Creatives are one of the most important aspects to improve in programmatic ads. And for improving results, it is crucial to run PDCA for the creatives. Running PDCA cycle is something that everyone who has been involved in programmatic advertising can think of. But haven’t you run into a situation where running the PDCA cycle is not effective? If the results are not properly considered and linked to next actions, you may end up having “similar creatives in every touchpoint”. In this post, Ito Hideyuki from the Creative Unit of IREP will tell us about the proper way to run PDCA cycles for creatives, focusing on the “C (Check)” and “A (Act)”.
The proper way of running PDCA cycle
– What is the proper way of running PDCA cycles?
Ito:
It is important to have a firm “intention” in each phase of the PDCA. For example, in the steps “P (Plan)” and “D (Do)”, you should think “what kind of messages should we communicate to whom?”, “what should we clarify to move on to the next step?”, or “is the distribution design based on the understanding of media and leads?”
The steps after getting the results, “C” and “A”, are also very important. If the results are not properly considered and linked to next actions, you may end up with “single PDCA” or “having similar creatives over and over”, limiting the range of creative development.
Being in a situation where “proposals were great but PDCA is not working in the actual operation” may be due to the stagnant “C” and “A”.
– What causes stagnant “C” and “A” in the acquisition creatives’ PDCA?
Ito:
It is usually due to the superficial consideration of the results. For example, if a banner featuring a celebrity gets good results, superficial insight will end up to the action of “mass-producing banners featuring the celebrity”. If we look deeper into the reasons why it earned good results, we can think that “the sense of safety (i.e., authority) of the product by using a well-known celebrity led to the results”. In such a case, next actions can have range of options including “expansion of the maximum level of expression to communicate safety, such as being No. 1”. Keep asking “Why?” is the key for having different options in the next step.
(Figure 1: Avoid PDCA stagnation by digging deeper into the reasons for results)
– Where should we look to determine the good/poor of a creative?
Ito:
Basically, evaluation indicators should be defined based on the KPIs of the measures. If the creative is for acquisition purposes, it should be evaluated by CVR, CPA, and conversions.
However, if we only make superficial judgments based on these indicators, the scope of creative development will be narrowed, and operations will shrink.
One of our jobs at Creative Unit is to look at all the indicators of an ads, both good and poor creatives, and to consider “what psychological and attitudinal changes in the user caused this change in the indicator”, and to link it to the next development.
For example, if there is a banner with a strangely high CTR among the poor creatives, it may be due to “led people not interested in the product by misidentification,” or “people were interested and clicked the banner but left because of bad LP”. If it was the latter, we can take actions to win good results by finding out why they left the LP.
Key points for reading the Creative Report
– How should we read creative reports to understand the results correctly?
Ito:
I think there are four elements: understanding the medium, understanding the target, understanding the creative, and understanding external factors.
– Let us go one by one. What do you mean by the first, “understanding the medium”?
Ito:
“Understanding the medium” includes “understanding the displayed area” and “understanding the distribution algorithm”. You need to consider where the creative was displayed for each medium, what kind of people are there, and what kind of creative tends to get good responses, as well as the distribution logic of the medium.
For example, Facebook ads and Yahoo! display ads have completely different characteristics, so there are many cases of having good results with one, but not with the other. Without an understanding of the medium, it is impossible to consider whether this is because “the intended message was not conveyed with one due to its displayed page” or because of “differences in audience insights of that medium”.
– What do you mean by the second, “understanding the target”?
Ito:
The fact that the numbers shown in the report is “the result of a user’s action with some kind of intention” gets inevitably weakened when we check reports by just looking at the numbers.
Because they are the results of “human behavior”, the accuracy of reading and understanding the reports will improve if we read them having understanding “target insights”.
– What do you mean by the third, “understanding the creative”?
Ito:
“Understanding the creative” can be divided into “understanding the result impact of each component of the banner”, “understanding the appeal axis”, “understanding the path”, and “basic knowledge of the creative”.
The first, “understanding the result impact of each component of the banner”, means that we need to break down the creative into elements and compare and evaluate each of them. For example, the components of a still banner are visuals, copy, tone and manner, layout, and overall design. Considering the result impact, “key visual” that first catches the eye and the “main copy” that triggers action are important. By keeping this scale of impact in mind, it becomes easier to assess “which factors are responsible for the result” when comparing the good creative A with the poor creative B.
“Understanding the appeal axis” and “understanding the path” are related to the “P” in the PDCA cycle. We should read the report thinking of “what was our intention behind the development of this ad?” and “what was the path to acquisition?”
“Basic knowledge of the creative” means just what it says. Without a basic knowledge of design, it is difficult to know what is likely to attract attention. It is important for non-creative staff to have some knowledge when reading the report.
– What do you mean by the fourth, “understanding external factors”?
Ito:
Results of banners are not only by “internal factors” alone. For example, results are affected by seasonal and trend factors, as well as by the competitors’ ad placement situations. Ignoring external factors in verification can miss-lead the results.
Linking to next actions
– So, the result analysis based on these points leads to the correct hypothesis?
Ito:
The numbers in the report are just “facts” and does not show “the cause of the results”. It is crucial to consider that point and I think that is where we come in. For running effective PDCA cycles over a long term, it is important to derive the best next actions after each PDCA cycle.
– It seems that there are many points to consider and very complex to derive those actions. Wouldn’t that be personalized? How should the team be structured to continuously lead to correct considerations?
Ito:
As you say, experience is definitely the key. But at IREP, we ensure our quality by establishing a forum for sharing knowledge gained from past projects and creative staff participating in “training programs for operations staff”.
It is also not necessarily true that “personalization is absolutely bad”. We believe it is fine to have members who are “strong in a specific industry or product” while ensuring a certain level of quality. We may get new insights through discussions among such members.
With the recent development of AI, there are trends to go after efficiency, such as developing solutions that can predict creative results in advance and distribute high-result-expected banners. But I believe that it is still our humans’ area to “use imagination to come up with unpredictable ideas”.
Examples of considerations and linking to next actions
– Do you have any specific examples of good considerations leading to next actions?
Example 1
Various creatives were distributed in a project of advertiser in the human resources industry. The CTR of a banner that showed “fulfilling private life to communicate the positive turnaround after changing jobs” was extremely high and thus the distribution volume was also high. However, the CVR via that banner was low, and the CPA for the entire ad group was rising.
Ito:
If the CTR is high but the CVR is low, we can make hypotheses such as “customer target to lead to the LP was wrong” or “there were no information in the LP that the target wanted”.
In this case, we determined that the former factor was the strong factor for low CVR. We hypothesized that “although we were able to communicate the image of a positive turnaround as intended, the CVR rate was low because we led users who were not that urgent to change jobs”. So, we added a copy to show the timing and urgency of the job change and delivered it as the next action.
As a result, the hypothesis was correct, and it increased the CVR while maintaining the CTR.
The key here is not to cut the creative because it has a poor CVR and CPA, but rather to ask, “Why is the CTR so high but the CVR so low?” By structuring the next action based on it, it led to the discovery of the new effective creative pattern.
(Figure 2: Uncovering new effective creatives by reflecting questions about results in Action)
Example 2
In the “Point Refund Campaign”, the advertiser conducted display distribution covering videos and still images. The video creative did not lead to overall acquisition and distributions did not increase.
Ito:
Many would end up “not to deliver video ads considering its production cost and poor results”. But we first took an in-depth look at why the video had poor results.
The still image ad was directly communicating “the point refunds”, while the video ad was “explaining the campaign in a story, including the benefits of earning points”. Based on the premise that “the campaign require a sudden increase in motivation and a change in attitude, like an impulse purchase”, we arrived at the hypothesis that “the video was packed with information that it gave users time to think whether or not to participate in the campaign before clicking, not generating impulse purchases”.
So, we deleted the information in the video to about the same level as that of still image ad, making the video “only showing a portion of the still image in motion, drawing more attention to the refunds”. This significant improved its acquisition, exceeding that of the still image ad.
As in Case 1 above, if the advertiser judged the creative based on numbers alone, this effective creative would not have been born.
(Figure 3: Considering the next action while digging deeper into effective appeal factors and user insights)
– Finally, give us comments to those in charge of promotions in the client companies who are concerned that their creative PDCA cycle may not be working properly.
Ito:
Creative PDCA requires understanding and consideration cultivated through daily work. It takes some time to develop such skills and expertise.
Of course, it is possible to achieve that in-house, but I think it would be better to hire an external company like IREP to structure the operational system.
If you feel that the PDCA cycle is not working, that the number of effective creatives is decreasing, or that CPA is rising over the long term, feel free to contact us.
Hideyuki Ito joined iREP in 2018. He has been responsible for advertising planning and creative work in various industries, including both BtoC and BtoB, such as human resources, education, and apps. Currently, he works as a manager in the department responsible for creative planning and production in the acquisition area. He received the IREP AWARD in the first half of the 2021 fiscal year internal commendation.
Want to know more about Digital Marketing?
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
