Data is accumulated over time. A large-scale enterprise has a large number of reports in excel. After a long time, I feel various pain points in excel making reports: what data collection is troublesome, the data in various systems cannot be connected, etc., I also thought of using a reporting platform or bi platform to manage data analysis in a unified way. On the one hand, it will reduce the pressure of various business personnel. After all, no one wants to spend too much time on reporting. In many cases, reporting After I finished it, I was no longer in the mood to look at the data in it. I just had a little energy, and you actually asked me to do the report.
But when we were going to start a bi platform, there were also problems. Our company had accumulated so many reports, from dozens of reports to hundreds of reports. If you conduct research, each department can come up with a bunch Come on, everyone is very happy with so many needs, thinking that this is good, and don't do these things by yourself in the future, but he didn't know that the project manager opposite him saw so many reports, and he felt in his heart. The shadow area can surround him now.
And automating all of their needs on the platform is a successful platform? It should not be. We carefully studied the requirements put forward by each department and found that a lot of content is repeated. For example: a commodity department wants to analyze sales performance, he cares about the sales of self-operated products on the day, cumulative sales, Sales volume of the day, cumulative sales volume, gross profit, profit margin, customer unit price, year-on-year growth rate, etc. b The Commodity Department also has a sales report. What he wants to see is the same as the above, except for the cumulative sales. But he insisted that the cumulative sales field should not be displayed, and must be distinguished from the department A. Why do I feel a little cheated? Is he bullying us outsiders?
The proposer of demand is based on individuals, at most various departments, but the builder of the platform should be based on the entire enterprise. The responsibility of the data platform should not only be used at the data level to centrally display data, improve the level of data automation and analysis capabilities, but also use the data analysis platform to unify the company's management caliber and integrate various data calibers. Unification, different analysis indicators are unified, so that the whole company has a unified understanding and judgment of one indicator, which can also reduce the communication barriers between departments in the enterprise and improve the communication efficiency between departments.
From another perspective, after unifying these contents, you can focus on the report function and make each report more targeted. As the so-called condensed essence is the essence, you have more than a thousand reports on the platform, I really don’t want to click Go in and see. Therefore, before starting a data platform project, it is necessary to plan these messy reports in a unified manner.
At this point, how should this plan be done? Let’s take the retail industry as an example. Don’t look at everyone’s reporting needs, but they all revolve around several themes: sales, inventory, procurement, logistics, membership, merchandise, Just look at the indicators through different latitudes.
So, the first step is to do the classification first. So before planning to sort out these messy reports, you need to think about how to classify and integrate these reports, see the figure below.
Some things are complex, and in order to be able to recognize that complex content, one might simplify it by categorizing it. What's more, it is easier to classify something that is not complicated like a report. For the time being, the reports are classified according to the above three dimensions. No matter how many reports are, they can always be classified into the above categories.
The second step: in terms of primary and secondary, three dimensions, will it be very chaotic, of course it will. Still need to choose a main classification dimension. A report often involves multiple business modules, such as inventory and sales, which are often put together, and often involve multiple user groups. For example, the sales performance report must be concerned by many people.
Therefore, it is recommended to divide the report function level, and describe the other two dimensions. form the following format:
- Data query report:
- Involving business modules: sales, inventory;
- Report name: real-time index query; user group: store manager (middle level), category manager (middle level)
Step 3: The report is sub-compressed, and the collected tables are marked, and marked according to the above method. The hardest part of this step is getting started. When you see a lot of unclassified reports, you feel like you don't want to start, but if you really start, it's over. After the division, there will be multiple reports under the same report type and the same business module, for example:
- Data query report:
- Involving business modules: inventory;
- Report name:
- Inventory structure analysis table
- Departmental inventory structure table
- Inventory structure analysis table (department)
- Inventory structure analysis table (brand)
- Inventory structure analysis table (single product)
- Single product turnover table
- inventory cost table
- Out of stock out of stock statistics
- Supplier out-of-stock out-of-stock statistics table
- Purchasing Out of Stock Out of Stock Ranking Table
- 。。。
这些在同一个报表类型、同一个业务模块下的报表，都是有压缩空间的，其实我们仅仅通过名字，也知道很多是可以合并在一起的。在这个目录下，只要满足了目的相同这一个因素，就是可以合并在一起。
目的相同，也就是看这个报表都是为了同一个目的，那么就不必分开来。如下图所示：
上面两个表，一个是缺货率一个是断货率，目的都是为了更好的管理库存，适时采购，减少商品的缺断货情况，那自然可以放在同一个表中实现，最多再参数面板上加选项而已。
再比如以下三个：
这几个表的目的在于了解商品库存情况，剩余库存结构，从而更好的管理库存，或者执行商品促销策略，降低库存成本或损失。所以上面几个表同样完全可以放在同一个报表中，甚至采用图表结合的形式，更友善的展示上述内容。
（报表的整合方法不绝对，更多的是要参考实际的场景的，但是依然这样写出来，只是提了一个思路，并且给大家一个敢于整合的信心，毕竟还发现了有人看见这些乱糟糟的东西就不想整合了的）
通过上面对大量的报表进行整合，报表一定会精简很多，但是这个过程也一定是不容易，比如我就知道项目经理一定是会遇到这种阻力的，业务部门a：你凭什么把我这个需求去掉啊，我之前一直都是那样看的，看了两年你让我改，还有人性吗！ 对不起真不是我没人性，是你这报表没人性啊，你还记得你入职的时候所许下的誓言吗？所以这次还请忍一忍吧，毕竟整合之后，统一公司的管理口径，也为以后报表平台的扩充建立一个好的基础，长期来看，对所有人都说利好的嘛。
另外，把企业所有的历史报表进行整合，自动化在报表平台中就结束了吗？其实也仅仅是个开始，当把这堆报表整理好后我们回头看，会发现其实报表类型很少的，绝大部分都是数据查询类的报表，最多有些日常管理类的。说好的数据分析平台来提升数据价值呢，怎么好像只是将日常的工作给自动化了一下呢。对企业来说，对数据的需求也不只是满足于日常的工作，同时期望于提升数据价值，所以必然会需要更多的管理报表和分析报表。所以，整理历史报表只是搭建报表平台的第一步，后面还有很长的路要走。既然是后面的路，那就先留在后面吧。这次不偏题了。
文 | jiago王
文章出自：知乎专栏《撩撩数据吧》