Commentary
Seeing Ad First Increases Action Odds Sevenfold
- by Jack Loechner , Staff Writer @mp_research, August 4, 2011
Users are three to four times more likely to act on an ad if it is the first or second one they see during their session, says the report. Ad effectiveness plummets as the user progresses through their online viewing.
And, repetition works to an extent. Ads shown five times or more to a user were 12-14 times more effective than ads shown less than five times.
Three criteria relating to the serving of online banner ads were examined:
- Placement relative to the page fold
- Moment of delivery within a user's session
- Frequency of exposure.
On the premise that advertising is all about capturing the attention of one's audience, the study tests the hypothesis that not all impressions are created equal, by evaluating the effect of three ad placement variables (page positioning, view order and frequency) on campaign performance (quantified in terms of click and action rates).
Three ad delivery parameters were examined to evaluate their influence on the number of resulting clicks and actions:
- Page positioning (above/below-the-fold): advertising delivered above the website fold is visible as soon as the page is loaded, i.e. scrolling is not required.
- View order: ads are assigned a ranking according to their order of delivery within a user's session, e.g. the very first ad to be delivered is considered to be in "first impression" position. View order relates to where in the "tail" inventory is positioned, e.g. early impressions are considered to be "short tail", while impressions delivered late in a user's site browsing session would be considered "long tail".
- Frequency: this refers to the number of times an ad is shown to a user over a fixed period of time.
The analysis revealed that when displayed above-the-fold, ads are almost 7 times more effective at generating a click than ads delivered below the fold. The ratio is virtually identical when considering whether an action was completed. These results support the findings of numerous studies based on eye tracking data, according to which users spend the vast majority of their time looking at information positioned within a page's initially viewable area.
Ads Delivered Above-Fold Get Better Traction | ||
Position | Click index | Action index |
Above fold | 17.9 | 9.5 |
Below fold | 2.6 | 1.4 |
Source: Casale Media, July 2011 (Click index: Number of clicks÷number of impressions x 1000; Action index: Number of actions÷number of impressions x 1000)
| ||
Absolute Impressions and Actions | |||
Page position | Impressions | Clicks | Actions |
Above fold | 1,728,347,297 | 3,094,349 | 164,169 |
Below fold | 54,087,739 | 14,264 | 742 |
Unknown | 120,402,698 | 17,705 | 772 |
Total | 1,902,837,734 |
|
|
Source: Casale Media, July 2011 | |||
The impressions sampled for this study are segmented into eight different tiers ranging from 1st-2nd position to 255th and beyond. The data corresponding to each tier shows that both clickthrough and action rates decrease rapidly as users progress through their online journeys: ads ranking in 3rd to 6th position see their click and action rates plummet compared with ads showing as 1st/2nd impressions (almost 3-fold and more than 4-fold respectively).
Ads Shown Early-On Perform Better | ||
Position | Click index | Action index |
1st - 2nd | 32 | 20.3 |
3rd - 6th | 12 | 4.7 |
7th - 14th | 9.5 | 2.6 |
Source: Casale Media, July 2011 | ||
This data suggests quite clearly that as users are exposed to more and more ads within their browsing session, those ads become less and less effective at capturing the user's attention, to the point of oblivion (a.k.a. banner blindness). The earlier an ad is shown to a user, the more likely it is to be noticed and therefore, effective.
This echoes a common practice in print advertising, where "early" pages, situated near the main editorial content, carry a higher advertising rate. Interestingly, the data shows that there is still value to extract even from very low ranking impressions. Although these will makeup some proportion of any inventory, they should be excluded from cases where an advertiser buys and values campaigns based on exposure alone.
Importance of Viewing Order | |||
View order | Impressions | Clicks | Actions |
1st - 2nd | 655313282 | 2095995 | 133175 |
3rd - 6th | 398362243 | 479545 | 18841 |
7th - 14th | 286068413 | 271858 | 7468 |
15th - 30th | 202354583 | 162232 | 3305 |
31st - 62nd | 123788168 | 67707 | 1124 |
63rd - 126th | 47171268 | 14327 | 216 |
127th - 254th | 10865305 | 2231 | 36 |
255th + | 4424035 | 454 4 |
|
Total | 1,728,347,297 |
|
|
Source: Casale Media, July 2011 | |||
It has been said that it takes nine times for a marketing message to move a prospect from a state of total apathy to purchasing readiness. The results of this study certainly lean in the same direction, as both click and action rates dramatically increase, almost 12- and 14- fold respectively, for ads that have been shown 5 times and over.
The Effect Of Repetition | ||
Repitition | Click index | Action index |
≥ 5 times | 174.4 | 92.5 |
≤ 4 times | 14.7 | 6.4 |
Source: Casale Media, July 2011 | ||
As in offline advertising, several exposures are required to achieve some degree of familiarity and to register with users. However, it is also a well known fact that over-frequencied ads can be counterproductive. To mitigate the effect, "frequency capping" mechanisms may be implemented to limit the number of times an ad is delivered to the same user or "frequency optimization" to determine the optimal cap for a specific campaign.
The Effect of Exposure | |||
Ad exposures | Impressions | Clicks | Actions |
≤ 4 times | 120402698 | 17705 | 772 |
≥ 5 times | 1782435036 | 3108613 | 164911 |
Total | 1,902,837,734 |
|
|
Source: Casale Media, July 2011 | |||
The report concludes with some final thoughts:
- Relying blindly on a single indicator, such as a reach table, or technique like hyper-targeting, could prove a mistake
- The basic requirement of getting an advertising message in front of eyeballs should not be taken for granted. Without delivery above-the-fold, early session placement and frequency optimization, campaign performance may suffer
- Diversification might mitigate the risk, but most of all, vigilance should be applied
For additional information, please visit here.
Jack Loechner is the Editor of The Center for Media Research's daily Research Brief. You can reach him at jack@mediapost.com.
Does this study allow for viewers clicking on top-loaded advertising by mistake? It happens fairly frequently, when a top-loaded ad unexpectedly (and automatically) expands downward into the website's content real estate. After intending to click on a link to a news story about the global financial crisis, a viewer can instead find him/herself redirected to a website marketing a new SUV.
I am so disappointed in both the center for media research and in casale. Research which uses the click as the relevant success metric for online ads is not only flawed in its basic premise; its misleading and downright damaging to the discipline of digital marketing.
There is no doubt that ads above the fold are better (more effective) than those below...but it is NOT because they are clicked on more often!
This piece is not illuminating...it is perpetuating a damaging myth...
I agree with John here, although this study still tells a compelling story when looking at "actions". The variables such as position, timing, and frequency all had a similar impact on "actions" as they did on clicks.
What's needed here is some clarity as to what this study considered an action. Is it a purchase, a download, a video play? I think that would paint a better picture and allow us to move our attention away from the oh-so-antiquated metric of the click...
Based on the tables above, I would say that this study didn't take rich media ad units into account, but rather standard placements like 728x90 and 300x250 banners. It is well-documented that ads placed above the fold have a better click-thru/action rate for the simple reason that websites are designed to have their most pertinent information "above the fold" - when's the last time you scrolled to the bottom of a web page?
What I find more interesting is that a consumer needs to see an ad 9 times in order to move from apathy to purchase. Frequency caps were created expressly for this purpose, as over served ads can lose their effectiveness, but they can also negatively impact a consumer's impression of the website, possibly hurting return traffic.
As far as measuring clicks - this is the metric 95% of businesses and agencies continue to use as a sign of success. We cannot expect research to show us something different until the digital media community determines what the next effective metric is and then makes it easy to track. Tracking conversions and/or specific actions (i.e. filling out an application, taking a survey) is often difficult as it requires the tracking pixel to cover at least two sites, possibly three. Having worked in digital advertising for a while now, I can assure you that invariably there is a weak link in that chain.
I know one particular company that will be thrilled to have this data as it will help them to further make the case for premium ad positioning (i.e. above the fold) and therefore allow them to charge a high CPM. Publishers will like it for that matter as well.
John, thanks for commenting. To build on your point, our intent was to illustrate with hard data that not all impressions are equally valuable -- by no means are we trying to propagate the myth that the click is king online! We actually encourage our clients to measure their campaigns using a variety of different metrics because we appreciate that the click only tells a small part of the story. The click is only one of the data points evaluated in the study. We also examined the effect of placement on conversion metrics, which revealed findings consistent with those found for clicks.
There are impressions placed in premium positions that deliver real impact and there are impressions delivered in positions that are rarely if ever seen (like at the bottom of a page, in an invisible in-frame, or in a gallery that the average user clicks through in a matter of seconds). Our intent was to use data to shed light on the concept of impression value, which, for many, is common sense as you pointed out. For publishers, making wise choices about the position of placements can have a positive impact on the overall value of their supply and for marketers, understanding the value of impressions up front can inform buying decisions that lead to enhanced performance.
Hi Jennifer and good point.
In our experience, the ad unit most susceptible to the top-loaded accidental click is the 728x90 (leaderboard), which is generally positioned across the top of the page. Our study normalized data across all ad units, so any inflation due to accidental clicks was screened out.
We further normalized our data by removing what we call “second clicks,” or accidental clicks that generally happen immediately after an intended click. They tend to inflate counts and otherwise dilute results. The study also looked at conversion data, which showed findings consistent with those found for clicks.
0.17% click through rate for above the fold.
0.0094% action rate.
Doesn't this say a big failure?
I would love to see this compared to radio and TV ads, the biggest competitors to online ads, or should I say that the other way around?
Although this is interesting, and some of you bring up good points. I would use this directionally, if we knew what the products were, sites, were there blogs involved which would make above the fold useless compared to an ecomm site. Did the clicks above the fold on the 2nd or 3rd page make any difference? Be nice to know the click stream metrics was on those. The other thing is did the ad appear in a new window when clicked or did it take them to the website? what is interesting is that the actions per click are very similar 5.31% vs 5.20%. I would ask which one generated a better ROAS.