Auction stat discrepancies where your traffic is lost

** Liu Chaoyang's journal.

The problem of discrepancies in traffic statistics has been plaguing the majority of bidding users. Whether it is Google or not, there is a certain difference between the billing traffic counted in the background and the traffic counted by third-party statistical tools, and some of them are as high as 30%. Through the comparative analysis of actual cases.

Let's take a simple case for comparative analysis.

Here's a set of data that we provide in the auction admin, with 173 clicks. The data includes search and network alliance data. Comparing another statistical tool, let's look at the specific data differences.

You can take a look, the statistical data here is 145ip. Why is the difference so big, and what is the problem?

Look for the root cause of the problem.

Here, we need to analyze the principle of bidding traffic calculation from several sources. As we all know, the billing mode is URL to trigger billing. In other words, when billing, the link must be triggered to calculate the cost. Practically all search engines are in this pattern.

Let's look at it this way:

In the end, we can find that billing, as long as this link is triggered, is a normal deduction. That's why we click on the same ad multiple times and go straight to the destination page without going through this link. Therefore, the malicious clicking behavior of the vast majority of software modes will be invalid and blocked.

Then, as long as such a link appears, then it will be deducted. Obviously, search engines don't lose traffic, so the problem of traffic reduction must be in other areas. Let's analyze and argue further!

The so-called "security alliance."

In order to resist the counterattack of 360, a security alliance was formed in September 2012, and many ** were intercepted and warned of risks.

360 browser, Kingsoft Antivirus, Computer Manager and other software all have risk warnings, which directly leads to the inability of traffic to enter**.

However, Kingsoft and other anti-virus software directly intercept the page, and the link that triggers the billing takes effect normally, but the user only sees a risky page. There is no doubt that after customers see such a page, 99% will close, so the traffic cannot be counted but the fee is deducted normally!

Practical solutions.

In fact, when we encounter such a problem, we lose not only advertising costs, but also customers. Such ** directly leads to a decrease in customer trust and greatly reduces the conversion rate. So, solving this problem has become very tricky and urgent.

In the actual combat experience, the advertising industry often encounters similar problems.

2. The content is encrypted, so that the interception system cannot intercept normally;

3. **** processing.

More exciting, we will decipher it for you in the following content. Stay tuned to the advertising circle! (The 7 groups of the advertising circle have been opened, welcome to join the 7 groups of exchanges 217868465 (Note: Advertising circle exchange and learning).

Peers are chaotic, ** in order to complete the performance chaos loss of customer advertising fees, so that customers can recharge, click and go, stay less than three seconds, statistical data statistics are not completely accurate. The reason for your own **, such as **instability, can't be opened, etc.!

Analyzing the promotion data in the background, it is nothing more than the following points.

In the search term report, first look at the words to see which words have the highest consumption recently, which words have the most clicks, and which words show more clicks and fewer clicks. Whether there is value in words with high consumption or more clicks, see if there are words that no one clicks on, analyze the reasons, and whether the ranking of words needs to be adjusted. Take a look at whether the matched keywords are useful, and block them if they are not useful.

Promotion time, how much to consume in the morning and afternoon or other times, depending on whether there is more consultation in the morning or more consultation in the afternoon, you can give more budget in the more time period.

If there are any ideas and URLs that are not displayed, you should take a look at those that are rejected, and the ideas that are not displayed can be decisively deleted and re-added.

What I said is still relatively simple, and for a comprehensive analysis, you can consult your customer service. If you understand the background or you can see it yourself for a long time, if you don't understand, it is recommended to let the customer service show it to you to prevent you from misoperation.

1 Determine the purpose of data analysis.

The purpose of data analysis is to make the account perform better, so we first need to know what problems the account currently has.

In this regard, we can analyze it through the marketing flow chart.

The marketing flow chart refers to the core data of the account centered on different dimensions, and calculates some auxiliary data according to the core data, such as click-through rate, dialogue rate, click cost, etc., to help us better discover diseases through the comparison between different data.

For example, the following diagram shows the marketing flow chart for an account.

This account is only promoted from Monday to Friday. From the figure, we can see that consumption is gradually showing a slight downward trend, but the overall click does not fluctuate much; However, the number of leads is gradually decreasing, and it is a slight increase at the end of the month.

For the time being, we can compare the data with the data with the worst (20-24).

The following figure is a screenshot of the data I have made.

From this set of data, we can clearly see:

1.The impression has decreased, but the click has not changed.

2.Total dialogue declined, but consumption increased slightly.

Overall, there is a downward trend in conversion.

Once the disease species has been identified and the purpose of the analysis has been clarified, the next step is to understand the direction of optimization.

2. Clarify the direction of optimization.

What kind of data do you typically analyze first when analyzing accounts?

Let's take the planning dimension as an example, let's analyze the data.

The image below is a screenshot of my account plan data.

The first plan. Comparatively speaking, it has the highest click-through rate, and the conversion and average price are also acceptable. But the conversion costs are a bit high. So, we can do traffic operations on the plan.

The second plan. It has the highest conversion rate, and the click-through rate is not too high or too low. So, we can be creative with the optimization of the plan.

The third plan. On the whole, the click-through rate, average price, and conversion cost are all acceptable. Therefore, we can expand the plan.

3. Determine how to optimize operations.

Knowing the disease and clarifying the direction of optimization operation, the specific operation method is the specific method.

According to the different requirements and goals of different industries, you can communicate with your peers more.