Join the Conversation Add Your Comment

  1. Martin

    Instead of just removing social sharing functions, I would first test removing only the counter, but only when there are let’s say less then five likes. Just perform an ajax call to FB before presenting the widget.

  2. wow! This is one of those articles that give you like a serious perspective and shakes you up to identify all that you’ve been doing wrong or about having the wrong approach. People blindly focus on getting traffic without caring much about conversions or dont seem to figure CTR out.

    In any case, an enlightening article.

  3. Interesting case studies. Not sure about the spontanious buyers argument for Brookdale’s. The fourth assumption looks probable. I have my own thoughts, but I’ll hold my tongue as I haven’t read the report. Not going to make any quick assumptions without doing any of the research ;)

  4. Denish

    I never tested a website, but here I am shocked, by removing social signals you can increasing conversions. but this is the real data so thank you so much for giving nice information.

  5. Bryan Clayton

    In depth analysis. Thank you for sharing these case studies. The trucking example is proof of a rigorous well executed CRO strategy.

  6. Sjove Linser

    I am new in testing, and your right always test and not use the generel guidelines :-)

  7. Changa Masomakali

    Hey Ott

    Thank you for this great post! You mentioned a confidence rate in some of the case studies. Could you expalain how you come up with that number and when do you typically use it?

  8. Tim Aucoin

    This is such an awesome post. Tip five was a little surprising, but it makes sense.

  9. Pawel Kontek

    Very good case studies. Number 3 surprised me the most! The conclusion is that you should always test everything.

  10. ISHIR

    Quiet interesting Post. I agreed with you. Sometimes using social sharing buttons unnecessarily everywhere doesn’t give good conversion rate, moreover websites selling products should focus more on selling rather than social sharing.

    By the way Good analysis/research. Keep up the good work.

  11. Jay Patel

    So the conclusion is “Less is more”. The Walmart case is really an eye opener. 98% increase on small devices is pretty bad ass. I wish “View detail” button removal conversion detail would have added some more figure to it.

  12. Keith

    More than anything the lesson from these studies is that testing of your site can not be an after thought. There is no way to tell by intuition what your visitors will like. Might that your visitors are all Facebook fans and love to use it as a log in. Might not. Have to test.

  13. @conversionhacks

    Good post Ott, I wish more big companies would post case studies of their tests for the industry. Would love to see a post from Amazon about all their testing for example.

  14. 2 things were counter-intuitive in these studies.

    1. I am surprised that offering Facebook login hurt conversions.

    2. The social icon study was also an eye opener.

    We need to remember that sometimes less is more.

  15. SEO Take Over Sk

    Great stuff, have not thought about the effect of the social media buttons on the ecommerce pages. They dont have count on how many times the page was shared, but might be distraction. Going to test it :)

    1. Ott Niggulis

      Good luck for testing, let us know of the results ;)


Leave a Reply

Your email address will not be published. Required fields are marked *

Current article:

6 Conversion Rate Optimization Case Studies With Surprising Results