Leave a comment
Get the GH Bookmarklet

AMAs

I’m Peep Laja, a conversion optimization fanatic. I run ConversionXL - the world's most popular conversion blog - and I'm the co-founder of Markitekt, a conversion optimization agency. I also teach and coach other optimizers + am throwing kickass conversion confernce called ConversionXL Live (http://live.conversionxl.com). I've worked with hundreds of companies around the world - small and huge - to improve their revenue per visitor, and can help you do the same. You can ask me anything, including stuff about * the process of conversion optimization * collection data, finding insights from the data * google analytics * running a/b tests * getting more people to take whatever action * persuasive design * running a conversion agency * building a blog * what's with the weird name I can also do quick site reviews (like 5 comments about your site). Follow me on twitter at @peeplaja and subscribe to optimization newsletter over at http://ConversionXL.com

  • TW

    Tommy Walker

    almost 5 years ago #

    Aside from A/B testing, what other kinds of tests can be run. Specifically for sites that are:

    A.) Low Traffic
    B.) Pre Product/Market fit

    • PL

      Peep Laja

      almost 5 years ago #

      Lets get one thing clear: testing is NOT optimization. If you don’t have enough traffic to test, you can still optimize.

      1. If you don’t have enough traffic (meaning it would take more than a month to run 1 test, calculate needed sample sizes in advance http://evanmiller.org/ab-testing/sample-size.html), then too bad. You can’t test faster, math doesn’t work that way. You could use bandit testing (what Google Content Experiments or Conductrics use as their testing algo), but it would still be way too slow way to learn.

      Testing is about learning, so if it takes more than a month to learn something, that’s not good enough.

      But you can still optimize. You might not have have a lot of quantified data, but you can always do user testing and heuristic analysis.

      Here’s a post I wrote on optimizing low traffic sites:

      http://conversionxl.com/how-to-do-conversion-optimization-with-very-little-traffic/

      Essentially you gather as much data as you can to be confident, and roll out all the changes at once – and hope to see a change in results you can believe in (e.g. used to get 10 leads / day, now it’s 20 leads / day). Smaller changes 10 vs 12 / day can be due to traffic fluctuations, so you can’t be sure that your changes made it better (could have made it worse too).

      2. Customer development is going to be your best bet – qualitative interviews. But can’t trust people when they say “I’d buy that” – people are unable to predict how they will act in the future, especially when it comes to taking out their wallets.

      You’d need to tweak the offer, and see if that gets more people sign up etc.

    • SE

      Sean Ellis

      almost 5 years ago #

      I've been thinking about this a lot lately too. This morning I actually wrote down the name "rollback test" (not sure if that's a thing) for tests that are too expensive to run as an A/B test. By too expensive I mean that instrumenting it as an A/B test takes way more effort than just doing it, seeing results and rolling it back if it's a mistake.

      Obviously this can be dangerous because there are a number of variables that can give you false positives/negatives that wouldn't exist with an A/B test (traffic sources, days of week...) Also, a "rollback test" requires discipline to define the report that you'll be making decisions on, and schedule some kind of notification to actually analyze the results.

      But I've done a lot of "rollback testing" over the years and feel like in aggregate it was an important form of testing that led to some successful companies.

      Curious if @peeplaja thinks it's ever valid to approach testing this way or if it should always be set up as an A/B test regardless of the cost.

      • PL

        Peep Laja

        almost 5 years ago #

        That is the way to do it if you have no traffic. BUT - the changes you see have to be big enough as the variables are not controlled with this approach.

        Look at your daily conversion rate fluctuation and stuff. So if you change stuff and it gets 10% better, it might be because you're actually getting better traffic, and not because of those changes. So you want to see like a 30%, 50% or even 100% lift :)

        That also means that you can't change little things. You need to be bolder with changes. If things get 2% better, you can't see it.

    • DL

      Dylan La Com

      almost 5 years ago #

      +1 great question

  • MB

    Morgan Brown

    almost 5 years ago #

    Thanks for doing this Peep. What's the number one mistake people make when trying to improve the performance of their website?

    • PL

      Peep Laja

      almost 5 years ago #

      The biggest issue is that people don't know what to do, or where to begin. So they try random stuff, use the spaghetti method - see if anything seems to work.

      It's typical for inexperienced marketers to think in terms of tactics (stuff they read in a blog post) instead of processes.

      People ask me "where do you start optimizing for ecommerce" or "what to focus on on a travel site". These are the wrong questions. You should start wherever you're leaking the most money.

      One you have a systematic approach to optimization (as opposed to ad hoc random shit), you'll get much better results.

  • LS

    Logan Stoneman

    almost 5 years ago #

    Hey @peeplaja , awesome to have you here!

    If you were giving a lecture to a group of college students in a senior level entrepreneurial class, what tools, books, or philosophies would you share with them? Your goal is to inspire growth, both in their professional pursuits in building their idea and in creating a personal growth mindset.

    Oh, and what's with the weird name :)

    • PL

      Peep Laja

      almost 5 years ago #

      You inspire with what's possible - through sharing stories. I could share many stories from my life, about working in Dubai, finding myself unemployed in Panama with no money, starting businesses and failing, always optimizing - until you achieve something you can be proud of :)

      And in terms of the mindset - I think one's success is determined by proactivity. Always taking responsibility and asking "what can I do to change the situation?". A great book to accompany that mindset is Stephen Covey's "The Seven Habits of Highly Effective People".

      About the name:
      Where I'm from (Estonia) I have a very normal, common name. And you'd be the one with the weird name! :) It's all a matter of perspective. It's like asking why the Chinese don't use the "normal" alphabet. So like the Roman alphabet is not "usual" except for the languages that use it, my name is weird in countries that have different names.

      It's pronounced Pep Laya, and it has always been a great conversation starter. At first I hated the way people keep pronouncing it in English, but then I turned it around, and made it something positive.

      • LS

        Logan Stoneman

        almost 5 years ago #

        Thanks Peep! As a follow up, would you suggest young professionals, or any age, to have international experience like you've had in Dubai or Panama? If yes, why and how can one take the 'daunting' leap.

        • PL

          Peep Laja

          almost 5 years ago #

          Oh man. Living and working abroad is the best thing a person can do.

          When you get taken out from your usual context - the people you're with, the restaurants you frequent, the environment you're in - and go to a new place - it's a chance to find out who you really are. The experience brings out your true self like nothing else.

          It also teaches you to question stuff. Things you deem "normal" are not normal in other countries. You'll get new perspectives, and you'll often find yourself thinking "oh I didn't know this could be done in another way".

          The more different the foreign country, the better. Moving from the US to Canada, or from Mexico to Guatemala is not going to be that different. So go further!:)

          How to go?

          A) If you're still in college, join AIESEC - the world's largest student organization (125+ countries): http://aiesec.org/

          I was a member, and it changed my life. The first time I went to live abroad was through AIESEC
          B) Volunteer with a non-profit
          C) Just pick up and go! Digital nomading ftw

          If you have a job that forces you to stay in the same place, know that it's an ancient concept and you don't have to keep that stupid job!

          • LS

            Logan Stoneman

            almost 5 years ago #

            This is awesome stuff @peeplaja - thanks for details and answering all the questions! Definitely solid advice for those having trouble finding themselves while they repeat the same things over and over again.

            I guess you could say the same thing about those who run the same A/B or conversion tests over and over again expecting different results.

            Cheers!

  • ET

    Everette Taylor

    almost 5 years ago #

    My man Peep, good to have you bro.

    If you could suggest one optimization improvement to test on GrowthHackers.com, what would it be?

    • PL

      Peep Laja

      almost 5 years ago #

      Sure. First tell me which metric do you want to improve?

      • ET

        Everette Taylor

        almost 5 years ago #

        Engagement per post, engagement being a comment or up vote.

        • PL

          Peep Laja

          almost 5 years ago #

          I'd test improving readability. Increase your font size, line height etc, maybe even change the font. Definitely no bold in the replies :)

          If stuff is easier to read, more people read and thus might engage more.

          • DL

            Dylan La Com

            almost 5 years ago #

            Great idea @peeplaja. Any particular font(s) come to mind that would be an improvement?

            • PL

              Peep Laja

              almost 5 years ago #

              Oh most standard web fonts are good. Pick a combo you enjoy reading on other sites (e.g. Medium etc)

          • NS

            New Svasti-Xuto

            almost 5 years ago #

            Great idea. To be honest, people not familiar with this type of idea board might find it difficult with what the up vote arrow here can do... Like the hamburger icon, people act to it differently.

          • DL

            Dylan La Com

            almost 5 years ago #

            Yo! @peeplaja thanks again for the idea.

        • MH

          Mark H

          almost 5 years ago #

          Add one of those iframe toolbars that show up on top with an upvote button and a comment button upon clicking an article.

          Because once we click the article link, we go outside the realm of GH and sometimes forget to come back and upvote or comment.

  • DB

    dan barker

    almost 5 years ago #

    hi, Peep, how is everything going?

    I have 2 quesitons: one straight-forward one & one rude one-

    1) My question is - aside from the bombproof content on ConversionXL - how much do you trust the case studies you see around conversion optimisation, and how much do you think is either PR or simply people reporting results where they have not quite understood the numbers & levers properly?
    2) Your blog ConversionXL has a massive brand in the space, your company name is less well known. Why haven't you rebranded the overall business to ConversionXL?

    I hope life is good, and hope to see you soon,

    dan

    • PL

      Peep Laja

      almost 5 years ago #

      Hey Dan, nice to see you here.

      1. If absolute numbers are not shared, I'm always doubtful. Especially if it's only "I tested it, and got +30% lift!" with no specifics given. Also, knowing how bad people are at calling tests, I bet half the case studies out there are bollocks - people making conclusions based on like 30 conversions or something.

      2. Good question. The idea has always been that I use ConversonXL only to build an audience - as what I sell / do might change down the line. I don't know if I'll have an agency 5 years from now, maybe I'll move into online education business.

      But I might change my mind. It's one of those things that is very difficult if not impossible to split test.

      Would I get more business if the agency was called ConversionXL? You tell me.

  • BL

    Brian Lang

    almost 5 years ago #

    Hi Peep, what are your thoughts on alternative approaches to interpreting results, as well as running A/B tests? Specifically, the frequentist approach (statistical significance & statistical power) compared to bayesian methods?

    • PL

      Peep Laja

      almost 5 years ago #

      I do recommend running bandit tests on stuff where you can't learn from about the audience (e.g. optimal button color, font size) - what matters is whether we're making money, and using frequentrist approach for tests where I can learn from.

      I'm excited to see what the new Optimizely stats engine is gonna be like, using http://en.wikipedia.org/wiki/Sequential_probability_ratio_test

  • RS

    Rob Sobers

    almost 5 years ago #

    Hi Peep,

    Two questions:

    1.) Do you have any evidence that persuasive/long-form copy works well for enterprise audiences that are used to the traditional feature/benefit listing approach?

    2.) What do you do when a consulting engagement doesn't live up to expectations? That is, you don't hit your conversion targets for the client.

    Thanks!

    • PL

      Peep Laja

      almost 5 years ago #

      1. Enterprise or not, everyone is a human being. Also, long form copy only works if it's good, not because it's long. Don't confuse length with depth.

      If it's a complicated product, long form copy is definitely helpful if they buy the product right on the website. If the goal is to get a lead / demo request, you'll need less copy. It's about the size of the request you're making.

      2. We part our ways. Every single conversion company out there has had cases where they failed to achieve the desired result. There are no magicians out there. It doesn't happen often, but it does.

      • WT

        Willie Tran

        almost 5 years ago #

        "Don't confuse length with depth." Awesome.

      • RS

        Rob Sobers

        almost 5 years ago #

        Thanks for the super honest answer to #2. Can be a tough pill to swallow and not many firms are willing to admit that.

        With #1, I think persuasive copy can repel some groups of humans who are particularly guarded against overly sales-y websites. I've found this to be true, at least.

        • PL

          Peep Laja

          almost 5 years ago #

          Now - overly sales-y and persuasive are the opposite things :) Hyperbole and your typical "internet marketing sales page" stuff are crap. Copy needs to be good. It's not persuasive if it doesn't persuade! Hype does not make copy persuasive.

          Yes, writing good copy is very difficult, especially good long form copy. Most copywriters suck ass. Don't hire cheap copywriters!

          • AC

            Adam Casey

            almost 5 years ago #

            Hi Peep, When you run test, do you create your own copy or do you hire a copywriter? Or does it depend on the test?

            • PL

              Peep Laja

              almost 5 years ago #

              I typically create my own copy. I hire help when I run out of ideas, or if I need long form copy written. @jenhavice is my go-to copywriter

              • JH

                Jen Havice

                almost 5 years ago #

                Many thanks, Peep!

                I have to agree with Peep on this. Being persuasive should translate into giving your customers everything they need to achieve their goals on the page. That means you've got to speak to their motivations while alleviating their concerns and hesitations.

                Sometimes that requires developing an argument that goes on for pages. Sometimes you can do it in a page. It goes back to understanding your customer and their current state of awareness not only of the problem you're trying to solve but of the solution as well.

                Don't confuse this with slimy copy that makes your skin crawl. Great website copy sells without making people feel like they're being sold to. It works with the design to make it as easy as possible for the visitor to decide and then take action.

                Think of persuasive copy in these terms and it may change your view on it.

  • SE

    Sean Ellis

    almost 5 years ago #

    Hey @peeplaja thanks for doing this. Which company have you seen that you are most impressed with their approach to conversion optimization? What is their team structure, cadence and use of external experts?

    • PL

      Peep Laja

      almost 5 years ago #

      Everyone is raving about Zalando, the European ecommerce startup. They came out of the blue, and in the last couple of years have grown bigger than almost everyone but Amazon.

      Their secret? They have like 80 optimizers on staff.

      I don't know details about their internal structure, just the legend :)

      Maybe you guys should do an AMA with their head of optiization (Jan Neutard, Head of Conversion Rate Optimization, Zalando Gmbh)

      • PB

        Plamen Barzev

        almost 5 years ago #

        I've never heard of Zalando (cuz I live in the stupid small country out of time), but their homepage is so SEO'd that it hurts.

        Still, the shop functionality is great. The bundles probably can help a lot. I can steal a lot ideas from this website, and then try them in different niche.

        Thanks for mentioning it Peep.

  • DL

    Dylan La Com

    almost 5 years ago #

    Awesome to have you on @peeplaja

    Are there any web design trends today that you've found are particularly bad for optimization and the sites are generally better without?

    • PL

      Peep Laja

      almost 5 years ago #

      If you consider automatic sliders still a trend, that's typically terrible due to distraction.

      VIdeo backgrounds also distract - and can hurt conversions.

      Parallax scrolling - same thing (depends on the implementation).

      • LS

        Logan Stoneman

        almost 5 years ago #

        I know AirBnb had/still may have video backgrounds on their home page. Are they just creating a brand or do you think there is might be a way those video backgrounds can be used effectively in conversions?

        • PL

          Peep Laja

          almost 5 years ago #

          It's just more difficult to do so. Everything can work - there are no thing that always work or never work - but the odds are just against you. Know your risks. And certainly don't do video backgrounds etc just because you think its cool. Your opinion doesn't matter.

  • GG

    Gail Gardner

    almost 5 years ago #

    What advice do you have for marketers who want to learn to do testing? Is there any efficient way to learn it? There is a shortage of talent in this area. How best can people who are interested gain knowledge and experience?

  • DL

    Dylan La Com

    almost 5 years ago #

    @peeplaja are you doing cohort analysis with Google analytics? If so, what are some tips for setting this up. If not, what other tool are you using?

    • PL

      Peep Laja

      almost 5 years ago #

      I haven't yet as it's kind of new, but will get into it soon. Mixpanel.

  • VV

    Viktor Vecsei

    almost 5 years ago #

    We run a high traffic (5M+ pageview) video sharing / tube site, think YouTube with niche content. Our aims are reducing bounce rates and increasing user retention - no product to sell.
    1. What pain points would you identify for a sites like this, what would you measure and how would you use that data to improve?
    2. What tools would you recommend beside Google Analytics for sites with huge visitors numbers and low customer value? We have too much data and our current budget won't allow purchasing premium.

    Thanks!

  • WT

    Willie Tran

    almost 5 years ago #

    Hey Peep!

    Thanks for doing this! Sorry to stray away from the questions you mentioned above, but I wanted to know what you thought of Austin, TX?

    How does it compare to San Francisco?

    Thanks!

    • PL

      Peep Laja

      almost 5 years ago #

      Austin is awesome. It's often referred to as San Francisco of Texas. It's much cheaper in comparison. Less pretty I think - but similarly green: in terms of nature, mindset of the people and smoking. It's cool to be weird. If your Whole Foods checkout person is not fully covered in tattoos, that's weird. Great music scene, lots of entrepreneurs.

      And - it's the conversion capital of the world. Look it up.

  • AA

    Anuj Adhiya

    almost 5 years ago #

    Hey @peeplaja
    Thanks for all the great responses.

    Let me ask the Peter Theil-style question:
    What's the one thing you know/believe to be true about CRO/Testing etc that most others do not?

    • PL

      Peep Laja

      almost 5 years ago #

      I haven't conducted a survey to know what most people think is true or not. So not really sure about this.

      If my views seem contrarian to some (e.g. every day without a test is a day wasted, test longer, ignore opinions, processes instead of tactics) , it's mostly not because of lack of education on their part, and less because my views are unique.

  • SV

    Sascha Pascal van Opzeeland

    almost 5 years ago #

    Hi Peep, thanks for sharing your expertise!
    Perhaps you could give me advice on this: In our SaaS we are thinking about switching from Google Analytics to a solution like KissMetrics, RJMetrics, or Inbound. Do you have any insights on what the strong & weak points are of these solutions compared to Google Analytics?

    • PL

      Peep Laja

      almost 5 years ago #

      Do NOT switch. Add. Always keep GA there, it's freaking amazing - and you never know when you need it. It does a bunch of shit other tools don't. There is no additional cost or load - no reasons for removing it.

      And if you run split tests, doing post-test analysis with GA is a must.

  • IM

    Ivan Muljevski

    almost 5 years ago #

    How, where and when do you learn new things?

  • SS

    sabrina saada

    almost 5 years ago #

    Hi Peep!
    If you could go back and do one thing differently when launching your first website/blog what would it be? Thanks for taking the time to answer these!
    ~Sabrina

Join over 70,000 growth pros from companies like Uber, Pinterest & Twitter

Get Weekly Top Posts
High five! You’re in.
SHARE
50
50