What is analysis doing?

September 7, 2016 May 12, 2022 Marketing , SWOT analysis, 4P analysis, etc.

What is analysis doing?

What do you think of when you hear the word "analysis"?

I think there are various images, such as "handling numbers", "difficult calculation", "organizing the current situation", and "giving answers".

Now, I would like to explain what analysis in business is doing.

(1-1) The first thing to do when analyzing

In order to analyze, you have to clarify two things.

One is to clarify what the "analysis" is, and the other is the purpose of the analysis. For example, if you want to analyze the number of sites, decide whether the number of PVs or the number of sessions is confirmed.

This is the analysis target. If you have browsed, it is a good idea to analyze the number of PVs, and how many sessions have been visited.
If you want to analyze the number of sessions here, then decide the purpose of the analysis.

If you want to make sure that organic from SEO is increasing, you can analyze the number of sessions according to the inflow route. If you want to see when the number of sessions increases, it will be analyzed by the time series of the number of sessions.

In this way, if the purpose of the analysis and the analysis is clear, the rest is just calculated. Half of the analysis decides what to see and how.

(1-2) What do you see and how?

You need a monosashi to evaluate the analysis results.

In other words, you need indicators and standards to evaluate what the numbers you have analyzed are. This monosashi is also divided into two.

One is "absolute evaluation" and the other is "relative evaluation".

Analysis is easy if there is absolute evaluation. All you have to do is evaluate it with the indicator that was determined in advance. Most indicators in this case are prepared in advance.

If there is a standard of "less than 1,000 per day visits = improvement required", if the analysis results are 999, the answer will be improved.

The indicators and standards are to give an answer from the analysis results in this way. This is also called "plan", "budget", and "goal" by changing the expression, but it plays the same role in evaluating the number of the analysis results.

Relative evaluation becomes a little complicated. You have to start by making monosashi. In order to make monosashi, we collect as many analysis as possible.

Then grasp the tendency from the whole data. For example, if you want to analyze what day of the week you can attract customers, first create a group on all data by day of the week.

If you compare the total of each, you will be ranked 1st to 7th. If you make the ranking as a monosashi, you can judge the day of the week that you can attract customers. You can also make the ratio a monosashi, not the ranking. If there is a total of 7,000 customers a week, an average of 1,000 people a day, and if there are only 700 customers on Monday, it can be evaluated as improving.

(1-3) Evaluation and concept of analysis

The concept of "evaluation" in absolute evaluation and relative evaluation is very different.

If it is absolutely evaluated, apply it to monosashi and evaluate the result. If you do not replace the monosashi, you should be able to evaluate on the same criteria. If the analysis results are good evaluation, it can be said that the data when the monosashi is made is desirable compared to the standard.

This concept has been applied to monosashi such as "plan", "budget", and "goal". Conversely, you can measure the feasibility of each by what kind of monosashi is made by "plan", "budget" and "goal".

Relative evaluation is a little complicated.

The result of relative evaluation is only "current visualization". Therefore, it is necessary to first determine, "Is the purpose of the analysis a correct analysis?" Specifically, you can make three decisions: "you can make monosashi with information that can be judged correctly," "you can not make monosashi with information that can be judged correctly," and "I do not trust the analysis results because it is information that cannot be judged correctly."

If you use information that cannot be judged, or decide on a monosashi, you often guide you to unwanted results. When it is not possible to evaluate, it is desirable to conclude that there is not enough information.

(summary)

Analysis is "express and organize the analysis in numbers", "evaluate with monosashi", and when there is no monosashi, "make a monosashi based on the trends and differences to be analyzed". This is the same for simple analysis and analysis like machine learning.

And keep in mind that there is a judgment that you cannot analyze with the current information other than the calculation results.

Marketing , SWOT analysis, 4P analysis, etc.

Posted by LSD_ADMIN