This is my 15th original article
Magic prescription
One day, there were two officers, both of whom had the same fever and headache, but hua Tuo’s prescriptions were quite different. After the original diagnosis, one of them was known to be superficial syndrome, which could be solved by sweating method. One is in heat syndrome, non diarrhea is difficult to treat.
This is a short story of Hua Tuo spread among the people. The core idea is to express hua Tuo’s marvelous medical skills. The same representation was cured with different prescriptions, which was unknown to outsiders
Hua tuo
Photo source: Baidu Baike
A nasty stomach ache
Again, there are so many things in modern medicine that have the same representations, different pathologies. The patient’s description is not clear enough, resulting in a simple “stomach pain”, involving dozens of causes. Below is the lilac garden collation of abdominal pain reason chart, the average person estimates to see dizzy.
Lilac Garden drawing
When we were doing data analysis, we saw a drop in GMV/ sales, which is the same as “tummy ache”. To a large extent, the enterprise’s internal business model, target group, external market conditions, policy factors; At a smaller level, the rules of the operations-level activity, the design of the product-level landing page and subsequent processes all contribute to the decline in the end result.
Medical solutions
Medicine has a very sophisticated solution to the complex causes of abdominal pain, and there is even a discipline called Diagnostics that studies how to determine the cause step by step.
But unfortunately, data analysis is a commercial activity, and there is no unified operating standard in the industry to deal with such a complex situation. Probably because the business subject of each analysis is too personalized.
You can look at the teaching content of “diagnostics”, including symptomology, history collection, various examinations, clinical diagnostic steps and writing of medical records.
Data analysts, do you ever feel the same? Is it very similar to how we do data analysis?
Symptomology — Business understanding
History collection — data collection
Various checks – various metrics, reports, temporary SQL boosts
Clinical diagnostic procedure – root cause search
Medical record writing — Data analysis report (excluding plan and calculation)
The figure above is a 4-step data analysis method sorted out by me based on my work experience, which is actually very similar to the teaching content of Diagnostics.
When I made this discovery for the first time, I was very excited. The knowledge of the two fields confirmed each other, so I gradually began to study how medical science finds problems, explores problems, identifies root causes and solves problems.
Steps to identify the cause
\
See watermark for the image source
The picture above is a medical diagnosis of “stomach pain” standard operating procedure, confirm one step, then further collect information, confirm another step, then collect information, gradually close to the truth. It’s the same thing we do with data analysis. \
Of course, very flexible people will immediately think: the reality is much more complicated than this, often a mixture of circumstances, so how to deal with it?
The branch of medicine that studies this condition is called Gerontology, and older people often develop multiple diseases. In “Union Gerontology”, the strategy for this situation is the principle of “individualization of the whole person and team-based diagnosis and treatment”. In the acute stage of the disease, “recovery” is the goal, while in the chronic stage, “stability maintenance” is the goal, and the overall goal is to maintain the functional status of patients. Get on The Bo Shining proposed “the first domino rule”, spread more widely called “20 rules”. In fact, it is the same logic: to find the most critical factors, acute factors should be solved first, long-term factors should be solved slowly.
Little desire
I have been working in the field of data analysis for more than 10 years, involving various fields of 2G, 2B and 2C, from ETL data processing, data warehouse construction and data center at the technical end, to statements and multidimensional analysis at the business end, and then to business analysis and business model. \
The data analyst is like the doctor of the enterprise. In contrast to medicine, the data analysis is far from systematic and standardized. I expect to be able to summarize a complete set of data analysis procedures from disease to root cause, just like diagnostics, so as to help everyone out of the box.
This is my little wish, and I will make slow but steady progress towards this goal. I will continue to create, continuous optimization, hope to help you, but also hope to get your support.
Highlights from the past
Real fuck | building data warehouse process
Field thinking | | build enterprise data analysis system data analysis how to facilitate the business
Retweet, like, watch, arrange?