Do a good job in data analysis and accurately solve the problem of delivery

In advertising, everyone uses the same operating background. Why do some enterprises have good promotion effect and low customer acquisition cost, while others have poor promotion effect and unstable effect? The gap lies in data analysis, which can be said to be an important indicator to measure the ability of advertisers.

Poor ROI, sudden increase in cost, continuous decline in conversion rate, and poor effect... All problems, if you want to solve them, should be attributed to specific figures.

Through data analysis, determine the direction of optimization and adjustment, make flow control more targeted, and accurately solve the current problem.

This article comes from the official course of Baidu Marketing. Click the link below to get the official course materials.

https://aisite.wejianzhan.com/site/wjzzt0cw/5e4c6dc2-b9ca-419a-b240-d7be65482afd

What is data analysis

Data is the cornerstone of all analysis. Data analysis refers to the process of processing and analyzing a large amount of collected data with appropriate statistical and analytical methods, extracting useful information from it and forming conclusions, and ultimately promoting and optimizing the follow-up business effect.

Four functions of data analysis: problem diagnosis, strategy formulation, description of current situation, and prediction of results.

Data analysis process

SEM data analysis is mainly carried out around three core links: account, page and transformation.

Data analysis first finds the core number, then determines the reason for the number exception, and finally determines the solution based on the actual situation of the account.

It is divided into 6 steps: determining goals, collecting data, sorting out data, analyzing and reasoning, adjusting and optimizing, and reviewing and observing.

Advertisers can find the core data and determine the optimization goal through the five links of the marketing link. The influencing factors of each link are different, so it is necessary to trace back to the source through different analysis methods and ways to dig out the problem.

Data analysis method

There are many methods for data analysis, including backward extrapolation, inductive analysis, deductive analysis, trend analysis, comparative analysis, etc.

This section focuses on the backward method, which is based on the principle of marketing funnel. When there is too little data in one link, go to the previous link to find the reason.

The marketing funnel is the same as the advertising marketing link, which is divided into five links: display, click, arrival, consultation and clue, corresponding to six core data: click rate, arrival rate, conversation rate, effective conversation rate, clue rate and closing rate.

Problems in any link will eventually result in low turnover. From the level to level of turnover, we can deduce the following: clue volume - effective consultation volume - consultation volume - number of arrivals - number of clicks - display volume.

If there is too little data in a certain link, it will jump to the previous link until the link with too much data is found, the problem of breaking down the elements and the reason is discovered.

By restoring the user transformation path and analyzing the effect of each transformation node, the backward method is an important analysis model that scientifically reflects the user behavior state and the user transformation rate at each stage from the starting point to the end point, and can be used preferentially in data analysis.

Baidu data analysis tool

Baidu system provides advertisers with a variety of data analysis tools, including background data reports, Baidu statistics, basic wooden fish, kaleidoscope and stargazing disk, and supports the analysis of advertising, user behavior, material effects, market industry and other data.

For SEM launch, background data report is the most important data source.

Baidu SEM data report can clearly present the corresponding effect data of each launch setting. High timeliness feedback on the fluctuation of launch effect. Each medium granularity data provides a path to quickly locate the problem, and supports convenient access to required data and analysis of various business resumption trends.

Baidu SEM data report includes promotion report, orientation report, creativity report, landing page report and special report, which can provide relevant data in the direction of account building, orientation setting, creativity, landing page, etc., to facilitate advertisers to conduct multi-dimensional data analysis.

A good operation is inseparable from data analysis. When advertisers encounter marketing problems, it is recommended to start with data and use data analysis to feed back the launch, so as to directly hit the key points and make scientific and accurate decisions.