Marketer’s Braintrust’s Facebook Ad Type Test Results

VERSION A
MarketersBraintrust-VersionAsm
How site visitors voted:
VERSION A (52%)
VERSION B (48%)
Actual Test Results:(Click back to see Versions A & B)
Version A, the Link Page Post Ad (LPPA), increased website visits from this Facebook ad campaign by 64.7% at a 99.9% confidence level.

Marketer’s Braintrust, an online marketing think-tank, conducted this test in-house and used the Facebook Power Editor to create the ads and gather the data. All data was analyzed using in-house tech.

This test used the same image and copy. Image page post ads (IPPAs) share a larger image where the LPPAs decrease the image height and ad a blurb about the page the visitor will be seeing next.

In general, most marketers have preferred to use IPPAs because people online are captivated by images. Also from an organic perspective, Facebook has traditionally given more weight to image posts. However, the testing team wanted to see which of these two ads would perform better in their niche (remember we aren’t saying ONLY use LPPAs over IPPAs – just test it).

Another important note, if you are to test any Facebook ads, make sure they are not a part of the same ad group. Facebook automatically assigns more traffic to the better performing ad in an ad group, which skews test results. Facebook ad testing in a single ad group isn’t about which performed best, but which performed best earlier.

Click here to see another, albeit older, Facebook ad copy test

> Tweet that you ‘Got it Right’ here.
> Tweet that you ‘Got it Wrong’ here.

25 Comments:

  1. This was an easier one I think. Version B’s image is a visual dead end that really invites people to overlook the first link, whereas version A adds an extra link & text after the image and so allows those who skim-read past the first link to then get another chance to see the second one. Good test though.

  2. Some follow up questions: When you say LPPA drove more traffic, do you mean visits or visits/impressions? Do the IPPA and LPPA cost the same to run, or is one more expensive?

  3. So, I went with my gut and voted for B, which was wrong. Then I had to figure out why (of course). After I voted, it dawned on me that my gut reaction said, “go with the image, that always generates more interaction/virality, etc). So, my initial thinking was flawed because this test wasn’t a choice BETWEEN image and link; it was a choice between image only, or image AND link. If I had taken 10 seconds to think about it, I would have chosen “image AND link”. Here, more is more. I guess going with your initial gut instinct should be reserved for the GMATs.

  4. Although it is the same image, I found that the larger image had a softer less agressive mood to it. In B the croping accentuates the arms muscles and the b&w gives it a Movers or Bouncers feel rather than SEO experts. Person #2 has the most relax feel which can only be perceived if his relaxed left shoulder is visible. Emotional response to visual stimuli is altered by the croping effects.
    The text below contributes to accentuate the reassuring feel after a first (top copy is aggressive) impression.

  5. I actually liked version A because the photo looked more professional. With the extra space on the left, it had a more professional feel, which makes the guys pictured look more professional.

    Version B was cropped and close up, which I felt made it look more unprofessional, taken by their friend on his out-of-focus camera phone.

  6. Actually, unless I’m mistaken, the Link Page Post ad actually does in fact have more copy: beneath it. This helps considerably to build trust, explaining a little about who these guys are. It’s not the same image and copy. The LPP has more copy, and the crop is different in the two images. Same source, different crops. An interesting test, though, and I like your testing notes re: Facebook ads. :)

  7. I’m guessing that both versions won because the real target market here is those of us reading about this Facebook test. In other words, how many of you clicked on the Marketer’s Braintrust link?

  8. @Megan Bush-the Link Page Post Ad drove more visits AND impressions. In this test, the LPPA cost 1.45% more than the IPPA. The cost will vary greatly by target group, as will the best performing ad type. You’ve got to test to see which ad type has the best ROI in your niche.

  9. Am I really wrong? I voted for B (Image only), but, in my defence I based my decision on Advert A being targeted at just ‘Business Owners’ with the copy beneath – which is only 1 target group audience in a sea of people. Advert B wasn’t limited to any group and therefore I thought it would have wider appeal and better engagement.

    However, it dawned on me (after I voted), that if the adverts were only published to business owner Facebook pages, then the advert that mentioned ‘business owners’ below the image (Advert A) would have more appeal to Business owners. Can you clarify where the adverts were published to?

  10. A little background on why we did this test: In the past, we’ve had phenomenal success with Image Page Post Ads. We recommended starting with that ad type as the control against which to test. However, last fall, Facebook increased the size of the image displayed in a Link Page Post Ad from a tiny thumbnail size to 400×209. This was a game changer in Facebook advertising. In most niches we work in, LPPA out-convert other ad types, but not in *all* niches. Sometimes IPPA (with it’s more simple format) still convert the best.

  11. I think it’s worth noting: page post link ads go directly to the landing page when the image is clicked on whereas page post photos open the image in a lightbox. The page page post photo requires more steps to get to the landing page which I think at least in part was the reason the page post link ad performed better.

  12. Tom –

    Good point. From a usability standard this is extremely annoying to me. Every time Mashable posts an image page post ad I click the large image instead of the link in the supporting text. This, as you pointed out, brings me to the image lightbox and not the desired content I was expecting. More often than not, I move on and don’t try to get to the content again (unless it is extremely compelling).

  13. Like Paul above I too picked B for similar reasons, assuming the ad was being shown to general users rather than business owners. Personally I preferred A, as I like to have the extra info and the extra link, but I assumed that Facebook users would prefer B. I should never assume anything.

  14. I don’t use facebook enough to know… it sounds like these were in two completely different locations? Some context would be helpful.

  15. @Paul & Benjamin – the audience targeting for both ads was exactly the same, and consisted of business owners. The only difference is the ad type. We used the same copy in both ads, but as you can see, the ad type changes the display appearance & the layout. HTH!

  16. I voted for A because I thought the additional blurb under the picture appeared more as a “friend” sharing a link than an ad. It looks like a news story someone shared with you. Which I guess is the point. :)

  17. Aren’t the test results are also likely to be impacted by the FB algorithms preference for the LLPAs?

  18. I think there are four factors contributing to the lift… My guess is the second link below the photo is the biggest as others have pointed out with being taken directly to the landing page a close second (Thanks for pointing that out Tom – great detail!). Third, is the photo differences and fourth might be the fact that Version A says “give us ten seconds” (which helps to further qualify how much of an investment the survey is). Interesting test – thanks for sharing.

  19. @Dara-Facebook’s algorithm doesn’t prefer LPPA, per se. LPPA don’t universally convert better, which is why I test ad types. Every time I start with a new business I test ad types, and then I test them again at least every 3 months. Facebook is such a young ad platform that it changes practically daily so frequently testing everything is necessary.

  20. I think the simplest explanation is that A looks more like a blog post someone has posted to Facebook looks like, whereas B looks more like an advert.

  21. Fascinating and very useful – more tests on social media formats please!

  22. @Caroline

    We’d love to see some more social media tests, unfortunately either 1) people aren’t testing (which makes sense, FB makes it pretty difficult to test properly) or 2) there is such a competitive edge people do not want to share.

  23. Like Andrea Warner, I do not use Facebook enough, so some context would have helped. This said, unless I see a cut-off image B, it is not true what you write “This test used the same image and copy”. B did not speak about the real need Braintrust are targeting neither mentioned the real benefit they offer (lower textbox). If my point is true, Daniel Ogilvy (founder of the popular ad agency) would have been very pleased with this test results. In fact, A resembles closely its preferred and tested schema for print ads (apart from the text above): 2/3 of the page filled with an image, the rest is copy.

  24. Hi Mirio, Thanks for your comment. I wanted to let you know that I live & breathe Facebook ads.

    As for the tests, let me explain it a bit more. We took the same image & copy and fit them to Facebook’s requirements for 2 different ad types.

    Facebook displays the exact same image & copy differently, depending on the ad type & placement.

    We wanted to see which ad type converted the best in our niche. As you can see from the results, the Link Page Post Ad converted better.

    Seems like Daniel Ogilvy was onto something. :)

  25. Andrea, accept my apologies. I confused you with RH, who wrote ‘I don’t use facebook enough…’ And, thank you for your explanation. Even though I do not use FB myself, I live and breath mass communication, and, if you haven’t done so yet, I encourage you to read what Ogilvy wrote. It gave me loads of food for thought. One thing I was wondering about concerns the lower part of your copy, which sounds catchy to me. In your shoes, I would now turn version A in the control group and test in B the lower copy placed in different positions, like above the pic or just below the first line of text. Best is never good enough, right ;)
    All the best for your endeavors, Mirio

Join the conversation:

Please note:
1. Comment moderation is enabled, I'll do my best to approve comments asap.
2. If you have any trouble submitting comments, please send me an email (anneh at whichtestwon dot com).



Sign up for new test results & how-tos:


Privacy Protected