Three Practical Data Analysis Methods of Search Engine Promotion

People who do Baidu bidding promotion are no strangers to data analysis, which is a tool for account optimization. Almost every month at the end of the month or the beginning of the month, all kinds of data should be counted and reported to the leaders.

First of all, we have a correct understanding of data analysis and a clear understanding of why we should do data analysis:

1. The purpose of data analysis is to find a clear optimization direction.

2. Don't analyze mythical data. Data analysis is to improve the promotion effect and find a clearer optimization idea for optimizing the account.

3. Through the purposeful collection of promotion data, according to the number and changes of various data, find out the problems existing in the promotion.

4. Combined with the problems found, the optimization method is found by data verification.

5. Data analysis is to solve the problem of poor promotion effect.

6. We learn not only analytical skills, but also thinking.

Second, analyze the data from the following dimensions: plan, region, product and platform.

Take our promotion data in September as an example: (Some data involve trade secrets and cannot be made public, please forgive me! )

In September, I spent a total of X yuan, X threads, X paid threads, X free threads and X total conversations. According to the statistics of conversation words, the conversation leaving rate is 72%, the conversation rate is 3.3 1%, the average arrival rate of Baidu +360 is 68.8%, the click rate is 2.6 1%, and the average click rate is 6. 18.

As can be seen from the data, the click-through rate needs to be improved. Last week, the click rate was adjusted to 3.37%, the average arrival rate of Baidu was around 95%, and 360 was only around 50%. One is that the base number is too small, and the other is that the arrival rate is too low.

1) dimension 1: planning data

There are fewer customers about account plan 360, so I will not focus on the analysis, mainly taking Baidu as an example. Their plans are similar.

The "UPS Southern Jiangsu" plan was newly established in September, and the key words are duplicated with other plans, so there is nothing to show and it has no effect. After that, it is planned to promote it on the mobile side. Areas can be relaxed.

The product series with relatively high cost is "computer room and building control", which has reached every clue of X.

The schemes with better performance are "distribution type, primary type and precision air conditioning", generally UPS power supply and precision air conditioning product series. Focus on it.

There is little consultation in computer room and building control, and their quantity is relatively small, and a lot of consumption is spent on big talk with low intention.

2) Dimension 2: Region

The regional cost this time is a bit abnormal compared with last month. The average clue cost is X, and the clue cost in Nanjing, Wuxi, Changzhou and Hangzhou is as high as more than X yuan. I looked at it, because the consulting volume in these areas was ok before, so the bidding coefficient was increased, which in turn increased the cost. In addition, Wuxi Changzhou has a low consumption base and only one consultation.

The most clues are Shanghai and Suzhou. Other areas are not very stable.

Among them, Shanghai spends X yuan, Zhejiang X yuan and Jiangsu X yuan. The lead cost is X yuan respectively. X yuan. X yuan. Shanghai used to have the highest cost, and this time it has the lowest cost!

Third, establish a general method to solve the problem.

There are many analysis methods, and the following are common analysis methods:

A, disorder analysis method (thinking)

We all have chaotic thoughts, and disorderly analysis can make our thinking clear, generally using mind maps. For example:

1, list the problems.

Make a list of problems that disturb your mind, or you can use a mind map.

2. Classification and summary

Classify and summarize complex problems.

Step 3 give priority to

It is important to know what these problems are.

Step 4 find out what the problem is

Find a problem that must be solved immediately.

B, logical analysis (sequential thinking)

Clear purpose:

First, identify a problem that needs to be solved, and make clear the motivation for thinking, such as low exhibition volume, low click-through rate or low inquiry rate.

Cause analysis:

Analyze in detail what factors will lead to this problem and what are the possible reasons, and list them one by one.

Depth analysis:

Only by determining which inducement is the key to the real problem can logical thinking see the problem more thoroughly.

C, flashback analysis (reverse thinking)

Time node: push back the completed project and make clear the final time node of everything.

Backward goal: define the ultimate goal and then define the key indicators of each step.

Reverse thinking: Looking at the problem from the opposite angle, you can see through many things.

Finally, it ends with two classic sentences by teacher Zhao Yang:

Open source first, introduce high-quality traffic as low as possible;

Reduce expenditure again, reduce the average price of keywords, and limit consumption in some dimensions.