One tiny little layout change netted us an almost 20% increase in on-page conversions.<\/p>\n
They say a butterfly flapping its wings can cause a rainstorm on the other side of the Earth. Well, today\u2019s story makes me think that maybe that\u2019s true.<\/p>\n
Conjurer of conversion, Rebecca Hinton, is back with another test you can try for yourself. This time it\u2019s all about social proof and where it belongs on your page.<\/p>\n
<\/a><\/p>\n But before you jump straight to the how-to, I\u2019ll also cover why you shouldn\u2019t <\/em>run off and do this without testing it first.<\/p>\n It\u2019s pretty non-controversial for both B2Cs and B2Bs to add social proof to their websites in order to boost conversion, right?<\/p>\n You know what I\u2019m talking about. Testimonials from happy customers. The logos of your biggest clients. Photos of your legions of fans gleefully offering up their firstborn children. (Too far?)<\/p>\n So why is social proof always tucked away at the bottom of the page? Or hidden on its own page like some secret Victorian wife in the attic?<\/p>\n \u201cWhile social proof is helpful, chances are it\u2019s not what your visitors came for. And you don\u2019t want to push what they came for right out of their radar.\u201d<\/p>\n That\u2019s Rebecca Hinton, the CRO strategist behind this and many other incredible conversion optimization successes at HubSpot. (If you\u2019ve been following this column, you\u2019ll know Rebecca\u2019s also responsible for the test that boosted paid ad CVR by 11%<\/a>. And I\u2019m certain this won\u2019t be the last one I share. She\u2019s that good.)<\/p>\n But while her point makes perfect sense, it poses a conundrum: Visitors actually need to, y\u2019know, see all those testimonials in order for them to work.<\/p>\n \u201cAccording to heat maps, only 50% of users scrolled far enough to see the social proof,\u201d Rebecca explains.<\/p>\n\n But work it does! Despite the fact that only half of our visitors saw them, the social proof sliders were still the second and third<\/em><\/strong> most-clicked elements on these landing pages.<\/p>\n So pop quiz, hotshot. If you move your social proof higher on the page, it pushes your marketing content down. But if you don\u2019t, nobody sees it. What do you do? What do you do?<\/p>\n Whenever we have a dilemma, Rebecca\u2019s answer will always be: Test it and find out.<\/p>\n For the control group, she kept the page as it was, with a social proof module containing both customer logos and written testimonials near the bottom of the page.<\/p>\n For Variant B, our CRO team separated the two. The customer logos became a small and unintrusive slider tucked right up under the hero banner, while the testimonial portion stayed near the bottom of the page.<\/p>\n\n This would hopefully strike the best of both worlds: More visitors would see some form of social proof, but the slim line of logos wouldn\u2019t push our marketing content down too far.<\/p>\n But is a small line of logos really enough to make a difference?<\/p>\n Turns out, the answer is not only \u201cyes,\u201d but a 19.5% increase in software signups that says \u201chell<\/em> yes.\u201d<\/p>\n Rebecca says it with a little more tact.<\/p>\n \u201cPutting a band with customer logos lends an air of legitimacy. Oh, they have customers with logos that I recognize? That builds trust.\u201d<\/p>\n Now, some number of you are ready to run off and slap some logos on your site just because \u201cHubSpot said so.\u201d So here\u2019s where I share the scary part of the story to convince you to test it first.<\/p>\n Rebecca had a hunch she could do even better, so she took a bigger swing with Variant C.<\/p>\n In this version, the logos still moved up to the top of the page, but she swapped the customer testimonials at the bottom for data points about customer success. Think of brief stats like \u201cAfter 1 year, HubSpot customers closed 55% more deals.\u201d<\/p>\n \u201cAnd I thought, incorrectly, that this was going to outperform the testimonials,\u201d she explains with grace and good humor. \u201cBecause when I read them, I find them to be very compelling. You attract 114% more website traffic. You generate 129% more inbound leads. To me, that\u2019s compelling. That\u2019s exactly what I want. Let me sign up!\u201d<\/p>\n Turns out that the opposite<\/em> was true. Variant C reduced conversion by nearly 10%. Oof.<\/p>\n Rebecca believes that the negative result is due to the fact that testimonials are from customers, while data points come from the company itself.<\/p>\n \u201cPeople trust people more than they trust companies,\u201d she says. \u201cThey have that healthy skepticism, which is totally fair. And that\u2019s why we test, right?\u201d<\/p>\n And that\u2019s why you should test it, too.<\/p>\n If you\u2019re ready to try this out, Rebecca\u2019s got some tips for you to consider.<\/p>\n \u201cI always think my ideas are good because they\u2019re mine. But my idea was to add the customer stats, and that lost,\u201d she shrugs. \u201cWhat helps is to have an insight based on data, rather than a hunch.\u201d<\/p>\n In this case, the data-backed insight for the test came from checking out heat maps. When Rebecca noticed that only 50% of our visitors saw the social proof module, the test she devised was a logical next step.<\/p>\n So instead of simply mimicking this test, take a look at your data and see what insights might inform a brand new test.<\/p>\n \u201cIf we were targeting a page that had a lot of returning traffic, or was deeper in someone\u2019s customer journey, I would not expect social proof to be as effective.\u201d<\/p>\n Part of the reason that this test worked was because it was on pages that targeted brand-new visitors. If your target page is talking to returning customers, they may not care about logos or the recommendations of other customers. That audience might actually be swayed by data points instead.<\/p>\n The only way to know is to consider what a visitor expects to encounter on a page like that, and then test, test, test.<\/p>\n \u201cIf we get a test win, we implement, then we wait two weeks,\u201d Rebecca explains. \u201cThen we find the data before the test launched (because you don\u2019t want to include test data) and we compare the before and after period.\u201d<\/p>\n Don\u2019t think of this as a second test \u2014 it doesn\u2019t need to be that scientific. This is more about making sure there are no unintended consequences.<\/p>\n \u201cWe\u2019re not looking to match up with the original test results. That\u2019s not realistic or reasonable. There are too many external factors. All we\u2019re looking for is to see that your results directionally<\/em> line up.\u201d<\/p>\n Since your newly made changes will probably affect a wider range of pages than just your test, there could be room for unforeseen problems.<\/p>\n \u201cIf you don\u2019t do that before and after check, issues can fester for months.\u201d<\/p>\nButterflies and Best Practices<\/h2>\n
Split Testing the Split<\/h2>\n
The Mysterious Variant C<\/h2>\n
Tiny Tweak Takeaways<\/h2>\n
1. Start with an insight.<\/h3>\n
2. Consider visitor intent.<\/h3>\n
3. Double-check your results after implementation.<\/h3>\n
How to Make a Logo Slider<\/h2>\n