How to make data analysis report? How to make Baidu bidding statistics report?

Many people who have just started SEM find data analysis difficult, especially when they see a pile of densely packed data, they don’t know where to start and are confused.

This is a confusion that everyone will experience, but as long as you continue to explore and learn, you will find that data analysis is not as difficult as you imagined!

But, be careful about the method!

I have seen some SEMers who, when analyzing data, would import keyword reports, and if they saw the click-through rate was low, they would quickly change the creative; if they saw the conversation rate was low, they would change the URL; if they saw that the keywords were not ranked or displayed, they would select all the keywords in the unit without saying a word and raise the prices in batches.

How can this be called data analysis? This is simply nonsense, worse than those bidders who can only adjust the price!

It is important to understand that data analysis is not done just for the sake of doing data analysis. When looking at a problem, one cannot just look at the details. One cannot simply put out a fire wherever it breaks out. One must consider the problem as a whole and solve the problem with a goal in mind so as not to get bogged down in a futile process.

To make it easier for everyone to understand, let's look at two sets of comparative data from different time periods:

(Generally speaking, in order to view the delivery effect more intuitively, we will use a comparative method to observe the data. There are two ways of comparison: one is month-on-month comparison, that is, comparison of the data of the previous cycle, such as this week's data compared with last week's data; the other is year-on-year comparison, that is, comparison of the data of the current cycle of the previous year, such as this October's data compared with the data of October last year.)

Looking at the data in the figure above, what can we see when comparing time B with time A?

Yes, at time B, operations of increasing volume (expanding the matching mode) and lowering prices were definitely adopted, because impressions, clicks, and conversations increased, and the average price went down.

But is there any actual effect?

No, the conversion rate did not increase but decreased, and the conversion cost increased by 100 yuan compared to before, so it can be said that it was a loss.

Now that we have discovered the problem, the next step is to find out where the problem lies. So what we need to do next is not to analyze, but to collect data and correlate the data.

First, open the Baidu promotion client, download the keyword data for the specified time and copy it into the Excel table.

Then export the conversation data and conversion data, and associate these two data with the keyword table (the associated data is associated using URL, and two functions are required, one is Countif, which is used to count the number of conversations and conversions; the other is Vlookup, which corresponds the number of conversations and conversions to the corresponding data source).

The following is an example of the sorted data:

After collecting the data, the analysis phase begins.

Generally, the analysis data will be analyzed from two dimensions: planning level and keyword level.

1. Planning level

The purpose of viewing the plan report is to observe the account from a larger perspective and understand the conversion capacity of each plan.

Therefore, we can directly use the sorted keyword report and make it into a pivot table to facilitate viewing the effectiveness of each plan.

The following is an example of a pivot table:

In this way, the conversion capacity of each plan is clear at a glance.

If the account has a small budget, you can suspend the plans with few or no conversions and save the budget for those with good conversion effects.

If the account has sufficient budget, you can increase the volume of those plans with few impressions to observe their actual conversion effects;

If it is a multi-product (multi-region) account, you can see more intuitively which product (region) converts better, and you can increase investment in that plan.

This is analyzing data from the dimension of planning. The analysis method of unit report is similar to that of plan, so I will not go into details.

2. Keyword level

Keyword analysis is a very important indicator in data analysis. No matter whether you find a problem in the plan or unit dimension, the final operation must be implemented at the keyword level.

When analyzing keyword data, you need to use the keyword report compiled above. I will not repeat the production method as it has been explained very clearly above. The only thing to note is that you must add the tracking code to the URL of each keyword.

From the above picture, we can clearly see the consumption, click, display, conversation and conversion data of each keyword. In this way, it is easy to judge the effectiveness of each keyword, and then it will be much more convenient to use the four-quadrant analysis method to analyze the data (the analysis of creative data is similar to keyword data, and the four-quadrant analysis method is also applicable).

◆ The first quadrant: high consumption and high conversion. If the average price is reasonable, this is a very good type of keyword, which means it is easy to be seen, easy to be clicked, easy to generate inquiries, and has a high conversion rate.

This type of word needs special attention, and you should focus on its ranking. If the average price is not ridiculously high and the conversion cost does not exceed the expected amount, don't lower the price easily.

◆ Second quadrant: low consumption, high conversion. Low consumption means fewer impressions or fewer clicks, and high conversion means the keywords are accurate enough.

This type of words should also be paid attention to. If the low consumption is due to too few impressions, then match it. Usually, you can just ignore it.

If the consumption is low because the average click price is too low, check the ranking. If the ranking is low, try to increase the price;

If there is a ranking, or it is between 1-2, then check the creative to see if the ad is not attractive enough.

◆ The third quadrant: low consumption and low conversion. Low consumption means fewer impressions or fewer clicks, and low conversions mean inaccurate traffic.

This type of words can be operated in large quantities, the matching pattern can be expanded, and more negative words can be used. After the volume increases, gradually observe the effects of clicks and conversations. If the clicks are poor, focus on the creativity; if the conversations are poor, focus on the landing page.

◆ The fourth quadrant: high consumption and low conversion. High consumption means many clicks or that the average click price is too high, and low conversion means that the traffic is inaccurate or the keyword itself is a traffic word.

These words are the culprits that lead to excessively high conversion costs, and they must be nipped in the bud. They are words that deserve special attention.

When analyzing and optimizing such words, first open the search term report to see if the matching of the keyword is messy. If it is messy, it is a rejected word. If it is normal, pay attention to whether the creativity and dialogue are attractive enough and whether there are any problems with the consultant's words.

If none of these problems exist, then see if the average click price is too high. If it is too high, lower the price; if it is reasonable, reduce the matching amount.

Some people say that words like this that don’t convert can be paused or deleted. If the budget is sufficient, I don’t think it’s necessary. Just run at a low price and in large volumes, and it will be fine. This is because the search terms brought by each keyword are different. Keyword A may bring search word B, and keyword B may bring search word C. This will not only make it easier to expand your keywords, but may also sometimes bring about surprising conversions.

The above is the method of data analysis. To summarize: when doing data analysis, you must first compare the data, discover problems through comparison, and then find problems by collecting data. From plans and units to keywords, we refine the data step by step, and finally implement it into specific operations to solve the problem.

If we break it down step by step like this, data analysis is actually not difficult. We only need to remember one thing when analyzing data: is the money worth spending? If it is not worth it, then there is a problem and you need to find out the reason.