Wide click-through rate variance?

June 11, 2009 10:39 AM

How do I account for the wide disparity in click-throughs among our banner ads?

Hi. I manage a the Web site for a newspaper. The paper's online presence is about 10 years behind where it should be...for example, when I arrived, they weren't measuring any statistics for the banner ads on their site.

I am now measuring impressions and click-throughs for our banners, and I have noticed something odd: there is a huge disparity in click-throughs among the banners.

For example, one ad for a local car dealer - a static, stale, non-action-inducing ad that has been running on the site for a year probably - is getting ten times the clicks of a new, animated, call-to-action-style ad that I designed last week.

I know a significant number of click-throughs on ads are from robots, not real people. But in this case, wouldn't spider/robot "click-through" rates be basically level across all ads?

So: other than people in our audience just really REALLY liking to click on car dealer banners, what can explain the vast click-through difference?

(And yes, I realize most online publications have moved beyond click-throughs as a good measure of online ad effectiveness. Like I said, we're a little behind the curve here. We're getting there.)

Thanks!

posted by Bud Dickman to Computers & Internet (13 answers total)

It could be that the new animated ad is ugly or off-putting, or for a company that your readers have no interest in, or the call to action isn't as effective as you think it should be, or that something about the new ad triggers ad-blocking software while the old one doesn't, or that the more popular ad is in a more advantageous position in your layout or shows up on more heavily-trafficked pages, or is the target of a robot designed to artificially increase clickthroughs on that specific ad, or any of dozens of other reasons.

The only way to answer this is to gather more data: do A/B testing of animated vs. static ads for the same company on the same page, for example. Look more closely at the clickthrough logs to see which ones are coming from robots and which from real users.
posted by ook at 11:04 AM on June 11, 2009

People don't click thru ads on purpose at high volumes, there might be something about that ad that throws people off - thinking they are clicking on something else.
posted by bigmusic at 11:05 AM on June 11, 2009

I don't know about you, but I use flashblock for "attractive, animated, call to action" style ads.

I could really care less about static ones though ;)
posted by shownomercy at 11:12 AM on June 11, 2009 [5 favorites]

Are you using any third party tracking on the ads? What's the system you are using to track them?
posted by wcfields at 11:14 AM on June 11, 2009

Seconding that the location of the ad may be a huge factor.

Is the new hotness ad you made also for a car dealer? People still go to newspapers for used car listings. The auto ad could simply be relevant to more site visitors than the new hotness ad you made.

Is the wording on the old auto banner ambiguous? Could people be mistaking it for a link to the classifieds?
posted by Number Used Once at 11:16 AM on June 11, 2009

I know a significant number of click-throughs on ads are from robots, not real people. But in this case, wouldn't spider/robot "click-through" rates be basically level across all ads?

Depends what the robot is trying to do and how smart it is.

Is the static ad linked the same way as the other ads? Robots react to certain identifiers, some only click on ads on certain networks, or react to certain key words. Or maybe someone unleashed a robot to drive up the clicks of car dealerships specifically.

It could also not be robots. It could be that the URL for the local car dealer is really hard to remember/google and people just go to your site and click the ad because its faster.

Or possibly your site is the first Google result for the dealership, so people visiting your site only are interested in that ad.

Maybe a popular site has the dealership mis-linked to your site, and so again the visitors from there go straight to the dealership.

Use your site analytics to check the path through the site of people who click on that ad. See where they came from and how that differs from other ads.
posted by Ookseer at 11:18 AM on June 11, 2009

A static, stale, non-action-inducing ad that has been running on the site for a year probably - is getting ten times the clicks of a new, animated, call-to-action-style ad that I designed last week.

I like static ads. Motion ads distract me from the article and make me angry. I've clicked on many interesting static ads, especially at special-interest sites (bike forums, fishing sites, etc.) but the only attention I give "call to action" ads is trying to figure out how to turn it off.

Make more pleasing and subtle ads. Flashy ads catch attention, classy ads catch clicks.
posted by Slap*Happy at 11:28 AM on June 11, 2009 [3 favorites]

Eye tracking studies show that readers actively avoid animation on websites. Maybe people just aren't seeing your animated ads. In my experience (sounds like we have very similar jobs) static ads with a simple message get a much better response than animated ads.
posted by geekchic at 11:40 AM on June 11, 2009 [2 favorites]

There have been many usability studies showing banner blindness. People avoid flashing, spinning, loud colored ads, because they are obviously ads. It is such a large phenomenon that web designers have to be careful not to invoke the same reaction in navigation and other important information pieces on their site.
As others are pointing out, you need more information to make a valid comparison:

Where are the ads located on the page? top, bottom, right, left, center (some combination there of)

What is the context in which the ads appear? Are they falling along subject lines ie. the car ad is showing up with articles about cars/auto companies and personal ads selling cars? What is the topic of the new ad? The more relevant, the better!

Also, the robots may not have had enough time to glom onto the new ad.
posted by Librarygeek at 12:03 PM on June 11, 2009

It could also be a matter of branding. If that car dealership is pretty well known/highly regarded people might be more comfortable going to their site to check inventory or whatever versus another ad for a newer/less trusted dealership.

Does the ad appear in the same place all the time? It could be a set of individuals who use the ad as a sort of bookmark. They read the paper daily online then occasionally use the banner as a link to the car dealer. Like salesman at rival dealers or even salesman at the dealership.

Really, trying to track these things will give you a gigantic headache.
posted by asterisk at 12:05 PM on June 11, 2009

Perhaps someone at the dealership (say, the advertising manager) encourages employees (or friends/family) to go to the ad and click through on a regular basis to drive up their own numbers to justify their ad budget?
posted by amyms at 12:09 PM on June 11, 2009

Assuming you're doing something along the lines of A/B, where page hits are randomly assigned to the old ad or the new ad, there's probably a couple of reasons. Firstly, people may find the new design so distracting they navigate away from the page. Secondly, the old design may not stand out from the content, and expect the link to contain newspaper authored information rather than an advertiser website.

Fraud is also a possibility. It's not unheard of for businesses to pay people to click on competitor's ads to drive up costs.
posted by pwnguin at 12:25 PM on June 11, 2009

Response by poster: Thanks all! I knew there was probably not one tidy answer to this. I just wanted some of the possibilities, and you gave them!
posted by Bud Dickman at 12:53 PM on June 11, 2009