Useful information | Data analysis thinking is essential for marketing promotion!

In this Internet age, we set up most of our marketing scenarios on the Internet.

When doing marketing on the Internet, we need to have Internet marketing thinking, which can help us sort out issues such as the direction, goals, planning, and execution of Internet marketing. Similarly, as an advertising optimizer, you also need to have data analysis thinking when doing analysis.

Data analysis thinking can effectively help us optimize our accounts and also improve our work efficiency.

For advertising optimizers, the backend operations are already familiar to them. But when operating, we should understand:

• What's wrong with the account?
• What is the purpose of the operation?
• Why do you do this?
• What results will be achieved after the operation is completed?

Most bidders frequently adjust the background in their daily work, such as negating words, increasing prices, reducing prices, adjusting matching modes, etc., but only a few people know why they do this and what effects it will bring.

This is why many bidders have worked for many years but have not become real advertising optimizers and are merely backend operators.

The reason for this problem is that we do not have a correct data analysis mindset when doing data analysis.

What is data analytical thinking?

I will explain it from 5 aspects below:

1. Find the problem
2. Analyze the problem
3. Develop a plan
4. Implementation plan
5. Review

First, let’s look at a set of advertising data that was originally operated by a client:

Find the problem

After looking at the data in the above figure, we can see at a glance:

Question 1: The number of leads in Guangzhou, Chongqing, Zhengzhou, and Xi’an is very small, with the values ​​being 40, 20, 4, and 6 respectively, and the average lead cost is relatively high.

Question 2: Chengdu’s average lead cost is the lowest among the six cities, but the number of leads is only 70. Although it is slightly higher than the above four cities, it is still relatively low compared to Shenzhen’s 211 leads.

Analyze the problem

After preliminary observation, we found the above problems, but this is just seeing the surface.

The most important point in data analysis thinking is: see the essence through the phenomenon and solve the essential problems in order to effectively optimize the advertising effect.

On the surface, the average lead cost is relatively high due to the small number of leads in Guangzhou, Chongqing, Zhengzhou and Xi'an, but we need to be clear about the fundamental reason for the small number of leads?

(1) Guangzhou

Let's first compare the data of Shenzhen and Guangzhou:

From the above data, we found that the lead rate in Guangzhou is only 60.6% , while the normal value should be above 80% .

Generally speaking, the reason for the low lead rate is related to the customer service's business capabilities, so I checked the Shangqiao conversation records and found that there are differences between the audience groups in Guangzhou and Shenzhen. Using the same sales pitch, consultants in Shenzhen can get leads and even close deals, but consultants in Guangzhou cannot. This is the problem.

The display volume and click volume in Guangzhou are about 1/3 of that in Shenzhen, but the consultation volume is only 1/4 of that in Shenzhen, which results in a consultation rate of only 4.85% in Guangzhou.

Based on the problem of low consultation rate, we conducted in-depth analysis and found that the landing page was not well made and the conversion ability of the landing page in Guangzhou was poor. In order to increase the consultation rate in Guangzhou, we can create a separate landing page that is suitable for the Guangzhou population for conversion. In terms of lead rate, we can provide language training for customer service staff.

According to the current 66 consultations in Guangzhou, if the lead rate can reach 80.5% of Shenzhen, there will be 53 leads and the cost can be reduced to RMB 368.83.

(2) Chongqing

Chongqing is actually similar to Guangzhou. The essential problem is that the consultation rate is relatively low, only 4.15%.

If the consultation rate can be raised to 7.03% in Shenzhen, the consultation volume will increase by 1.66 times and the consultation cost will drop to RMB 165.81.

(3) Chengdu

We analyzed the problem of low number of leads in Chengdu:

• Number of leads = number of inquiries × lead rate
• Number of consultations = number of clicks × consultation rate
• Clicks = impressions × click-through rate

According to the formula, the number of leads is related to the lead rate, consultation rate, click-through rate, and impression volume.

From the above data, we can find that Chengdu’s lead rate is the highest among all cities at 87.5% , the consultation rate is 10.10%, and the click rate is 2.92% , all of which are relatively good values.

However, Chengdu’s display volume is only 23% of Shenzhen’s, so the essential problem is that Chengdu’s display volume is relatively low .

If the results are good, we can try to expand the volume and bring in more conversions .

(4) Zhengzhou and Xi'an

Finally, we analyzed the reasons why there are few leads in Zhengzhou and Xi'an, and found that the combined consumption of these two cities is only slightly higher than that of Chongqing.

The traffic volume in the two cities is relatively small, so we need to further expand the volume so that we have enough data to support our analysis of the causes. Only with sufficient data can we correctly analyze the problem.

The problem of these five cities seems to be that the number of clues is relatively low, but when actually analyzed, each is different. This is what we call seeing the essence through the phenomenon , and the essential problem is the problem we need to solve.

Don’t make any adjustments to the backend as soon as you see a problem. This is a common problem among bidders. Blind adjustments will sometimes only make the situation worse .

Develop a plan

Based on the analysis of the above problems, we began to develop specific optimization plans:

When formulating an optimization plan, we must clearly understand the purpose of the optimization plan, and we need to quantify the goals and set KPIs (key performance indicators), which will help us review and test the results later.

Implementation plan

After that, we will follow the plan and will not go into details here. But it should be noted that before execution, we need to set a time period to test the effect.

During this cycle, if the effect is minimal, we will implement the next plan. After the entire plan is implemented, we will start to review.

Operation log

When we operate the background, we must make records before the operation to clearly understand what we have done.

This can avoid operational errors, and when problems occur, you can find the cause. You can also clearly see what operations have been performed and whether the operations have any effect.

This is what a rigorous advertising optimizer needs to do, and it can also avoid a lot of unnecessary trouble.

Review

Review goals : What are the KPIs of the program?

Effect comparison : Has this KPI been achieved?

Analyze the reasons : Why is it not implemented?

Summary and optimization : What adjustments are needed afterwards?

Looking back at what we did? Understand the essence of the problem from this.

Reviewing is a process of self-analysis. We need to review key events and conduct in-depth analysis of problems that arise in key links. For example: Why did it not achieve the expected effect? Is there an operation problem? Is this the wrong direction? How to solve it?

When reviewing, you need to constantly ask yourself questions, find answers from the review, and understand where the problem lies. Reviewing can effectively avoid making the same mistakes in the future, turn experience into one’s own ability, and improve work efficiency.

This is an essential skill for a qualified optimizer and also a way to continuously improve yourself.

Summarize

The purpose throughout the entire data analysis thinking is: to deeply understand the essential purpose of action .

Never optimize for the sake of optimization, but perform data analysis and optimize ads in order to improve advertising effectiveness!

Source: Growth Superman