Measuring the Effectiveness of Television Commercials with Irep

Irep Inc.

Hi. I’m Tomita from Irep’s Programmatic Television Commercial Promotion Division. I’ve been asked by the editors at DIGIFUL to write up a second column for this series. Let’s begin by summarizing the previous issue.

With the reorganization of Irep by Hakuhodo DY Group, we’re now able to handle television commercials. We began discussions internally as we built up our track record. The last article also discussed how we are able to provide television commercial services that enable decisions to be made based on data-driven planning/measuring.

We called the service “A Scientific Approach to Television Commercials.” One of the defining features of the service is how it combines Irep’s data analysis power with the history of mass-media advertising that Hakuhodo DY Group possesses to create a new PDCA cycle for television commercials.

Recently, we’ve developed a programmatic digital mass advertisement service from the programmatic television commercial service that provides an integrated PDCA cycle with internet-based video ads.

This service consists of three initiatives. The first is television commercials and OTT ads. The second is television commercials and internet video ads. The third is television commercials and the lower region of internet ads.

This article will explore the methods Irep uses to measure the effectiveness of television commercials.

Evaluating Television Commercials Based on Actions and Performance


Conventional television commercials ordinarily used measurements like actual viewership rate (GRP/PRP), delivery/stream-based measurements like delivery rates, and measurements based on consciousness such as advertisement awareness, recall, and usage consideration from surveys.

On the other hand, for ads on the internet, evaluations were based on actions and performance, such as clicks and on-site conversions. We at Irep use similar methods to conduct the PDCA cycle. We’ve established a method to apply the PDCA cycle based on actions and results for television commercials. We currently use two models to conduct the PDCA cycle.

(Figure 1: The results of television commercials are visualized based on actions and performance)

– Measurement Method 1: Short-term Spike Detection Model

This measures how much traffic occurred on your website for the few minutes immediately following the airing of your commercial. For example, it measures which of the following two patterns had a higher rating.

  1. A slot with a 10% viewership rate that resulted in 100 visitors to the website after airing.
  2. A slot with a 5% viewership rate that resulted in 300 visitors to the website after airing.

This method can be used to measure the contribution of each slot and time of day to the traffic of a company’s website, and thus serves as an indicator for advertisers to determine which air slot and time of day has higher ratings. However, even if the results for a particular day of the week and time are good, there may be differences among broadcasters, and even if the broadcasters are the same, there may be differences due to differences in the genre of the programs. If several different types of advertising materials are aired, there could be differences for each material.

Based on the results of this analysis, it may be possible to increase the amount of traffic to the company’s website by proceeding with zoning and revision of estimates (negotiating which slots to air).

However, there are caveats to this approach. Figure 2 shows that more than 80% of the programs aired one or two commercials per program during the airing period. If we were to analyze the airing slots, we would have to pay attention to the ratio of the number of commercial airings per program.

In other words, it’s important to keep in mind that the analysis will be conducted with a small number of trials, and that the days of the week and times with good performance do not generally mean that the slots have high ratings.

(Figure 2: Points of caution for the short-term spike model)

Therefore, we adopted a method to increase the reliability of our analysis. We analyzed which factors are important in the way commercials are evaluated for airing, and made the evaluations not coincidental, but highly reliable based on evidence, thereby enhancing the reproducibility of subsequent planning and buying.

We determined the important factors for each project, such as the day of the week, the time of day, or the program genre, to provide highly reliable analysis results and a PDCA cycle.

– Measurement Method 2: Time-series Statistical Model

The time-series statistical model is a model that statistically calculates how much the KPI was lifted during the airing period. In addition to calculating site traffic, this method enables the calculation of conversions within a particular website. It’s also possible to calculate how much the television commercials contributed to conversions, as well as the CPA of TV commercials.

This involves the use of estimating counterfactuals through causal inference. This method statistically calculates the counterfactual of what the KPI values would have been if the advertisement had not been aired or distributed.

The difference between the actual value and the counterfactual value is judged to be the KPI value resulting from the advertisement airing.

(Figure 3: Grasping the counterfactual)

This method requires the intentional creation of regions where the ad is not to be aired or streamed. Using the KPI trends in those areas as data, the counterfactuals for the areas in which the ad is aired or streamed are calculated.

You may have noticed by this point that this method is not limited to television commercials. It can be applied to all advertising channels, whether online or offline. For example, it’s theoretically possible to visualize the effectiveness of internet video ads and even newspaper ads. 

We at Irep have used this method to compare the results of television commercials and TVer ads with those of standalone television commercials. As a result of several demonstrative experiments, we’ve also seen cases in which TVer ads have increased the effectiveness of website traffic.

(Figure 4: Case study of improved traffic to website through combined use of television commercials and TVer ads)

By measuring and modeling the ROAS for both television commercials and YouTube ads using this method, it’s also possible to calculate the appropriate budget to be allocated for television commercials and YouTube ads. Many of you may be asking what the right answer is when it comes to allocating budgets between television and digital ads. Using this method may lead you to the right answer.

(Figure 5: Case study of allocating your budget between television commercials and YouTube)

Monitoring on the “x2 supported by TV AaaS” Dashboard


Both of the two models we introduced are included in our proprietary development tool, “x2 supported by TV AaaS” (hereinafter, “x2”). The projects we’ve supported can be monitored on the tool’s dashboard. The details of x2 will be discussed in the next column in the series.

We don’t think that the two models we discussed today, the short-term spike detection model and the time-series statistical model, will meet all of the demands of the advertisers.

However, as we mentioned in the previous article, we at Irep have access to the tools and vast amount of data owned by Hakuhodo DY Group.

To respond to the demands of the advertisers, we can choose from various tools and methods. We believe that our analytical skills, cultivated through the PDCA cycle of advertising on the internet, will enable us to respond to a wide range of needs.

Summary


This article explored our approach and methods for measuring the effectiveness of television commercials, which support Irep’s programmatic digital mass ads. We hope that you now understand Irep’s high capabilities when it comes to television commercials.


Want to know more about Television Commercial?
Contact us to discuss how we can work together.

Irep Inc. is an award-winning global digital marketing agency based in the San Francisco Bay Area. Our headquarters are in Tokyo and our network spans more than 20 countries. In Japan, we are ranked No. 1 for performance-based marketing. We also offer highly specialized market entry, as well as integrated marketing and localization services. Since 1997, our data-driven solutions have effectively led our diverse international clientele to continuous success in Japan, Asia, and beyond.


Irep Inc.
LinkedIn: https://www.linkedin.com/company/irepinc
Email: info@irep.inc
Address: 900 Concar Dr. Suite 400, San Mateo, California 9440