[AMA] I’m Michael Aagaard, Senior Conversion Optimizer at Unbounce. Ask Me Anything

If you’re serious about conversion rate optimization, then you’ve likely heard of Michael Aagaard: founder of ContentVerve.com, international speaker and, as of this year, the Senior Conversion Optimizer at Unbounce.

His pragmatic approach to conversion rate optimization, focusing more on research and real results rather than mere “conversion lifts,” has made him a respected authority in the field. (Oh, and he’s a handsome and a super friendly guy to boot. He’s been described by his colleagues as a “norse god” and a “lightening bolt in human form.”)

Simply put: he’s the guy you want in your corner when you’re prepping your next A/B test hypothesis.

Well, on November 30th, we made that happen — Michael answered your burning conversion queries that disturb your slumber (to make sense of the marketing mayhem that fills your waking hours).

Unbounce’s Senior Conversion Optimizer Michael Aagaard.

Check out the comments below to see the Q&A, where Michael answered questions about conversion research, formulating a hypothesis and what it’s like to transition from being a consultant to working in-house.

default author image
About Amanda Durepos
As Unbounce’s former blog editor, Amanda Durepos curated and crafted content to help marketers be better at their jobs. When she wasn't managing the editorial calendar for the Unbounce Blog, she was wrangling internal and external subject matter experts, helping them educate other marketers in an entertaining and actionable way. You can follow her on Twitter but she’s mostly just making puns.
» More blog posts by Amanda Durepos


  1. Lydia

    My boss *insisted* on keeping the full site navigation on each landing page (one is above). How bad was this? How important is it to limit the movements of the incoming lead?

    • Michael Aagaard

      Hi Lydia,

      As a rule of thumb, I’d keep distractions to a minimum. So in most cases I’d remove most of the global nav to facilitate a targeted landing page experience. However, that doesn’t mean that you always have to remove every single link.

      In some cases users like to click to the Home Page to get an fuller picture of the brand behind the landing page. In other cases they start clicking to get the information they need in order to convert.

      It also depends on the purpose of the landing page, the brand itself and whether you are dealing with a high-scrutiny product.

      I’d do some quick research to get insight on how the nav is affecting the user experience. You could go to your landing page report in GA, click on Second Page Report and see which pages users are visiting after the landing page and where they are exiting the website. Also I’d set up some click and scroll maps to figure out where people are clicking and how far they are scrolling.

      This insight will help you understand how users are interacting with your landing page and the nav and whether removing it will improve the user experience.

      Looking at your landing page, my gut reaction tells me that you need to give users more information on the landing page itself. My biggest question was, “What happens at the consultation?” Personally I’d start clicking around the website to get that answer.

  2. Brian O'Meara

    For inbound lead generation (downloading an asset/eBook)… Which converts better? … small form on the landing page or just a button (to a small form)? Also, when the form is completed … do you simply open the asset or take them to a thank you page where they have to click again to download the asset? (which do visitors prefer) Typically my first button says “Download eBook” so what should second and possibly third button/link on subsequent pages say?

    • Michael Aagaard

      Hi Brian, Hate to give you the fluffy answer “It depends…” But it really does depend :)

      I’d analyze your current setup to get some insight on how it is performing. This’ll give you an idea of whether you are currently doing the right thing.

      Moreover, I’d do some user testing to find out if there is an alternative that would perform better. Do some mock-ups and invite e.g. 5 people from your customer base to go through two different scenarios (one with the form on the page and one with a second step) to get input on how the two different experiences affect their motivation.

      If the initial feedback suggests that e.g. changing to a two-step process is better, I’d move forward with more user testing or an actual A/B Test on the landing page (given that you have enough traffic to run tests).

      In relation to giving users the asset itself: it definitely feels better as a user if you get the asset as soon as possible. However, whether you can deliver on that depends on stuff like company policy – double opt-in on email submission, etc. Also consider your own tracking setup, in some cases it is much easier to track conversions if you have a stand-alone confirmation page.

      As far as choice of button copy in a situation with several steps, I’d use “Get My E-book” on step one and then “Download My E-book Now” on step two.

      Hope that helps!

  3. William

    Great way of interacting! I am curious about your ideas about the following:
    -Using the scrum method (for example 2 weeks) for a high traffic platform (sample size will not be a problem). What’s is your experience? What are the best practice?
    -What do you think are the most advanced but crucial analysis every cro-specialist must perform (ad hoc/recurrency )?

    • Michael Aagaard

      Hi William,

      I’m not 100% familiar with SCRUM, but from what I know about the process, it could be a very interesting setup for CRO and A/B Testing. The main thing being that you have to make sure that 2 weeks is truly a representative test duration that will give you real insight.

      Quantitative analysis using Web Analytics setup to constantly understand the state of the online business and to identify new opportunities. This is an ongoing process that never really stops.

      Qualitative research to better understand where your customers are at – interviews with customers themselves and the people in Sales and Support are extremely valuable.

  4. Joe Faillace

    Hello Michael,

    Awesome of you to share your time and knowledge with us! What can you share with us on how to use Google Analytics Aagard style?

    Tip of the cap,


    • Dan

      I hit the little down arrow thinking that would show comments or something. I think it down voted this question when I should have up voted. When I tried to correct my mistake it basically said something like “too late, u already voted.” Maybe someone can fix it? If not, my apologies to everyone on team analytics

    • Michael Aagaard

      Hi Joe – I use Analytics a lot! I see it as the bread and butter of quantitative CRO Research.

      My approach is to start with the big picture and then zoom in.

      So first off I look for general stuff to give me an idea of what strategy to assume – traffic and conversion numbers are crucial here.

      There might be just enough traffic/conversions to run tests (calls for radical treatments). There might not be enough traffic/conversions to run tests (calls for radical changes and tracking effect over time). There might be tons of traffic and conversions (you can test even small changes and be able to detect impact). As you can see already, these are three very different approaches.

      After that I dig a little deeper and look at performance across: devices, browsers, channels, days of the week/month, etc. This helps me understand the external factors that have impact on the user experience.

      Then I move on to figuring out where in the conversion funnel things are going wrong. I use a basic manual step drop analysis for this purpose and it gives me a clear idea of where the biggest leaks are.

      This insight tells me where it makes most sense to start – which in turn helps me build an actual CRO strategy. At this point I move into more specific details on the individual sections I need to optimize. Depending on the page itself, I look at stuff like bounce rate, exits, next page report or navigation summary, new vs returning, device mix and many other things depending on where the data takes me.

      Let me know if you have more questions sir!

      • Joe Faillace

        Pure wisdom :-) thanks for giving some insights into your methodology!

        Much appreciated,


        P.S. rumor is that you have some top secret GA templates that you share every blue moon…what’s an Unbouncer gotta do to get a hold of ’em?!

  5. Andrew Miller

    At what points in the research, hypothesis formation, test planning, design, and deploy stages do you think clients should be involved? For context, we’re an agency hired to help solve specific Conversion Rate challenges but frequently get bogged down at the design phase because it suddenly becomes “real” to our clients when they see our proposed pages and turn into design critics.

    • Michael Aagaard

      Awesome question Andrew!

      Before joining Unbounce, when I was doing CRO consulting on large projects, I found it helpful to do a kick-off workshop with the client where I’d get all the important stakeholders together (people from I.T., marketing, sales, support, etc.). At this workshop, I’d explain my process to them and make sure that they all understood what I was hired to do and how I was going to do it. This way everyone gets it from the beginning, you can set realistic expectations and get employees excited about the work at hand.

      Moreover, you can show them that you are serious about business and not just another Snake Oil Salesman and thereby gain their trust (which is crucial).

      After that I’d start my research process and get the necessary stakeholders involved (just the ones I need to get the input I need). Then I’d invite stakeholders back for workshop to present all the insight, get their feedback and plan the next steps.

  6. Trevor C

    I am all about digging through data to find gaps in our conversion process to determine my next hypothesis for a test. I never thought I’d say this, but I love data and finding ways to improve the status quo. Often times we get a lot of paid traffic to our landing pages, where the users for the most part all behave the same way. This makes it tough for me to dig deep and find those gaps/inefficiencies.
    To get to my question, I have a hard time just “willy nilly” coming up somewhat random things that I (my moral compass) feel will do better, or looking into best practices. Should I feel that way, or should I just do what I need to do to make things work? Or do you have any tools or suggestions I use to dig even deeper into data? Currently using Adobe products for data, Hotjar, and a few other tools.

    • Michael Aagaard

      Sounds like your doing fine with the quantitative stuff (i.e. finding out where things are going wrong). However, it seems like you might be lacking insight into why users aren’t converting.

      I recommend digging into some more qualitative research to get this insight. Asking questions, doing interviews, running feedback poles are all super powerful methods that yield incredible insight. Within the last two years I’ve started spending more and more time doing qualitative research. You’re basically getting your own users to tell you how to make the conversion experience better.

      My general approach is to uncover exactly what they want to do and how they want to do it – as well as what’s keeping them from doing it.

      • Laba

        Hi Michael, this is exactly my problem, but it’s not clear to me you would ask “questions, interviews, feedback poles” to all the customers who didn’t convert. I don’t have their contact information. I only have the contact info for customers who liked my page. Much appreciated!!!!

  7. Don Osborne

    When starting with a new client, what does your personal “check list” look like to prepare for, implement tests, and evaluate results to improve conversion? I imagine you’d ask the client what their goal is, so let’s assume the client’s first goal is to increase conversion on a simple landing page / opt-in form, but after some questions the client reveals that his/her real goal is to increase the # of warm leads and sales through their funnels, starting with both organic traffic from search (via “top of the funnel” content) and paid traffic from Facebook and Adwords. You discover the client in fact has an email list of prospects with 5,000 names collected in the 12 months.

    What direction do you start in and what destination do you hope to arrive at? Care to share your internal “shopping list” as you start with a new client? (E.g., we’re going to bake a cake, so we need flour, eggs, sugar, an oven … .) I’d love to “see inside” your internal thought process (aka, from high school, “show your work”),

    • Michael Aagaard

      Hi Don – thanks for your question sir!

      I outlined two of my processes in my replies to Andrew Miller and Joe Faillace. They basically cover what you are asking here. Please read these and let me know if you have any follow/up questions.


  8. Marcello

    Hi Michael,

    I know you’re amazing with forms, and in this period I often notice websites, especially ecommerce sites, do the impossible to add christmas decorations and snowflakes all over their pages, even on signup pages.
    While that’s nice and pretty, one thing I believe they never asked themselves is:
    Does this actually improve conversions?
    Having misplaced elements on the page (unless you sell christmas products on that specific landing page) should be done in a specific way? Won’t it add to distractions on a page where the real target is to get the conversion?
    And how would you USE these type of decorations on a landing to actually drive users to convert?



    • Michael Aagaard

      Hi Marcello – interesting question!

      In my experience there is no one-size-fits all answer. It depends on multiple factors like the brand itself, the product, the execution, user motivation, etc.

      I certainly don’t think that there’s some inherent psychological phenomena that makes users convert when they see Christmas decorations all over the place :)

      If I were with a company planning on going nuts with Christmas stuff, I’d do some user research first to get an idea of how users will react and how it will impact performance.

  9. Ami

    I have been having these back and forth discussions with my colleagues about what is the most accurate and truthful way to calculate conversion. (On a landing page and for an online shop). We don’t seem to reach a common ground and there are quite a few articles out there giving diferent formulas (true conversion etc.) What is your opinion?

  10. Tom

    Michael, do you feel happier as a optimizer for SaaS company than when you was freelance?

    • Michael Aagaard

      Hi Tom – thanks for the question.

      I’ve honestly never been happier working than I am now :) But that has a hell of a lot to do with Unbounce and the culture here – it is an amazing place. I can imagine many other SaaS companies where I wouldn’t be happy.

      Also, I’d been freelancing for so many years that it was a welcome change of pace. It is fantastic to get to opportunity to really dig into one brand and really get a deep understanding of how to move the needle – rather than scratching the surface with a lot of different clients.

  11. Raul Tiru

    Michael, Thank you so much for doing this! Tweeted it ;)

    What worked for me in the past was letting people know WHY I do what I do instead of telling my features (Simon Sinek).

    What is you Biggest takeaway for conversion optimization. I know it’s not focusing on small variables ;)

    Looking forward to hearing from you soon!

    • Michael Aagaard

      Hi Raul – the most important thing I’ve learned about CRO is that process is king!

      Having a solid process in place is the closest thing to a “works every time” solution. Research is by far the most important step in my process. It is all about the insight. The better you understand you conversion problems, the easier it’ll be to find the right solutions.

  12. Dennis

    When analyzing your test results, what type of statistical analysis do you rely on to show you (and your clients) that the data you’ve collected can be relied on to make an informed business decision?

    Are there any specific statistical tools you prefer to use?

    • Michael Aagaard

      Hi Dennis – great question.

      For really solid test data (using a frequentist approach), here’s what I want:

      The pre-calculated sample size (users).
      A sample of 400 conversions or more (actions).
      Test duration of 2-4 weeks (depending on business cycles and traffic).
      A significance level of at least 95%. Power of 80%. P-value less than 0.05%.
      No overlap in the confidence interval.

      But there is no eternal exact truth, it is all relative and there many “depends on” factors.

      I love http://abtestguide.com/calc/

  13. Ed

    Do you need to know statistics to a high level? And which is the more important: stats/analytics or design/psychology? Cheers, Ed

    • Michael Aagaard

      Knowing statistics is very important. I wasted a lot of time running shitty A/B Tests and conducting horrible research simply because I did not have sufficient knowledge of statistics.

      Most of what we do in CRO has to do with stats in one form or another. The better you know stats, the better you’ll be at CRO.

      But having said that, you of course need to know a lot more than just statistics. And design/psychology is super important as well. Stats/analytics help you understand what going wrong and where it is going wrong. Psychology helps you understand why and design helps you come up with great solutions.

      If you can’t accumulate all these skills yourself, it is very valuable to team up with someone who is strong with the aspects you are lacking.

  14. Audrey Bloemer

    Hey Michael, I wanted reiterate a question that’s already been asked a few times. “Getting started” with CRO recommendations is often one of the hardest parts of the process. In your opinion, what are most important metrics/reports/analysis to use and review when you are starting the CRO process. You shared a great “second page” GA report at CTA Conf, any more tactics like that you can share? And how you use the data! Thanks!

    • Michael Aagaard

      Hi Audrey – thanks for the question.

      I think the answer I gave Joe Faillace in an earlier comment will answer your question. Read that first and let me know follow-up questions you have :)

  15. Tim K

    How much does “click count” – the number of clicks required to complete a process – really matter?

    • Michael Aagaard

      Hi Tim, it’s all relative and depends on the individual case. In my experience it has to do with perceived complexity. In some cases a conversion experience can seem more complex with a more clicks. In other cases, it can seem less complex with a higher click count.

      A good is example are checkouts. I’ve seen many one-step checkouts backfire heavily because the one-step variant has so much information and so many fields on one page that it becomes extremely crowded and difficult to decipher.

      It crazy how quickly people establish rules for things that really are highly relative. I prefer to use common sense and data to inform my solutions : )

  16. Alex

    Michael, what are your favorite bands and music groups. Also do you listen to music while doing CRO? How are you liking Canada? ?? see ya in 2016!

    • Michael Aagaard

      Yo Alex! I have a broad taste in music, but I’m a Punk Rocker at heart. My all time favourite band is the Ramones (got a Joey Ramone action figure on my desk and I used to play bass in a Ramones cover band).

      Bad Religion, Misfits, The Clash, Samiam, Teenage Bottlerocket, Menzingers and Glasvegas are other punk favourites. I also love Metal and Hard Rock like Slayer, Pantera, Guns n’ Roses and Danzig. I’m also into classics like The Stones, Neil Young, Cash and Dillon. Even classic West Coast Gangsta S**t is on my play list. Albums like Doggy Style, The Chronic, Straight Outta Compton and Eazy Does it are amazing.

      I have periods where working to music helps me focus and tune out distractions (now that I’m working at Unbounce with 130 people, this is becoming more frequent), and other periods where I seem to need silence to concentrate.

      See you in 2016 man!

  17. Pascal

    Hi Michael,

    thanks for sharing your knowledge.
    What is the best or your favourite check-out of a e-commerce (e-retail) site at the moment and why?

    Thanks so much

  18. Geoffrey

    Thanks for the amazing opportunity to share all the doubts and hesitations! So many exciting questions already have been asked, and I barely can wait to get the answers.
    And here is my question: Some people really hate popups. The pop-up form that triggers after a few seconds interrupts visitors and can indeed be perceived as annoying. BUT why do so many great marketers and successful e-commerce sites use them? Can they actually improve conversion rate?
    Thank you so much, for your help!

    • michael aagaard

      Hi Geoffrey – sorry for the late reply here :)

      Ah the dreaded pop-up. It is in my experience quite a complex subject as there are so many variables to take into consideration. I’ve seen lots of pop-ups that were timely, relevant and not annoying at all. Vice versa I’ve also experienced lots of pop-ups that really bugged me. It’s all about context and delivery.

      I’ve run several successful experiments where pop-ups have worked well. But in those cases I put time into doing my homework so I had a relevant offer and a timely delivery of the pop-up. But I have never experienced massive lifts like the ones you see the vendors promote :)

      I’d say it is definitely worth experimenting with.

  19. David

    Hi Michael,

    do you have any experience in publishing articles to readers in their native language ?
    I’m interested if there is a higher CTR or more interaction if visitors found articles in their own language.

    Thank you! David

    • Michael Aagaard

      Hi David, I don’t have any hard data on this. However, it makes sense that reading in your own language will make the experience more relevant.

      But than again there are tons and tons of variables to consider: is it educational content, is the target audience multi/lingual, are you selling something, is there a local market, is it possible to translate the text and properly get the message across, are you doing this for SEO purposes (local vs. international), etc.

  20. Stacy Sutton Williams

    Similar to Dennis’ question – what type of statistical significance do you look for? We have some niche clients that don’t have a ton of traffic. And they have super low starting conversion rates. When I use Optimizely’s sample size calculator, in order to get to a 95% confidence level, we’d have to run a test for 6 months! This isn’t very practical. Any suggestions? Thanks!

  21. Jakarta Karoseri

    Hi Michael,

    I read some articles on your blog & i love it :)

    Do you have tips when you create outline for your blog post?

    Thank You

    • Michael Aagaard

      Hi Jakarta – thanks for the kind words!

      When Creating an outline, I think of my overall goal with the blog post – what do I want the readers to take away? Then I think of how to best tell a story to get my point across – what examples and data can I use? I also do some keyword research to get an idea what phrases people are using to find info on the topic I’m writing about.

  22. Tom@Miditech

    I am testing few landing pages right now for one of the health product. I am confused if I should use full site navigation or use no navigation on LP

    • Michael Aagaard

      Hi Tom,

      As a rule of thumb I’d try to remove distractions from my LP, and generally full site nav can be distracting on a page where you want to convey one specific message.
      But that doesn’t mean that you can never ever have any nav on your landing pages – that is a useless generalization. You need to find out which links area actual distractions and which links are actually helpful.

      E.g. I often see users clicking like crazy on the company logo on a landing page. Via qualitative interviews with users, I’ve found out that for many people it is just normal behaviour that they want to see the home page to confirm that the brand seems trustworthy.

      Click maps, second page report, navigation overview are tools that can help you get this insight.

      Also think of it as a trade off between how much content you need/want to give your users on the LP itself. If 90% of your users are clicking “Pricing” in the global nav, that’s an indication that your potential customers need to know more about your pricing – now it is up to you whether it is most beneficial to have that info on the landing page and remove the nav link or to let them click through to the actual pricing page.

  23. Sven

    Hi Michael,
    I liked how you introduced yourself to unbounce :)

    Here is a little follow-up to the last AMA from Oli…
    I asked there:
    “How to optimise the main landing page of a (complicated) product for developers?
    (SEO CRO Usability)
    Generally LP Attention Ratio should be 1:1 (only one link/CTA on the page).
    But this is a complicated product, so generally speaking no prospect coming new on the page (by Google Search) will immediately click on the main CTA “Download Free Trial”,
    they need more informations about technical details.
    => For this reason (and for SEO reasons) main questions are linked to pages with additional details. This kills the Attention Span completely (about 60:1 or even more :(
    So what?!?”[…]
    Oli answered:
    […]”Something that can be very effective in this type of situation is to remove all of the links that go elsewhere on the site – so a dedicated landing page. But leave a sticky navigation in there. So instead of going to sub pages, the nav brings you further down the long page to the content they are looking for. This way the Attention Ratio is still effectively 1:1 or 2:1 – because the nav links aren’t “leaks” away from the page.
    This lets you do heat map analysis to see which of the nav links people are clicking on, and removing any that are not getting attention. You can then try removing the content that that nav link was related to as it might not be important in the purchasing decision.
    In terms of CTA position, it’s important not to be afraid of moving it down the page so people can have a chance to read first without being distracted by the “ask” being too aggressive. For this you’ll want to do a scroll map to see if people are even bothering to scroll, and how far they are getting down the page – which can influence where you place the CTA.
    Hotjar will let you do both of these things.
    Regarding SEO, if you are running campaign traffic (paid ads,email, social, display) then you don’t want to pollute your data with organic traffic. In that case the setup you have might be okay (for organic) and you should try the single page with anchor sticky nav for your campaign traffic only.”

    So here is an followup-idea about how to have some “internal links” (for SEO) and keep the Attention Span still focused:
    (which could be suitable both for SEO and campaign-traffic)
    – ok we’ll remove most external links from the landing page
    – we structure the Landing page into several information blocks containing much white space between and guided navigation
    (like explained in unbounce ebook “attention driven design”
    on page 35/36 “Whitespace” — do there exists live examples for this??)
    – navigation will be a sticky navigation which works just as anchors (like described by Oli)
    – we try (as described by Oli) to find/optimize the most important blocks of information for the landing page
    – BUT additionally for the (maybe four) most important features/topics we’ll create special pages
    – so if for a certain information (maybe “licensing”) the “short block” on the landing page does for a certain visitor not provide enough information (to convert), there are maximal four (4) links to click on which will lead to special pages containing more information about this feature/topic.
    – these special “additional info pages” are designed in a way that they do not include complete website navigation either (so have a very focused attention span, too!) and are designed in a way to encourage to EITHER go back to the main LP and go on there OR (via an included main CTA conversion – trial download) execute there directly the conversion

    Do you have any experience with such a “two-level” “microcosmos” design/concept?
    Are you aware of websites/Google-Search-Landingpages using this kind of concept?

    Thank you very much in advance for any hint,

    • Michael Aagaard

      Hi Sven – your question is 650 words long and consists of references back to questions you asked Oli, answers he gave you, follow-up questions, multiple iterations and ideas based on this as well as several links to external pages.

      All in all this makes it very difficult to figure out what your really asking here, and I simply don’t have time to go through it all in detail to figure it out.

      If you can boil it down a little and ask me a direct question, I’ll have a better chance of being able to answer you in a blog comment :)

      • Sven

        Hi Michael,

        sorry – the question is complex. It is about a two-level landing-page concept AND optimising the landingpage like explained here on page 35/36 “whitespace”:

        Ok, if “both elements” are too much, the most reduced question is:

        Do you have any experience with such a “white space” design/concept?
        (as mentioned in the cited unbounce-whitepaper on page 35/36)

        Are you aware of websites/Google-Search-Landingpages/Real-Word-Examples using this kind of concept?

        This way it is hopefully simple enough :)

        Thanks anyway,

  24. Agen Taruhan Online

    Hi Michael,
    I liked how you introduced yourself to unbounce.
    Do you have tips when you create outline for your blog post, Do you have any experience in publishing articles to readers in their native language ?

    • Michael Aagaard

      Hi Agen – thanks for your question. I answered both these in previous comments. Please look at the questions from Jakarta Karoseri and David to see my answers.


  25. Peep Laja

    So who all live in that epic beard of yours?

    • Michael Aagaard

      Elvis Presley and Bob Marley have been living permanently in my beard for almost 12 months now (fantastic tenants!). Tupac has expressed interest in moving in. But I heard that he can be a quite noisy at times, and I’m concerned that the other tenants will move out.

  26. Dan

    For short campaign flights that do not have the time or budget for sufficient CRO of ads, pages, or other assets, how would you approach the build out and delivery of the campaign? Would you just hope for the best, point out the pitfalls and recommend fixes, or insist on delaying the campaign until changes are made for instance? Let’s also assume that you’ve been given an initial set of ads or LPs that are so bad they are basically broken and the campaign must go live tomorrow.

    • Michael Aagaard

      Hi Dan – fantastic question!

      I think all marketers fear that situation (maybe we’ve even been in it)!

      If it is fundamentally broken and everything indicates that the campaign will suck, I’d do everything in my power to delay the campaign with just enough time to fix the worst stuff.

      If delaying is totally out of the question, I’d focus all my energy on fixing as much as possible before launch. I’d try to get key stakeholders into a room so I’m as close as possible to the people that can actually implement the changes. I’d walk them through the glaring mistakes and hopefully get total backup for fixing as much as possible as soon as possible.

  27. Top Buffets

    Hello Michael , in his opinion how you can gain new customers or visitors to a niche portal when we are not selling a product but selling other companies services? In our minds we do not see the possibility of applying the landing page strategy Thanks for the opportunity

  28. Amrdeep Athwal

    Hi Michael,

    Oli gardener says that one of the things that impresses him most about you is the amount of research you do when building hypothesis.

    Some of the biggest uplifts I have delivered have been when I have spent days digging through analytics data, qualitative data such as surveys/session recordings, reading a whole site top to bottom and even becoming a customer/client of the brand in question to find the pain points.

    Apart from doing all that what other recommendations do you have for building your hypothesis where do you go get insights and how do you separate the signal from the noise.

  29. Brad

    What is the min amount of traffic you should have before considering CRO?

    • Michael Aagaard

      Hi Brad – that is a fantastic question that I know a lot of marketers are asking themselves.

      One thing to remember is that CRO and A/B Testing are not one and the same. A/B Testing is a tool you can use in your CRO process, but it is not mandatory. You can easily do CRO without testing.

      Generally you can say that the more traffic you get, the more data you’ll have, and the more data you have, the easier it will be for you to get insight to drive informed hypotheses and treatments. This is the most important part.

      When it comes to the question of when you have enough traffic to start testing, you need to consider how many conversions you get – not just how much traffic. You need a certain number of actions or outcomes before you can run meaningful experiments. As a rule of thumb, 1000 conversions pr. month (of whatever metric you are measuring) is a good starting point for doing advanced testing with fairly deep segmentation. You can use this tool to calculate your testing capacity:

  30. Mike Mozg

    Hi, Michael!
    What can you say about LPO for Asian markets and in Japan in particular?
    Maybe some hints and advice?

    • Michael Aagaard

      Hi Mike, sorry man but I don’t have specific any experience doing LPO for the Asian market.

      My best advice is to do your homework and conduct research into your target audience: who are they, what are their drivers, what are their barriers, why would they buy your product and why not, demographics all the groundwork. If you have campaigns running make sure that you’re constantly doing research to get insight into how your campaigns are performing and how you can improve them.

  31. Spencer

    What is the future of conversion optimization?
    What kind of information would you like to learn more? (mobile opimization & testing, app optimization & testing or something way different?

    • Michael Aagaard

      Hi Spencer – there’s a lot of talk about what the next big thing is. However, I’d say that in general there are so few people who get the basics, that I’m not too worried about the next thing – I’m worried about getting people to simply understand the basics of CRO – statistics for example :)

      I’m always trying to get better and learn new stuff or get deeper insight to stuff I already understand. Lately I’ve been reading up on statistics a lot. And within the last year, advanced analytics has been my thing. The next thing I’d like to get better at is code so I can do more stuff without having to ask developers for help.

      One of the big things that I look forward to though is getting tools that will help us get user-level insight across platforms, browsers, devices, etc. So we can get a better understanding of what the real user journey looks like. But that’s going to take a long time and we need to have something much more solid than cookies.

  32. Nick

    Hi Michael, thanks for your time today!

    What are your thoughts on multi-form CTA sales landing pages?
    – Unbounce only allows one form per page, is this not limiting?

    What do you suggest as a work around?

    Thank you :)

    • Michael Aagaard

      Hi Nick – I’m not a big fan of the “stalking CTA”, the kind that’s there in every single screenful asking you “ARE YOU READY TO CONVERT NOW! IT’S BEEN IT’S 150 PIXELS SINCE WE LAST ASKED YOU!” :)

      Personally this experience gets even worse if you’re throwing the same form at me multiple times on the same LP.

      I’d recommend that you feature the actual form once on your page, and “anchor” CTA buttons that’ll take you to the form if you click on one of them somewhere else on the page. But as part of my research I’d ask myself the question “Where does it really make sense to feature the CTA?”, “Where does it naturally fit into the decision-making process of my potential customer?”, “What is the tipping point between helping my potential customers and bugging them?”

      I can’t think of any cases where I’ve seen a significant benefit to having the same form featured multiple times on the same LP.

  33. Laba

    What variables and data should I be looking at post campaign to analyze a landing page (with variations) that had very poor conversion rates? The only advice I seem to get is to run A/B tests. Is there anyway else.

    • Michael Aagaard

      Hi Laba,

      I’d make sure to dig into your analytics data and look at stuff like performance across devices, user types, channels and browsers to see if anything stands out.

      Did a particular device perform much better than others? Did one channel constitute 96% of conversions? Did you have tons of people clicking your ads but very few converting on the landing page? If so maybe you have a problem bridging the gap between ad and landing page.

      Averages lie, so even if you are running A/B Tests, you need top integrate your data with your web analytics setup so you can dig into segments and understand the details.

  34. ron

    I am told over and over to have an offer in exchange for the form fill/get contact info. I just don’t know where to put that offer. lets say offer is ebook.
    1)I can’t strictly advertise the offer in all my digital marketing, so my above the fold needs to match the marketing message. In this case “we will find your XYZ/we will curate..”
    2)Building on that, if my form is above the fold, then there is no room for the offer above fold, and/or if I stick in in there the above the fold becomes cluttered.
    3)If I put the offer on the bottom form, the people who scroll down will never see the offer
    4)Should I put a popup email capture with the offer (get rooster), or will that annoy?
    So confused!

  35. Jared

    Hi Michael,

    What amount of data would you say is beyond small sample size and considered adequate or substantial enough to draw a conclusion regarding a landing page’s effectiveness (or lack thereof)?

    Thanks in advance!

    • Michael Aagaard

      Hi Jared, it’s all relative and depends on what data you are looking at, what kind of decision you want to make, and how much risk you are willing to run.

      Generally, the bigger the sample, the more accurate the data (assuming that your data-collection methodology is sound). Sample sizes also depend on whether you are conducting qualitative or quantitative research.

      When analyzing split test data, I want the pre-calculated sample size, as well as at at least 400 conversions to be somewhat confident in my observations.

      If I’m digging into a segment in analytics, I’m suspicious of samples under 100 (I’d prefer 300 – 500 in any given segment).

      If I’m running a Feedback Poll on a landing page asking a specific question, I’d want around 100 replies (I’m looking for input here not eternal truth).

      If I’m conducting a series of qualitative interviews with customers to better understand their decision-making process in regards to a single product or service, I’d start with 5-10 customers.

  36. Gary Downey

    Says the webinar started at 1pm today EST but how do you access it?

    • Michael Aagaard

      Hi Gary,

      This is not a webinar. This is an Ask Me Anything Session where people can ask me questions by leaving a comment. So there is no webinar to access.

  37. Mazarine Treyz

    When you AB test sales pages, what are some things that you change that you’ve found lead to the quickest increases in conversions? For example, button color, adding a video, changing the headline at the top of the page?

    Thanks for taking the time to answer my question!


    • Michael Aagaard

      Hi Mazarine,

      Answering your users’ questions and increasing their motivation to convert are the things that really move the needle.

      In order to figure this stuff out, you have to put some time into doing your research.

      Your research might tell you that your headline is totally off, in which case I’d work on developing and testing a new headline.

      Your research might point to the fact that your users need an easy explanation of a complex aspect of your offer and that video is the right format to deliver this explanation. In this case, I’d develop a video and test it.

      I wish there was a simple answer to your question. Unfortunately there isn’t, and I’d be a terrible CRO Pro if I told you that there was one. It really depends on the given conversion scenario. Research always pays off, and doing it up front will save you tons of time in the long run. The better you understand your conversion problems, the easier it will be to solve them.

  38. Philip

    I have some questions about using tools like Optimizely or VWO for A/B testing.

    1) Choosing the sample in ecommerce
    When the tool randomly assigns the participants to the test groups, you cannot guarantee that the resulting distributions of e.g. revenue per customer within the groups are comparable.
    If you have some extreme shoppers in a group, it might impact the results. How big of an issue is that? What can be done about it?

    2) Using a ‘redirect experiment’
    With the above mentioned tools, you can either modify a page using JavaScript or build the variation with its own URL and redirect the participants to the correct URL.
    For redirect experiments, the server has to first dispatch the original page and then load the variation. So you lose maybe 200-300 milliseconds for the first page load.
    Doesn’t that put the variation at a significant disadvantage? Would you advise against using redirect experiments?

    3) Doing the statistics yourself
    At the CTAConf panel you said you did the statistics yourself instead of letting Optimizely/VWO do them. This year the two services updated their statistical models to sequential and bayesian models, respectively.
    Do you still calculate the statistics yourself? Can you match those statistical models? What are the advantages to doing it yourself?

    • Michael Aagaard

      Hi Philip – thanks for your questions. See my answers below:

      1) Choosing the sample in ecommerce
      Personally I don’t look at the results in the testing tool. The reporting is far too limited. I make sure to integrate my test data with my web analytics setup so I can use the full segmentation, dig deep and get the full picture instead of just the averages. This will help you understand the sample pollution as well.

      2. Redirect Experiments
      If you got a visual flicker going on it is definitely an issue. An issue of uncertainty, because you don’t know what impact the flicker has and that makes all your data shady. It can also affect your GA data if one variant URL loads before the other. The placement of the snippet in the source code and other technical aspects have an impact. I have hardcore developers helping me solve this stuff, so I rarely spend too much time fixing these issues due to the fact that development and code are not my primary competency : )

      You might find some inspiration in the ConversionXl post: http://conversionxl.com/server-side-vs-client-side-ab-testing-tools-whats-the-difference/

      3) Doing the statistics yourself
      Yes I still do it manually. You have to if you want to get your data from your Analytics setup instead of your testing tool. You can easily do your own sequential and Bayesian calculations if you have the right tools. I do.

      The advantages to doing it yourself is that you are no longer depending upon the limited data your testing tool spits out. I can look at whatever metric/dimension I want and get the data. The disadvantage is that you really need to know what you are doing. I’ve spent a long time learning all this stuff and I still have far to go. Luckily I can draw on some amazing team members to help sanity check what I’m doing here at Unbounce (Data Scientist with a PhD in String Theory and Data Analyst with a Master’s in Statistics).

  39. June Swatzell

    Hi, my friend Chuck Salina just published a text book and made this website – he asked me to help him improve it. Where should I start? My background is non-technical founder. http://www.powerless2powerful.com/about/

    • Michael Aagaard

      Hi June – first find out what is meant by “improve it”. What are the goals of the website? What is the current state of the website (get hard data on this)?

      You can only make something better if you know what better means and what the current state is.

      Getting this insight is your starting point.

  40. Daniel Chen

    I am interested in knowing which metrics you choose to focus on and how you communicate the hypotheses, story, insights and/or methodologies so I broke up one big question into 3 parts:

    Could we see an example of a report you would give an advertiser for a multi-channel ad campaign?

    In your analyses, which conversion attribution model(s) will you generally use, ie. last touch, first touch, multi-touch, impression assisted view-through or a combination?

    And as one last follow up question, how do you account for the impact of marketing efforts that are difficult to measure, such as billboards or radio?

  41. Johnathan Dane

    Hey Michael!

    First off, thanks for taking the plunge across the Atlantic and helping us North Americans out. I feel like I get higher conversion rates just because you’re closer to California now ;)

    I have a quick question for you:

    If you have 30 minutes to do conversion research prior to build a landing page from scratch, what questions would you ask to get the highest chance of conversion glory in your first attempt?

    • michael aagaard

      Hi Jonathan! Thanks for the kind words buddy!

      I’d ask questions like: Who is going to visit this LP? Which channel are they coming via? Which device are they using? What is their motivation for visiting the LP? Why would they yes or no to the offer on the page? What are the top three questions users will ask when they deciding whether to say yes or no? What information should we give them to help them answer those questions?

      These are some basic questions that I’d ask the client before starting on the LP if I only had 30 minutes.

  42. Abrar Shahriar

    Hi Michael,

    I am really enthusiastic and practitioner in CRO. I am eagerly wanted to know some advice / tips from you.

    1. What are the techniques that you would follow at your early stage of life in CRO?
    2. What are the tools that boost conversion at your point of view?
    3. What motivates you to take it to the next level or what is the reason behind your success?

  43. THB

    I can’t think of any questions that haven’t been asked yet, but thanks for the awesome info!