Our CRO team made a change that lifted the performance of our paid ads by almost 11% \u2014 and they didn\u2019t have to make any deals with supernatural beings to do it.<\/p>\n
So while I\u2019m a little bummed that I don\u2019t get to use my Ouija board, the good news is that you\u2019ve already got everything you need to try this out for yourself.<\/p>\n
In fact, this tactic is all about what you don\u2019t <\/em>include on the landing page. Below, I chat with our sorceress supreme of conversion rate optimization to find out what that cryptic advice actually means.<\/p>\n <\/a><\/p>\n But first, a grave warning \u2026<\/p>\n Before you go cutting content from your website, a word of warning:<\/p>\n \u201cYes, this works for us,\u201d says Rebecca Hinton, CRO strategist and principal marketing manager at HubSpot. \u201cBut it may or may not work for you, so you always want to test it.\u201d<\/p>\n Rebecca\u2019s tests have helped my program hit triple-digit growth, so I\u2019m going to politely insist you take her word on that.<\/p>\n At HubSpot, we never dive into changes without having the proof to back it up, and neither should you. Your audience could react very differently from ours.<\/p>\n The tactic I\u2019m about to share came from the results of a rigorous experiment, and later on, I\u2019ll show you how to run one just like it.<\/p>\n Okay, now onto the good stuff.<\/p>\n The first change, surprisingly, was to stop sending paid ad traffic to our product pages. Why? Because those pages have too many jobs already.<\/p>\n \u201cYour website has to appeal to all of your customer personas,\u201d Rebecca explains. \u201cPeople who are new, people who are seasoned, people who are already customers.\u201d<\/p>\n That adds up to a lot of content. And for visitors who landed on your site via a paid ad, it’s a lot of distraction.<\/p>\n To illustrate her point, Rebecca gives the example of a user clicking on an ad that says \u2018Download our ebook.\u2019<\/p>\n\n \u201cIf you were to send them to a website with full navigation, maybe they get distracted, maybe they feel like \u2018Ah! I just wanted the ebook!\u2019\u201d She throws her hands up in the air in mock frustration.<\/p>\n \u201cBut with a dedicated landing page where the primary CTA is about downloading the ebook, now they\u2019ve had a logical experience.\u201d<\/p>\n So her team set out to make a dedicated landing page for each ad being tested. But, as I mentioned above, what\u2019s on those pages isn\u2019t nearly as interesting as what isn\u2019t. <\/em><\/p>\n And what isn\u2019t there is about 90% of our website\u2019s navigational links.<\/p>\n Here\u2019s a screenshot of what one of our product pages currently looks like:<\/p>\n\n Like your Gran\u2019s holiday dinner, there\u2019s a little something for everybody.<\/p>\n Now here\u2019s the paid ad landing page for the same product:<\/p>\n\n If we stick with the food metaphor, this one would be a working lunch. You get exactly what you came for and you get it fast.<\/p>\n On the dedicated landing page, visitors can only sign up or request a demo. (Or leave, I suppose. But let\u2019s think positive.)<\/p>\n \u201cThe point of a landing page is to focus users, so we don\u2019t really want to be linking them out everywhere,\u201d Rebecca says.<\/p>\n And the proof is in the results \u2014 which I\u2019ve actually been underselling, because one regional market saw an incredible 83% increase in CVR.<\/p>\n Even if you\u2019re not sold on nixing the nav, you should still be using dedicated landing pages. Rebecca explains why:<\/p>\n \u201cIf we send paid ad traffic to product pages, we can\u2019t do any CRO testing. I don\u2019t own those pages.\u201d<\/p>\n Chances are, your team isn\u2019t the only one with an interest in your product pages. That can limit what you\u2019re allowed to change, add, or experiment on.<\/p>\n By creating a dedicated landing page, you\u2019re also creating a sandbox you don\u2019t have to share. (The dream of every middle child.)<\/p>\n \u201cWe\u2019d call that a win even if the results were flat because it opened up green space for future testing.\u201d<\/p>\n But now that I\u2019ve covered what not to include, what should <\/em>you put in those shiny, new green spaces?<\/p>\n Since the details depend on your business and what you\u2019re advertising, you\u2019ll need to do a little experimentation. But Rebecca\u2019s got some tips to get you started \u2014 and they upend what I\u2019ve always heard about A\/B testing.<\/p>\n Most A\/B testing guides tell you to pick one small change at a time. And if you\u2019re just trying to optimize an already high-performing page, that\u2019s sound advice. But to get these results, Rebecca tossed that out the window.<\/p>\n Start with wildly different versions that will quickly identify trends within your users\u2019 preferences.<\/p>\n \u201cYou want to take big swings, and say, \u2018These pages are radically different, and it looks like our users are more attracted to this one.\u2019\u201d<\/p>\n Once you\u2019ve got a clear winner, then you can narrow in on smaller details like color choices, CTA language, image placement, etc.<\/p>\n\n Many landing pages mistakenly swing to one of these extreme opposites:<\/p>\n \u00a0Marketers assume conversion will happen on its own, and include too little content. \u201cThink about the journey starting from seeing your ad to taking the action you want them to take.\u201d<\/p>\n While you want a clear path to the CTA, your landing page should also include content that helps guide that journey.<\/p>\n That may take the shape of testimonials, trust indicators, customer stats, or other kinds of social proof. It may be language that romances the call to action. It may even simply be basic company info.<\/p>\n \u201cSet the stage before you dive into a specific product.\u201d<\/p>\n The exact details will depend on what you\u2019re advertising, but no matter what you include, make sure it creates a logical path to conversion.<\/p>\n But remember that the journey doesn\u2019t stop at the conversion.<\/p>\n \u201cWhat\u2019s the post-conversion experience?\u201d Rebecca asks. \u201cIs it a thank you page? Is it a purchase confirmation?\u201d<\/p>\n If the landing page is dinner, your post-conversion confirmation is dessert. Nail this, and your visitors will come back for more.<\/p>\n I\u2019m definitely guilty of this one. If you only have two choices, one of them is going to win, right?<\/p>\n Not necessarily. You could have an inconclusive test with equal results. You could fail to get statistical significance. Your visitors could reject both <\/em>choices.<\/p>\n \u201cIf you have the traffic to support it, test a couple different landing pages,\u201d she says.<\/p>\n More variants won\u2019t necessarily guarantee a winner, but they will help you work through your options faster.<\/p>\n Just be sure you\u2019re not spreading your traffic too thin. Which brings me to the next point \u2026<\/p>\n With too small an audience, your results could just be random chance. Did landing page two really convert better? Or did it just happen to get the visitors who were ready to click?<\/p>\n To know that, you need to make sure your test reaches statistical significance<\/a> (the probability that your results are due to real factors and not chance.)<\/p>\n Without considering significance, Rebecca\u2019s test may have sent us in the wrong direction entirely.<\/p>\n While her experiment increased the number of signups<\/em>, it actually appeared to reduce the number of demos <\/em>by 11.6%.<\/p>\n However, while the team was 99% confident in the signup conversion results, they only achieved 64% significance for the demo results.<\/p>\n (There\u2019s no magic target for statistical significance, but the higher the number, the more confident the results. Imagine crossing the road if you were only 64% confident a car wasn\u2019t barreling toward you.)<\/p>\n So if we hadn\u2019t considered significance, we may have been spooked by the reduced demos and chosen the wrong landing page.<\/p>\n\n Okay, but how do you figure out what\u2019s significant?<\/p>\nMeasure Twice, Cut Once<\/h2>\n
What She Cut<\/h2>\n
How to Make Landing Pages that Land<\/h2>\n
1. DON\u2019T test one element at a time. Start with big swings and radical changes.<\/strong><\/h3>\n
2. Consider the journey, and not just the destination.<\/strong><\/h3>\n
\n \u00a0Marketers assume they need to convince every visitor and include too much content. <\/p>\n3. Don\u2019t assume you\u2019re going to get a winner.<\/strong><\/h3>\n
4. Pay attention to statistical significance.<\/strong><\/h3>\n