other thing like, I come from a very consumer background at might, so when I hear all this conversion tests, its always a haha moment where you just like 'Oh thats the reason why they do it, thats the reason why theyre testing this, this is why this failed and this is why it worked' So with split testing as a foundation of Omniconvert, have to ask, you guys have got some massive clients, youve helped Samsung, youve helped T-Mobile, Avon, I was really surprise by Avon, to be honest with you cause I didnt think that they had much of marketing plan going but that being said, clearly they do [laugh]. I think every SAS companys dream to have multi-billion dollar companies, you know theyll own your stuff with your services for growing their business cause its like going 'Yes! I have a tier 1 company of the world using my services and Im getting paid for it.' So whats the trick? How did you sort a approach this guys or did they approach you?

Valentin: Weve approached most of them, so it was outreach but the trick here is to be perseverant and patient, so weve been waiting for 3 years, we still wait for some of them. And another trick is to find the real needs of the persons, which are the decision maker and other important piece to find out which type of persons can block the whole process because for us, for instance, the CTO, so the engineers, can shut out or shut up the whole process, so they can stop and they can say no. 'the script is not fat enough and this is affecting our websites performance. Its a long road ahead, you need to have enough energy and patience to go there and to understand whats their emotional trigger that everyone has. Because I do believe that B2B is actually driven by emotions, so B2B sales, are, there are still emotions over there so you do a good connection with 1 person from enterprise, company, there you go! So in the end its still the connection, the empathy and being there with your prospect.

Marcus: I think its pretty interesting when you split testing your outreach and by that I dont mean 'are you gonna drink coffee or red bull while youre doing outreach cause its a long process, really, I guess with the messages, split testing timing, obviously like you mentioned that finding the right people, trying to identify where the roadblocks are and who the roadblocks are. Have you sort a found a way to test, I imagine its really difficult because every company works in a different hierarchy that everybody has a different emotion about it too. Which is why you say B2B marketing or B2B outreaches is very emotionally driven, I 100% agree with you; but did you end up having any significant wins from split testing your outreach campaigns, so to say?

Valentin: we havent got something really smashing, I remember that 2 years, the appropriate person used to work; you send an email to anyone in the company that youre connected with LinkedIn or connected to someone on a different department and then you ask for the appropriate person then he states who that person is, and then to refer, to the guy which told you about the decision maker, and you start a conversation based on that. But now hes not working to go anymore. So unfortunately I dont think theres a recipe that you can use and so is working but the most important thing here is to take into account that the companies are buying because of the 4 factors right? The companies buy to make money, to save money, to stay out of prison, I mean the owners owe the company or to make things much more easy. So this are the 4 reasons, so if you cant address the reasons of the company, it doesnt matter if you address the decision maker. And then, you have also the decision maker in the company, which can be very stubborn, hes not into connecting with people, because imagine, I dont know, like Sheena and Atlasian, right? So, they dont have this sales team to teach their products so they said few months ago, that they havent hired no one to do sales. Because the product addresses the CTOs and the guys which are technical and sales guy could not talk with the—

Marcus: with the CTO

Valentin: Youre right and a technician or a technical guy so easy. So you need to understand in which market are you, Theyve done the same, so no sales initially but we are addressing marketers and marketers they look cheap. We are addressing also the marketers also the CEOs of this companies we can find ways to grab their attention and then its a matter of, I dont know there are supposed to be perseverant and persuasive.

Marcus: One of the coolest thing I noticed with Omniconvert all over, I guess your website even a lot of the articles you guys write,and even now youre mentioning in the outrage and it comes down to personalizing the experience for the person. And I noticed that sort of a resonates really really well with Omniconvert, and to go a little bit deeper than that I mean more than just using that persons name at the appropriate time and when not. You know, the personalization is a pretty active time, you know when you go down that route you can be apply an email, copy that design, So I mean, what do you think the fundamental values of user personalization exist in a marketing plan?

Valentin: Yeah, it has a lot to do with Cialdinis 6 principle of persuasion. You need to be relevant and you need to show sympathy, in order to that, you have to adapt to who your costumer is because tailor-made approach is always being 1 sized approach, and the trend, the latest trend is, youve seen personalization on Coca-Cola right? , so they personalize your, the bottle. And the, youve seen that on Starbucks, they write your name on the coffee cup. So personalization is, it is a big trend and it has to do with the marketing and with the costumer experience officer as well, so CMO and CXO, they are going to lead the growth of the big companies. And if you dont deliver experience because we are getting out of the economics of commodities and we are getting out of the information era, we are getting to the knowledge era and to the experience era. So if you are not giving me the value and the experience, we are hunting, we as humans, we are hunting this small positive experiences which are building the trust and we are in this dynamic world that, that I as brand, I need to be relevant to the audience because that how I am different from the others. And the personalization, it is a big caffeine in this industry thats why we, thats one of our key point of difference against our competitors. I think everyone should take personalization seriously because otherwise youre just singing the same song to everyone.

Marcus: Yeah, I completely agree on you on that, I think particularly on people doing outreach or using the same website input or the same website content, you know dont get me wrong, theres a good reason why templates exist, its fast to set up, people are familiar with it, consumers are very familiar with it, particularly, lets say for instance, EBay for instance, that shopping experiences is also similar to Amazon which is also similar to Alibaba, so everybody is used to that same sort of format. And I think thats when templates have their place, but when it comes down to making it very specific to the user or for the customer, I think thats where the differences with Omniconvert is you guys have that personalization perspective that a lot of analytics dont actually, I guess you could say, Im not gonna say comprehend, comprehend is not the word for, but they dont use as a point of focus. And I think thats the one defining factor of anyone in analytics is as much as everybody goes by the data, theres still something as far as, you know people will call that God instincts but I think its coming down to, what you said earlier about increasing someones level of happiness. Basically when you give someone a dopamine hit without giving them an actual dopamine hits. [Laughs] And I think, from what Ive seen with you guys, and I find this to be a really really cool campaign, wheres after somebody actually signed up, you sent the working email but youve also taken it one step further and giving them, I think it was over 50 ideas or strategies for them to immediately start implementing so they actually have a reason or purpose or almost like a guide on how to actually start from scratch with analytics on their website, with split testing, with increasing the conversion rate of their business, I think that sorts of what stands out most to me, as far as what companies are doing. Because, I mean you could sign up to google analytics for instance, all theyre giving you are just courses of going 'oh, this is how you do A, B, and C' theyre not really trying to go that one level deeper that try to help you, I guess come up with what you need to actually do and I think thats a really really big defining point. So I got 1 last question for you which is, its gonna be a fun one, its actually came from Tim, Hes our head of warrior actually, so you ready for this one? [Laughs] yeah alright, its an interesting for you, So when your, when youre testing on the top of the funnel right? So like landing page, your outreach, lead generation, youre focusing on specific elements like a CTA, like an Ad, and like graphics and what not. So you set the goal already that, lets say for instance you have 5% of conversions and you want to take it to 10, youll do everything in your power to move it from 5% to 10% right? So with that, regardless of the successes, you know, the very specific out like the 4/5 that come out as failures, from those, thats a lot of failure [laughs] How do you measure the effect of the failure rates of those tests at the very top of the funnel, and as a result of that how do you compensate those results when you get to the latest age of the funnel?

Valentin: Yeah, thats a good question [laughs]

Marcus: thanks Tim, [laughs]

Valentin: So this is the cost of testing, if youre testing you need to be prepared to risk and you need to be prepared to lose but if you have a sample, lets say, ING for instance, they are a huge bank here in Europe, so theyre, I dont know, 20 countries. They are testing with Omniconvert; they have a sample of 10% which is not effective at all so they dont do any kind of test and then they are testing everything because they are very open to find out the winning version, so if they can test on the other 90% of the visitors, they can get to the best version much more faster, so in this manner, they can, of course they are analyzing the whole impact with the, of the testing in Google analytics because we are degrading Google analytics, and thenthey will be sure to see that 'what if I wouldve thought to test it, so what happens after 6 months of testing' The ugly truth is that, most of the small entrepreneurs and smaller companies are not testing properly because they are being deceived by their own limit of testing. So they struggle and they dont find a winner and they say 'No! now Ive lost 6% of the conversions with this test but its a cause that you need to pay and the good news is that, the quality of each test is not as important as the quantity of test you are doing. So the velocity of testing is the one that shows how fast youre gonna get to a much more better version of your website, so that, at the end it pays off, but you need to stay in the game so, its like, if youre not testing that means you are wasting money. But if you stop testing, you took testing only for 3 months, thats it youve lost the whole amount of money because you havent reach the finish line like the marathon. If you dont start at all, you are happy but if you abandon after 5km, but if you stay the whole thing at the end of that marathon, man, Ive run in the marathon, Ive seen, its like pure ecstasy, right? So thats split testing as well and at the end is gonna pay back.

Marcus: Okay, I think, its not an easy question to solve and I think youve managed to help on that one quite well, Ill have to ping Tim about that actually [laughs] But look Valentin, I really appreciate your time today, I think in all the people in the world I wanna ask about split testing, Im very very happy to have you hosted in the Warrior Podcast, so what will do is I think well leave it off there. For the warriors out there, Ill be sharing a lot more content with Valentin coming up very very shortly. Again Valentin, its a pleasure to have you on the warrior podcast and we look forward to having another chat with you again soon, hopefully we can pull a panel together of conversion rate optimizes and share some of our top secrets together.

Valentin: [laughs] Okay Ill be glad to do that. Thanks as well Marcus, I want the best for all the warriors, may you be impressed about your results!

Marcus: Yeah alright, Thanks Valentin.

Valentin: Thank you Marcus.