Best Practice Mythbusters #1: When Social Proof Doesn’t Work

One man’s best practice is another man’s conversion killer. The first instalment of “Best Practice Mythbusters” looks at social proof. In theory, showing off how popular you are should grok AIDA – grab Attention, spark Interest, create Desire and push visitors toward Action. But there are cases where social proof lowers conversion. But is there more to it? recently featured a case from which tested customer ratings vs. price:

Click images to enlarge

The “big brand price” version (control) won the challenge by “so much that had the losing variation been rolled out without testing first, it would have cost the company hundreds of thousands of dollars in revenues per year.” The test accounted for both online and telephone conversions.

At the time of writing, 63% of marketers voted for the social proof version…

Another example comes from

The middle version without social proof converted at 2%, the alternate versions at 1% and 0.9%, respectively.

The secret to social proof

Depends on…


Both experiment results, upon closer examination, are not so surprising when you consider the context. It comes down to the user’s primary motivator or primary FUD. FUD meaning fear, uncertainty or doubt. users likely would appreciate knowing how other users rate sellers. But ultimately, the customer wants to get the best deal. It’s not that social proof is not important, it’s just not as much of a motivator as saving the most money.

In the case of DIYthemes, how big your marketing list is is irrelevant. Subscribers are more concerned about email privacy. A follow-up test with an anti-spam callout could outperform, with or without social proof. (Screenshots via KISSmetrics).

Whether you are using testimonials, fan counts, Like badges, or star ratings – ask yourself if there isn’t a more important influencer of trust and action than social proof.


DIYthemes’ case may demonstrate how design affects conversion. Simplifying forms, reducing text, making things look easy all aid in conversion. Again, further testing (including challenging social proof against privacy) is required to answer these questions.

Putting the proof in social proof

The post-modern, post-marketing consumer is skeptical. Certain forms of social proof are more trustworthy, and therefore can have a higher impact on conversion. For example, testimonials that link back to a customer website is more believable than anonymized praise, like from “–Judy, Ohio.” Facebook and Twitter counts are verified by a third party, where simply throwing a number up on a site is not backed up by anything.

Bottom line

As each instalment of this series aims to illustrate, it’s the application of so-called “best practice” that matters. Social proof may work, or may not work, and variables like context, design and type of social proof affect the results you will get.

Related Articles

3 Responses to “Best Practice Mythbusters #1: When Social Proof Doesn’t Work”

  1. The Money Supermarket case doesn’t surprise me at all. It is a price comparison site after all. Customers are looking for the lowest price, especially for a commodity like car insurance, this is borne out by the low levels of loyalty in that market.

  2. Charles says:

    At first, I assumed the MS site had used “lowest amount first” on both, with the addition of customer rating instead of the generic “view” CTA.

    Although the user intent is a shorter path – i.e show me details and let me buy – I’d suggest the “lowest amount first”, supported by a proper customer rating (not just “how easy it was to give them my money”) would be something to test.

  3. Hi Linda,

    First of all – great work – and another great post!

    I know you havent had anything to do with the moneysupermarket test featured on WhichTestWon – but I think that its kind of obvious why the social proof variation bunked here.

    Its a price comparison site – hence price sensitivity is at the tip of the tongue – and the social proof variation has a huge price difference. The price variation however not only has the same price range as the list in the background – it also has logo match.

    I think its stretching it quite a bit claiming that social proof doesnt work there – havent you at all considered the possibility that its simply constructed way off base?

    I might be missing something really important here – so please correct me if I’m wrong :)


Leave a Reply

© 2014 Get Elastic Ecommerce Blog. All rights reserved. Site Admin · Entries RSS · Comments RSS