Banner Ad Image Test
| VERSION B | |
How site visitors voted:
VERSION A (71%)
VERSION B (29%)
|
Actual Test Results:
(Click back to see Versions A & B) Version B – with a right-side button, sans headshot – boosted on-demand webinar sign-ups by 65%. (Note: Never judge a banner test only by clickthroughs; as demonstrated here, what really matters is the conversion activity downstream after the click.) Optimeister (recently acquired by QUISMA), ran this A/B test in-house, placing the banners on their own site and blog, using Vertster’s testing technology. So, visitors were highly qualified responders. The copy in both cases was identical and fairly compelling. For example, the buttons read, “View now for free.” The headshot in the losing version was of German analytics and testing expert, Dennis Kruger, of whom some visitors would have heard. Still, in this case, an expert’s headshot didn’t do the trick. Kruger told us, “This clearly shows that including a picture of yourself is something that needs to be tested! In our experience, unless you are super hot and wearing a bikini, or super famous, the picture is better left off your page. What is interesting to note is that moving the button to make room for the image could have actually helped the headshot version because in western countries we read left to right and left button placements often outperform right placements.” To which we reply: Dennis, anytime you want to test wearing a bikini in your banners, we’re looking forward to publishing the results. |
Great example!
Do you have any data regarding the CTR rates of both banners?
FOILED AGAIN! LOL!
Excellent test Anne! How interesting it is to see conventional usability wisdom get tossed onto it’s head.
All Hail Split Tests.
Interesting test – i got it worng!
Interesting indeed. I thought humane faces resulted in more clicks. Apparently not.
I had this wrong, I thought the picture would bring more click. hmm
You got me on the last one, but I’m back! Reasons I picked were same as they stated, left-right eye path… and no bikini.
I hate ads with people’s faces or words from CEOs. Honestly, does anyone even care what those people have to say?
Hi – I too picked the one with the face incorrectly. On thinking about it more, though, I believe the right face in the right position, or perhaps another image that supports, enhances and makes the message more compelling would do better than the one with no pic.
Our feeling was the right banner for ‘conversions’ but we choose the left one for some reason -
Version A is definitely more inline with best practices.. yet it underperformed!
I would be interested to see the CTR for both versions as well. The CEO photo version could have possibly brought in more clicks but were less qualified leads.
Thanks Anne for featuring my test!
At first I was suprised by the results as well – but when you think about the target group – experienced Online Marketing managers – it becomes clearer: They just do not want to see faces and “CEO keynotes” but real helpful information.
Ill look for the CTRs but I do not know if we have them still stored.
If you have any questions regarding this test, feel free to contact me: dennis.krueger@quisma.com – or easily via this comments thread.
Hi Anne!
Another fun one, but there’s one test that still needs to be performed.
Conventional wisdom backed by decades of testing says to include a relevant photo, especially of a face, however our left-to-right reading leaves our eyes on the PHOTO as the final position, rather than the action device (the button).
I assure you that showing an average-looking (non-model) person does NOT suppress response over NO photo at all. On the contrary, it adds personality, reality, warmth, and connection. Always has… regardless of the product or service. There are very few mentally healthy people who “don’t want to see a face.” No bikini required.
I’d move the photo to the left… the button to the right… and retest. That’s very likely your new winner.
Success!
Drew Eric Whitman, D.R.S.
Direct Response Surgeon™
Author of,
“Cashvertising: How to Use More
Than 100 Secrets of Ad-Agency Psychology
to Make Big Money Selling Anything to Anyone”
http://www.Cashvertising.com
I agree with Drew on the layout.
The photo is engaging and good for getting attention to the ad – but the position within the visual flow could be tested.
Also maybe test less text when there’s a photo to reduce the otherwise increased friction.
And a doctor/scientist look instead of a bikini.
Hi, As ever, tests are always most illuminating.
What was the absolute number of differences in the sign-ups measurement metric. By what confidence interval is the result valid?
I voted B because I belive if the eye is catched by an image (anonymous for some people), it prevents people from reading the copy.
I vcoted for B because I thought it was straight to the point and there was too much going on in A with the photo. Plus, he wasn’t in a bikini…
B, it is obvious to me as I do eye tracking. The head attracts people’s eye gaze and detracts from the button. I wrote an interesting and popular post about faces and eye gaze here: http://usableworld.com.au/2009/03/16/you-look-where-they-look/
Very cool! I love detailed explanation here.
In our testing experience and the testing experience of our widely respected peers like Tim Ash, images typically do not increase conversion. Flow is critical as-is relevancy. In this case, the image clearly didn’t add value. Images that do add value tend to be when they appropriately bridge a gap in understanding.
Very surprising result, at least to me.
How do you like using the Vertster service? I am thinking about using it for my agency.
I worked with Omniture Test&Target for over a year in several projects and was not really satisfied. It is expensive and the support is slow. Vertster is different. Of course it has not the full functionality of T&T, but is has many advantages!
The price is good, it is really easy to implement, the support from the Vertster staff is fast and really great and there are some nice features T&T do not have (e.g. recalculate segments).
As an agency you are really flexible using Vertster. You pay only 0,0075 US-Dollar per test visitor. No minimum charge. You can skin the Vertster interface to your corporate design and have it available at your own domain.
If the customer has very special requirements and we need T&T we switch this project to T&T – and charge the customer MUCH more than in a Vertster project because of the license cost and the additional work we have.
About 80% of our projects are done with Vertster (and some with GWO, as we are a Google Conversion Professional).
You can find the new Vertster user group on LinkedIn at http://www.vertsterug.com
Doh! I thought the headshot would have won. Good lesson to learn I guess.
Nice test!
Social comments and analytics for this post…
This post was mentioned on Twitter by EssenteeWeb: Currently reading http://whichtestwon.com/banner-results/trackback. Conclusion? In this case, leave out the headshot!…
[...] was missing from the original experiment was the element of control. In truth, this experiment could have been testing any of the following; [...]
Got this one wrong.
Thought a picture would add credibility.
A game of MILLISECONDS…
When I saw the face I knew that would overpower the few milliseconds the average viewer gives and then they would go back to non-banner content. Especially since there was so much text to read. Had there been less text, then I would have voted for the face banner.