This article mainly introduces the business unit (mainly the Operation department) how to better the data platform without perfectProduct Managerpropose data requirements to satisfy some of your own data analysis.
first, who will mention the data needs?
In General, more is the data demand for operations. But there are other business units, such as finance, billing, marketing, technology, and other business units.
second, what are the data requirements?
• Product data: PV, UV, page bounce rate, DAU, retention rate, event conversion rate, user attributes, and more.
• Business data: Number of registrations, downloads, number of orders, cancellations, registration conversions, order conversions, and more.
third, how to raise data demand?
For a chestnut: The National Day will be on the line of a new business promotion page, then the leadership needs to see this new business on-line after the effect of how?
Time Planning
Although the national Day on-line, but as a data demand, need to have time planning awareness. Do not wait until the active page goes online, and then go to the product, data group to mention the relevant data needs, but to learn to put forward the requirements. In particular, the active page of PV, UV and other user behavior data, because if there is no demand before the launch, development generally will not be actively buried point statistics, so on-line found no buried point, it is too late. In addition, although the business data can be extracted from the database, but because the demand time on-line, left to the data team processing time is not much, often lead to urge you to the data, you always urge the data team to help extract the basic data.
Analyze Dimensions
On- line an activity page, as an operation you need to know what the leader wants to see when the effect is on-line, and what data will be rendered. What basic data is needed for this process, and how to use the underlying data for analysis. The most important of these is the source of the underlying data. So it's important to be clear about the dimensions of the underlying data when you need to ask for data in the pre-operation. Avoid the data that the underlying data does not want to analyze, and reduce the time cost of communicating back and forth with the product or data team.
Hypothetical scenario:
operation: I need all the order data during National Day ...
Data clerk: What data do you want to order? Order number, User ID, ordering time, booking pickup time, booking time, order status??
operation: Oh oh, I want the order number, user ID, ordering time, booking pickup time, booking time, order amount.
Data Clerk: is the order data between 10.1-10.7th for placing orders? What is the order data for an appointment between 10.1-10.7th? Or is the actual pickup time between 10.1-10.7th and the order data?
operation: I would like to place order data between 10.1-10.7th in a single time.
Data clerk: Do you want all the orders? Or just make a reservation for a successful order?
operation: Just make an appointment for a successful order.
... ..
The above communication took half a day, this time the data clerk to the order number, user ID, place a single time, an appointment to pick up the goods, booking time, the order amount of the data to operate. Operations to grasp the analysis, the process of analysis found that the order amount is significantly larger than the average amount of daily orders.
operation: How big is the amount of the order you gave me? Did you pull the wrong one?
Data clerk Troubleshooting Half-day ...
Data Clerk: No, this is the order amount.
operational thinking half-day ...
Data clerk: Did you not exclude the discount amount Ah, the order amount to exclude the use of the preferential amount.
Data clerk Heart 10,000 grass mud horse ...
The above scenario, which can be said to be a data clerk or more responsible, will go and operate to confirm what data is needed. But it takes a long time and the processing efficiency is low. So how do you avoid this kind of situation? Requires operational clarity of the underlying data analysis dimension.
Taking the above scenario, you need to specify the time dimension, the order dimension, and the special circumstances of the required fields. It is better to be able to understand the order form characters commonly used paragraph, on the one hand can be specific to inform the Data Clerk what field to improve processing efficiency; On the other hand, it is easy to discern whether the data provided by data clerk is abnormal. It doesn't matter if you don't understand the table structure, but you need to tell exactly what type of data you want and what the special circumstances are.
purpose of expression
when it comes to data requirements, operations can simply describe the purpose of this data requirement. Like who wants this data? Why do you need this data? What is this data used to analyze?
There are three main advantages of expressing the purpose:
• Whether the demand is reasonable, the product, the data clerk as a demand receiver, can be based on this demand to assess whether the demand is reasonable, unreasonable situation can be called back to demand.
• Demand prioritization, product, data access to the needs of a lot of, how to schedule your needs, depending on the needs of you who proposed? What data do you want to analyze? What problems can be solved? The assessment requirements are prioritized and then scheduled for processing.
• Whether there is an optimal space for demand, for the purpose of operational expression and the underlying data required, in order to better meet the expressed purpose, the underlying data is not an optimized space, easy to provide better basic data for operational analysis.
Iv. Summary
to put it simply, operational data requirements are better than users ' demand for products. Because the demand is put forward by the internal staff, occupy certain initiative. So we need to clear the time, dimension, reason three elements, so that the product, data clerk to complete this data needs efficiently and accurately.
The above is a combination of their own work, due to their own company encountered the related problems, so there may be some limitations. You are welcome to exchange and learn more.
Source: Everyone is a product manager
Some tips from the operations department on data requirements
