Effective account design for Google Ads search ads (September 2018 edition)

The trend towards automation in Google Ads is on the rise

Recently, we've seen the words "AI first" and "machine learning" (hereafter referred to as "machine learning") used in many places, but to what extent are you benefiting from the machine learning features that have already been implemented in Google Ads?

As user behavior becomes more complex with the rise of smartphones, manual optimization of Google Ads is becoming harder to keep up with year after year. On the other hand, Google Ads continues to update daily to optimize ad delivery with machine learning and maximize performance. In order to benefit from machine learning, it is important to understand how Google Ads works and carefully consider how to build the foundation and structure.

So this time,

• Understanding the structure of Google Ads "Search-based advertising"
• "Hagakure Structure" is an account design that is said to facilitate machine learning

We will explain about:

*The previous "Google Adwords" was renamed "Google Ads" in July 2018.

Search Ad Structure and the Role of Each Item

First, let me briefly explain the structure of Google Ads search ads.
Google Ads search ads are made up of four main elements: "campaigns," "ad groups," "keywords," and "ads." What can be set at each level is as follows.

campaign

• Daily Budget
• Bidding Strategy Audience
• Delivery schedule setting
• Distribution area settings
• Negative Keyword Settings
• Ad rotation etc.

Ad Group

• audience
• Exclusion keyword settings, ad rotation, etc.

*The following applies to the settings of exclusion keywords and ad rotation, which can be set at either the campaign or ad group level.
・Settings in campaign → No settings in ad group = The settings in the campaign are applied to all ad groups. ・Settings in campaign → Settings in ad group = The settings in each ad group are applied.

keyword

• Match Type
• Exclusion keyword settings, etc.

Ad text (URL)

• 3 Headlines
• 2 Descriptions
• Display URL, final page URL, etc.

What is the Hagakure structure recommended by Google?

Now that you understand the components of search advertising, the next question is how to assemble them. Perhaps many people who have run search ads have created and run multiple ad groups in one campaign.
Of course, being able to finely divide and manage ad groups has its benefits, such as being able to finely adjust bids for each keyword, but in the current Google Ads environment, which encourages the use of machine learning, we recommend the simple structure advocated by Google, known as the "Hagakure structure."

The Hagakure structure is a simple structure with one campaign and one ad group for each URL (landing page).

Why one ad group?

1. Ease of managing operations
2. To aggregate operational data into ad groups

This is for two main reasons.

Google Ads' machine learning optimization is basically performed on an ad group basis, so if you divide your ad groups too finely, it will take a long time for each ad group to accumulate operational data, slowing down machine learning optimization.

A common pattern of dividing ad groups into smaller parts is that even if the keywords you want to acquire are the same, you divide the ad groups by match type, and multiple learnings are made for the same keywords. Of course, learning will progress better if you consolidate the data in one place.

Let's create a campaign using the Hagakure structure. For example, if you are creating an online shop selling luxury furniture, the structure would look something like this:

What's important is a simple correspondence between your campaigns, ad groups, keywords and ads and the content on your landing pages.

So, you might be wondering, what if you have a dining table and a coffee table, for example, and each has a different URL (i.e. different landing page content)?

To be honest, it is difficult to say that "this account structure is optimal" because it depends on the budget for each product and the number of types of products, but if the URLs are different, there is no problem in dividing them at the ad group level or at the campaign level. We recommend that you understand what can be set in the above campaigns and ad groups, and then choose a method that is easy to manage.

To reiterate, the important thing is that keywords and ad groups are grouped together for the URL.

Account operation flow designed with Hagakure structure

If you are planning to build and operate a campaign using the Hagakure structure, please refer to the operational procedures that we use.

1. Maximize your clicks and see the numbers

   Select Maximize Clicks from your campaign bidding strategy and start by finding out how many clicks each keyword can get within your budget.

2. Implementing the PDCA cycle for advertising

   Once you can see the number of clicks and CTR to a certain extent, the results of your ads will become clear. Ads that are not performing well may not match the user's interests or may not be relevant to the landing page. Increase the number of text patterns that are producing results or create new ads, and take the next action.

3. Optimize your CPA with target cost per conversion

   Once you have earned a certain number of clicks and are able to obtain a certain number of conversions, the next step is to optimize your CPA (cost per acquisition). Select "Target CPA" from the bidding strategy and set your target cost while referring to the conversion costs you have earned so far. If you have enough conversion data when setting up, the recommended target CPA will be displayed at the bottom of the setting screen, so please refer to it.

summary

This time, we explained about search ads and the "Hagakure Structure" recommended by Google. In order to improve the performance of your ads, it is becoming increasingly important to pay attention to and gather information about Google's features that are advancing machine learning. There are rules and patterns for optimizing machine learning, so by actually running the program and conducting various tests, you can not only promote machine learning but also deepen your understanding of machine learning, which will lead to maximizing your results.

Finally, we also accept consultations regarding advertising operations based on EC operations. Since we have experience in operating EC sites, we believe we can offer advice on advertising operations, so if you are interested, please feel free to contact us.