"Although we are running web ads, we are not achieving the expected results," and "While CPA is rising, we do not know which metrics to improve." Many marketers face challenges like these. The old operational methods of setting up a campaign and considering it done not only fail to produce results but also waste budget.

Modern web advertising has transitioned from relying on broad demographics to a strategy of continuously optimizing based on data. The key to success lies in agility, which allows for rapid responses to change.

This article explains six new common sense practices that are essential for future web advertising operations. From the basic PDCA cycle to automated bidding utilizing AI, and new technologies to survive in the post-cookie era, we will comprehensively introduce a framework to maximize results. By the end, you should have a clear guideline for analyzing and improving your company's ad performance.

1. "Setting it up and considering it done" is the cause of failure: The PDCA cycle of web advertising operations that continuously delivers results

Effective web advertising operations are not simply set-and-forget. To continue achieving results, it is essential to run the PDCA cycle, a continuous improvement process. PDCA consists of four phases: Plan, Do, Check, and Action.

The reason many campaigns fail is that the Check and Action phases are often overlooked. Particularly in the Planning stage, if there are no clear criteria for evaluating success or failure based on specific metrics, the Check phase depends on the subjective opinions of those involved, and the Action phase ends up being arbitrary measures.

To ensure continued success, it is important to systematize PDCA not as a personal experience guideline, but as a "model for operations" that can be shared across the team. This preserves consistency in measures and accumulates long-term insights within the organization. A disciplined verification method, like the A/B testing introduced next, will become a powerful engine in the Check and Action phases.

Professional Tip: Routine Evaluation in the Check Phase To sustain the PDCA cycle, let's establish a routine for the evaluation phase on a weekly and monthly basis.

• Weekly: Check short-term metrics like bid prices, budget consumption pace, and CTR per creative, and make adjustments.

• Monthly: Review mid to long-term strategies regarding budget allocation per medium and the effectiveness of targeting sets, and link them to the next Plan.

2. Moving from "Who" to "When and Where": Targeting strategies to win in the post-cookie era

The targeting of web advertising has evolved from merely specifying user attributes (demographics) to capturing user intent and context (situations).

First, the target audience can be broadly divided into "active segments" and "latent segments." Active segments consist of users who have specific needs and are actively searching for information. In contrast, latent segments refer to user groups that may potentially become customers in the future, even if they do not have conscious needs yet. Depending on the type of advertisement, the preferred target segments can vary.

In particular, retargeting, which involves re-approaching users who have visited a site once, was a very powerful method utilizing third-party cookies. However, the trend towards enhanced privacy protection is shaking the foundations of this strategy. According to research, consumers find ads to be "annoying (63%)" and "persistent (47%)," and this user sentiment is driving changes throughout the industry.

As a result, traditional retargeting is becoming obsolete due to privacy regulations. The key successor to this user-tracking strategy is contextual advertising. This method shifts focus from tracking the user to analyzing the content and context (context) of the web pages users are viewing and delivering ads that are highly relevant to that content.

One study found that contextually relevant ads are 43% more memorable compared to less relevant ads, making this a very effective strategy for enhancing brand awareness and purchase intent.

3. The strongest ads are always unfinished: The technology of A/B testing to improve effectiveness at low risk

A/B testing is a method where multiple patterns (A and B) are created by changing one element of an ad copy or image to compare which produces better results. It is an essential technology for making data-driven improvements in ad operations.

A/B testing has several key advantages.

• Affordable improvements: Since effects can be verified through minor changes like altering text or color schemes, there is no need for large amounts of work or additional budgets.

• Clear results: Since outcomes can be compared using clear metrics like click-through rate (CTR) or conversion rate (CVR), it becomes easier to identify what led to improvements.

• Low risk: Changes are made gradually, so even if a tested pattern fails, it minimizes the negative impact on the overall campaign.

The most important rule for A/B testing is to make changes one at a time. If multiple elements are changed simultaneously, it becomes impossible to determine which change affected the results, preventing the drawing of correct conclusions.

[Specific Example of A/B Testing: Display Ads for a Nail Salon] Create test patterns focused on only one element as follows.

• Layout: A pattern that emphasizes an enlarged close-up of fingernails vs. a pattern that shows more textual information.

• Image: Photos showing the ambiance of the salon vs. close-up shots of finished nail designs.

• Text: Test by changing the appeal points. Whether to emphasize price benefits like "20% OFF popular designs for first-time customers" or quality and experiential value like "Professional techniques with a custom nail experience that transforms your fingers" affects responses.

• Color Scheme: A luxurious design focusing on black vs. a glamorous design focusing on pink.

4. The bidding showdown between humans and AI is over: How AI optimizes ad operations

Modern web advertising operations have completely shifted from manual fine-tuning to automated optimization by AI. In particular, the use of AI in setting ad bidding prices has become the norm.

With manual bidding, operators set bid prices for each keyword, whereas in automated bidding, AI analyzes numerous signals (devices, regions, time zones, user behavior history, etc.) in real-time at a speed and precision that is physically impossible for humans, optimizing bid prices for each auction.

In Google Ads, Smart Bidding, which is particularly focused on maximizing conversions, is mainstream. Depending on the ad's objective, an optimal strategy can be chosen.

When implementing automated bidding, it is essential to understand the Learning Period. When applying a new bidding strategy or making significant changes to settings, AI requires a period to collect and learn data (typically 1-2 weeks) to find the optimal delivery pattern. During this time, AI analyzes KPIs such as CTR and CVR, which are explained in the next section. Performance is likely to be unstable, so frequent changes to settings should be avoided, and it's important to observe the AI's learning process.

5. Numbers tell a story: Identifying bottlenecks in ad effectiveness measurement from the relationship between CTR and CVR

To improve ad performance, it is insufficient to simply look at a single metric. By interpreting the relationships between multiple metrics, you can identify the actual bottleneck (issues) that the campaign faces. Manual analysis is essential for diagnosing bottlenecks, but executing large-scale improvements based on those diagnostic results falls under the expertise of AI's automated bidding mentioned earlier.

First, let's understand the four most important KPIs in web advertising operations.

• CTR (Click-Through Rate): (Clicks ÷ Impressions) × 100 - A metric measuring how much the ad creative attracts user interest.

• CVR (Conversion Rate): (Conversions ÷ Clicks) × 100 - A metric measuring the persuasiveness of the landing page (LP) or offer after ad clicks.

• CPA (Cost per Acquisition): Ad Cost ÷ Conversions - The cost incurred to acquire one conversion.

• ROAS (Return on Ad Spend): (Sales ÷ Ad Cost) × 100 - A metric showing how much revenue is generated from the ad spend.

By analyzing the relationships between these metrics, we can make diagnoses such as the following.

When "CTR is high but CVR is low," the problem lies not with the ad creative but with the landing page (LP) after the click. Users may have been attracted by the ad, but the LP's content was likely not as expected or not compelling enough.

[Common Issues and Hypothesized Causes]

• Issue: Low CTR

  • Hypothesized Cause: The ad creative (images or copy) does not resonate with the target, or the targeting settings are misaligned.

• Issue: High CTR but Low CVR

  • Hypothesized Cause: There is inconsistency between the ad message and the LP content. The LP's design may be poorly constructed, making it unclear what action the user should take next.

• Issue: Conversions are occurring, but CPA is high

  • Hypothesized Cause: The cost per click (CPC) is rising, or the targeting range is too broad, resulting in many unnecessary clicks that do not lead to conversions.

• Issue: CPA is within the target but ROAS is low

  • Hypothesized Cause: While conversions are being obtained, they are mainly for low-cost products or services that do not contribute to overall revenue.

6. Preparing for the common sense of the future: Responding to new privacy-protecting technologies (Topics API)

The abolition of third-party cookies represents a structural change in the digital advertising industry. To adapt to this change, understanding and preparing for new privacy protection technologies is required.

At the center of this is Google’s Privacy Sandbox. This initiative develops new technical standards that allow for ad delivery and effectiveness measurement while protecting user privacy.

One of its core technologies is Topics API. Traditionally, ad platforms tracked users, but with Topics API, the browser itself protects user privacy while providing anonymous interest topics to advertisers. Specifically, instead of cross-site tracking of individual user behavior, the browser assigns a broad interest "topic" (such as "fitness," "travel," "cooking," etc.) based on the user’s recent browsing history. Advertisers can deliver highly relevant ads to these anonymous topics, without infringing on individual privacy.

This transition is not a distant future issue. For example, Yahoo! Ads will start implementing Topics API from January 2024 and will begin observing and obtaining topics.

Now, marketers need to focus on the following three things:

• Strengthening Utilization of First-Party Data: Enhance the collection and organization of customer data (email lists, CRM information, site login information, etc.) that is directly collected without being affected by cookie regulations. This will become the most valuable asset.

• Re-focusing on Contextual Ads: Increase investment in contextual ads linked to the page's context, rather than user data. This is an effective targeting method even under privacy regulations.

• Improving Creative Quality: Since targeting granularity may become coarser, the ability to create compelling creatives that move users will be more important than ever. Differentiating through the appeal of the message and design quality will directly impact results.

Conclusion

To achieve results in modern web advertising operations, it is essential to understand and implement the six new common sense practices. Advertising is not a one-time setup; continuous PDCA cycles are fundamental. Targeting has evolved from demographics to user intent and context, and daily improvements are supported by data-driven A/B testing. Furthermore, at the core of operations is automation by AI. To evaluate results accurately, it is necessary to interpret the relationships of each metric and adapt to the industry's shift towards privacy protection.

Modern web advertising is no longer merely a budget battle; it is a competition of agility. Success is defined not by the initial campaign setup, but by the speed and intelligence of the optimization cycles.

However, integrating data from multiple media, optimizing budgets using AI, and executing daily improvement plans based on data is a complex and time-consuming task.

What I would like to introduce is the AI marketing and ad operations optimization platform "Cascade". Cascade functions as an AI analyst that automates the complex data integration discussed in this article, allowing you to focus on strategic decision-making. It automatically integrates and analyzes data from several ad media, such as Google Ads and Meta Ads, identifying wasted ad spend and growth opportunities while proposing concrete improvement plans to maximize ROAS. It is a powerful tool for efficiently applying the new common sense discussed in this article. If you are interested, please check the details.