Leaderboard vs. 468 Banner

Case Study of Impact of Ad Size Upon Earnings

[Hobbs]

One of channels has been running Leaderboard for a long time, it has been generating around 20% of my site earnings, beautiful targeting, great CTR, 10 times my site average eCPM, perfect blending with the leaderboard below logo and navigation but above content. It was a beauty.

I wanted to test for many things:
1) See the CTR impact on smaller ad size
2) Give image ads a second chance
3) Lower the ad footprint on the page, and have them ads compete for a smaller area, hopefully resulting in higher EPC.

I did a 2 weeks test switching from Leaderboard to 468x60 banner size which resulted in low CTR, low eCPM, low EPC and half earnings!

2 things actually changed for this channel:
1) Its blending was affected with white space left and right of the banner
2) It became infested with CPM image banners, which are beautiful to my eyes, but obviously not for my visitors.

This is becoming a rather expensive test, so I am giving it only another 1 week of testing, where I will restrict the Ad unit to Text ads only to kill off them image ads and isolate the real cause of the drop.

3 things to conclude at this stage of test:
1) If it aint broke don't fix it
2) Blending quality can be worth half your earnings
3) What you may think looks nice or logical could have nothing to do with your visitors taste, so do not stop testing.

Will update this post in one week with the impact of restricting Ads to text only and doing away with CPM image ads.

Note: I am not advocating Leaderboards, if 468's blend better in your page then 468 it should be.

[jetteroheller]

For this, I want 600x80 or 600x90 format

My content area is always 600 width.
So I have no opportunity for the leader board.

But the 468x60 is to small to work good.

[Hobbs]

3 days later:

After restricting 468 banners to text only, thus eliminating image CPM ads I am seeing the following:

- Much much higher EPC and eCPM
- Still low CTR
- Overall very good earnings comparable but less than when the 468 banner was a leaderboard.

Conclusion:
a) Yes less ads per ad unit will increase eCPM and EPC
b) CPM image ads are neither profitable nor popular for my site

In 4 days I will increase the 468 to leaderboard thus returning to original format before testing, this will give the 100% blending effect and compare to measure for only the blending impact.

I am posting this because I promised to do it, but I do not plan to update this thread further after the final test due to lack of interest.
(Sorry for bringing it to the top again)

[21_blue]

Hey, Hobbs, I'm interested.

I didn't see your original post 2 (two!) days ago - it must have got lost in the avalanche of messages at the time.

By the way, I don't think images are necessarily CPM. And I suspect if you gave the Google bot long enough it would work out that image ads don't provide you with a good return. I say this because we saw a lot of image ads initially - in pre CPM days - but they gradually disappeared

[Jean]

due to lack of interest

This is interesting, I just didn't have any comment to add. Thanks for doing it and letting us know your findings.

[Hobbs]

Hey 21_blue, Jean,
I will update in a week, just thought no one was interested in the outcome, and it was an expensive test for me, 10% of earnings is a hefty amount, if you add it the normal earnings wave you get seasick!

I am quite sure those images I saw were CPM, and in a previous extended period test I did 6 months ago, Google did seem to favor them over better paying CPC ads, probably advertiser retention strategy? Don't know.

Anyway, good to see some 1 and 2 dollar clicks for a change.

[annej]

I am very interested. I was just thinking of changing some leader banners to 468s because they would 'look nicer'. Your results so far are very helpful and I will be interested in hearning more as you test things.

[bghtn]

Much interest here. Thank you for posting.

[gendude]

I'm looking at 468 banners as well, as they would give me a cleaner looking site. Thanks for posting your results.

[damianf]

I too am quite interested in this... currently using mainly 468 banners but will now try leaderboards for a while to see if I can increase revenue.

Cheers,
Damian

[europeforvisitors]

My CTR, eCPM, and earnings went up quite a bit when I switched from 468 x 60 banners to leaderboards back in 2004 or thereabouts. I've used both ad formats in the same location (right below the top navigation bar), so the ad unit has been the only variable.

I think that--on my site, at least--the leaderboard has performed better because:

1) It's more visible, and...

2) It gives readers more choices, so they have more opportunities to click on ads that interest them.

BTW, 468 x 60 banners are fading from the scene (my rep firm tells me that some big agencies aren't even designing for the 468 x 60 size anymore).

[Hobbs]

I have concluded this experiment with results that are very specific to my own site.

a) Blending blending and more blending, it makes or brakes the bottom-line earnings.

b) While 468 (or smaller ads in general) ads can increase the EPC due to more competition on limited space, it can lower your CTR and overall earnings.

c) Targeting higher eCPM is better handled by working on better quality content to get better targeted ads than fiddling around with ads, sizes and locations.
Time is better spent on growing and improving your site.

d) There are still many ways to increase earnings, but focusing solely on eCPM is not one of them.

[johnblack]

Hi Hobbs,

Many thanks for sharing your 'experiment' with the community.

Don't ever confuse a lack of response with a lack of interest - I've just caught up with this thread and it's all good info. Keep up the good work.

Thanks

JB

[Hobbs]

Thank you johnblack, I did not want this point to become the focus of the discussion, back then when I posted that I will not update the thread, I was only pissed at myself for lowering my income with experimenting, and everyone is right, no response does not mean lack of intrest..Anyway I will always experiment and post when I have time, call it curiosity with a tiny bit of natural greed ;-)

[greatstart]

I also switched to leaderboard, and noticed an improvment in my CTR. I guess 468x60 is now a thing of the past, as many users are running larger monitors with 1024x768 and higher resolutions.

[jomaxx]

FWIW, I did a split test last year and decided to stay with banners in preference to the leaderboard. The CTR was no higher and the EPC was less, although perhaps not to a statistically significant degree. I put it down to the effect of running more separate ads on pages which unfortunately don't have a deep pool of relevant advertisers.

[21_blue]

b) While 468 (or smaller ads in general) ads can increase the EPC due to more competition on limited space, it can lower your CTR and overall earnings.

I've come to a similar conclusion recently experimenting with a 120x600 tower and 120x240 vertical banner.

c) Targeting higher eCPM is better handled by working on better quality content to get better targeted ads than fiddling around with ads, sizes and locations. Time is better spent on growing and improving your site.

This is a good point. If one "fiddles", the best outcome one can hope for is to increase earnings up to a ceiling that represents the maximum potential of the site. If one adds content, however, the ceiling is raised.

There is a role for fiddling, but it shouldn't take the lion's share of time nor be the main strategy for income growth.