Setting KPIs for Web Advertising Initiatives ~ Dentsu Inc. Private Marketplace

"How valid is it to use CPC (cost per click) as a KPI?"

This was the fundamental question when planning Dentsu Inc. PMP project, as discussed previously. Even now, many advertisers adopt CPC as a KPI, especially for web advertising aimed at awareness building and attitude change. But should CPC inherently be a crucial metric?

If CTR is 0.1%, are 999 impressions meaningless?

Compared to Europe and the US, Japan tends to place greater importance on the CPC metric. Web advertising assumes "site visit = awareness" and seeks efficiency in driving website traffic. Naturally, for direct campaigns where online outcomes directly translate to revenue, clicks and conversions hold extremely high value. However, even for campaigns aimed at awareness or attitude change—not immediate conversion—the focus is often solely on actions after the click.

Consequently, major issues frequently covered daily by outlets like eMarketer and AdExchanger in the West—such as where ads appear (from a Brand Safety and Ad Verification perspective), whether ads are properly displayed (Viewability), and whether ads are genuinely seen by humans rather than fraudulent bots (Ad Fraud)—are often downplayed in Japan. The quality of the content being provided is often overlooked, with the primary focus instead on how high the CTR is for the ad slots within that content (i.e., how low the CPC can be kept). Unfortunately, this seems to be the current state of web advertising in Japan. One thing that can be said with certainty is that publishers providing high-quality content will eventually become exhausted if this situation persists.

If this were based on sound reasoning—such as research showing that "ad placement quality has no correlation with brand awareness, and impression-based awareness is inherently weak, so web banner ads should primarily pursue click efficiency"—then it would be perfectly acceptable. However, the reality is that, except for a few pioneering clients, this has hardly been verified at all. As a result, for example, if the CTR is 0.1%, the 999 impressions displayed before getting one click are completely disregarded. Can we call it healthy growth when market formation progresses by overlooking factors that might potentially hold significant value?

Awareness & Attitude Change & Reach to Target Users as KPIs

Methods for verifying awareness and attitude change are also evolving daily. Solutions that can somewhat meet advertiser demand are emerging, ranging from simple surveys to SSP (single source panel) studies spanning devices like TV, PCs, and smartphones, and even real-time analysis of awareness and attitude change. Particularly when the objective is not harvesting but rather awareness acquisition, attitude change, or target reach, as some domestic advertisers are discussing and trialing, it may be necessary to set reach quality as a primary KPI. Based on this data, a PDCA cycle should be advanced regarding media and advertising menus.

Various reasons likely contribute to the current CPC bias. A significant problem was the lack of the above perspective on the part of advertising agencies. It's difficult to say they were making proposals that properly considered the quality of reach, something not measurable by CPC. Moreover, a fundamental premise was the widespread, somewhat vague perception that banner ad impressions lacked any real awareness effect. This perception was very accurate. As Google announced late last year, approximately 60% of banner ads were not even seen by users ( 56.1% of display ads weren't even shown ). Ads that aren't seen obviously cannot have any awareness value.

*Viewability Survey (December 6, 2014)
https://think.storage.googleapis.com/docs/5-factors-of-viewability_infographics.pdf

 

Recently, technological solutions have emerged for various RTB issues, including viewability. PMP is one such solution. First, establish the premise that impressions are reaching users to enable impression-based awareness and attitude change surveys, then rigorously verify performance across each media outlet. Furthermore, we conduct performance analysis per site not only during conversion-based verification (as described in Part 1) but also based on awareness and attitude change. We provide this feedback not only to advertisers but also to publishers. This enables publishers to implement their own advertising PDCA cycles, exploring how to enhance awareness effects. This approach transcends existing schemes and formats, leading to advertising development and ultimately benefiting advertisers. If such a cycle becomes established, wouldn't it create a market different from the current one?

What surprises me most in conversations with amnet (UK), Dentsu Inc. Aegis Network's trading desk, and major international media players is that CPM (cost per thousand impressions) figures differ by an order of magnitude compared to Japan. This reflects the existence of value metrics beyond CPC, such as awareness/attitude change and target audience reach, which publishers themselves recognize and actively seek to maximize. While I don't necessarily believe rates should immediately rise to UK or US levels, the fact that many advertisers are purchasing high-CPM ads instead of low-CPM ones is worth considering.

As stated in the previous article, the objectives of the Dentsu Inc. Private Marketplace are as follows.

We hope this project will serve as an opportunity to foster a healthier mutual understanding between advertisers and publishers.

Next time, I plan to write specifically about PMP. In the somewhat disorganized programmatic space, I want to clarify where the area Dentsu Inc. calls PMP fits in, and what benefits it can offer advertisers and publishers that differ from RTB.