Report again? How much can you?

Hello, Habr!



Many employees of large companies, I think that almost everyone, has to do and submit reports. The overwhelming majority of this type of activity causes a negative reaction. Let's take a look at what reports are in essence, why they are needed, and most importantly how much such pleasure costs.

Why do you need reports at all?

To make any meaningful decisions you need information. In the simplest case, when you are personally directly involved in the process or are physically in the same room, all the necessary information comes to you automatically via audiovisual channels, even if you do not really want it. You know what the junior is doing, how often and on what issues the enikey is pulled, how many times a day and how well the cleaning lady washes the floor, you can see why the code slows down in the profiler right on your machine and you know how many spare mice are in the closet. There is practically no need for formal reporting in these situations. You will immediately see the problem and be able to solve it. This works great in companies where there are several employees and everything is physically in the same room. Everything changes dramatically in the presence of a physical barrier or,when the company grows and you cannot personally take part in all the processes. But the responsibility does not disappear anywhere. Being in another room, you no longer directly see what is happening, but you must make the right decisions in a timely manner. Based on the examples above, you will probably want to regularly receive information like this:

• How many tasks do we have in the queue, what is in development, how many are closed for the period?
• How many calls were made to the helpdesk, what type, how many were successfully closed?
• Cleaning schedule.
• Logs of your application from the user device.

Some of this information will be generated manually (it makes sense to fence the cleaning accounting system only for a large cleaning company). Some are in the form of raw unstructured data (device logs). But the basic information will be required in the form of reports in the form of tables or graphs that are convenient and most clearly and reliably reflect the reality. (Examples of graphs that do not clearly and inaccurately reflect reality are given in this wonderful article ). Only in time to learn that the helpdesk employee is playing the foolis very overloaded, adequate measures can be taken (expanding the staff, paying a bonus for shock work, trying to eliminate the cause of mass appeals). Without knowing this, it is impossible to organize the process normally. This applies not only to managers. Developers, devops also need to know what is happening locally for users, how the network is loaded, servers, the number of requests to the microservice per unit of time, etc. In fact, these are the same reports and without them it is impossible to solve their problems.

Basic requirements for reports

Let's formulate the basic requirements for reports:

1. Visibility and ease of perception.
   
   
   
   There is little sense directly from 5 GB of raw logs, they will need to be processed and aggregated at least before we know something useful.
2. , .
   
   
   
   « 19/20 ». ! , 19 « », 20 . «» , .
3. .
   
   
   
   . IP , % , . .
4. Timeliness.
   
   
   
   It will be unpleasant to know that the channel began to slow down a week ago and half of the users have already dumped, leaving angry reviews. And the problem could be solved by plugging the patch cord in 30 seconds. And we ran out of mice in the warehouse, they didn't buy it on time. A trifle, but also unpleasant.
5. Acceptable cost of creation and use.

Let's dwell on the last point in detail. If for the previous ones I can partly rightly be called "Captain Obvious", then for some reason, for some reason, there is often a complete disregard of reality. Let's tear off the veils!

Everything costs money

Again, a little bit of capital. In our world, practically nothing is free. Buns do not grow on trees, Santa Claus does not exist . Any work requires money, time, and some other resources. If you need a report on the number of earthworms per sq. meter, then we can estimate its value. Let's estimate the costs:

1. Creature. It costs C money. Here we will also include various improvements, modernizations - we will designate C _m
   
   
   
   This can be writing a code that generates a report from the database. Coming up with a form in Excel for manual filling. Teaching a dumb, alternatively gifted employee how to properly drive numbers into a cell and save a file.
2. Regular preparation. It costs P money / day (well, or another period, not the point).
   
   
   
   An employee must take a shovel and dig up a few square meters of earth. Count the detected worms and drive the data into the plate
3. Useful life (depreciation). T days. This also includes how much revision will be used, modernization - T _m .

You can use one report without changes for several years, or you can come up with new indicators and change the display form every day.



In total, this report costs per day

$$

$$/T+ / T+$$

Let's take an example. Let it cost 10,000 galactic credits to develop a report. They will use it for about one and a half Earth years (let's take 500 days for simplicity), then they will come up with something fundamentally new. Also, during this period, a minor revision of 1000 credits will be required. Daily manual data entry (not everything can be obtained from the database automatically) costs 100 credits. In total, we have:

10000/500 + 1000/500 + 100 = 122. If there are 243 working days in a year, that is, it costs us about 30 thousand a year. The formula is generally simple. Those interested can drive the data into a spreadsheet and play with the parameters. If you come up with new columns, require you to change the color of the graph weekly, or the report itself is done manually for half a day, it is easy to get costs in the hundreds of thousands. At this point, it is worth seriously thinking: is this information really that important and is it worth the money?

An example of a very ineffective report from long-term personal experience

Situation. The company has two main non-overlapping directions in one office: wholesale trade in grain (we buy from farmers in the region) and sale of diesel electric generators (the main clients are oil workers in very distant regions). At that time, no one really had mobile phones. Everyone called up on ordinary wired phones. Calls to the region (tariff zone <100 km) were very cheap. And in the Krasnoyarsk Territory, the Far East, etc., it is quite expensive. The company worked with good profit, but the communication costs were quite serious in total. While the main source of costs was clear to everyone, the CEO demanded a breakdown by department. We had no PBX logs. Elektrosvyaz (now Rostelecom) knew how to send information in the form of a paper roll. In total, my colleague took more than a week of working time,to analyze all this and generate the required report. The meeting where they were supposed to look at this report and draw conclusions looked something like this:



General: "Yura (head of generator salesmen), what did yours say so much?"

Yura: "Well, so we are working!"

General: "Oh well, well done!"



That is, in fact, they did not learn anything new and useful at a high price. The situation resembles





For adequate guidance, you need to have information in the form of reports. If the report is inconvenient, you need to think about how to improve it, perhaps add missing information. But at the same time, one must clearly understand that it will cost money. Maybe a lot of money. The correct question should be: "Is getting information X worth the sum of Y?"



Examples:



"Is it worth knowing the color of mice in the warehouse (column in the report) 5000 rubles?" (the answer is rather negative, since this information is required so rarely and of little use that its cost is close to zero)



“Is the knowledge of the complexity of the tasks performed in the last release and planned in the current 50,000 rubles worth? (The answer is rather positive, since we must understand as early as possible that there is a risk of not being on time. In this case, we can find a solution by discussing the problem with the team lead and the Customer or by hiring a freelancer for some tasks).



There is also an important point: information can lose relevance. That is, the recipient is no longer interested in it in whole or in part, but we continue to generate reports. However, the recipient has no particular motivation to cancel them.



If common sense is not enough, at the corporate level, the process can be effectively structured as follows. The recipient of each report must “finance” them within the budget of the unit. That is, if you need reports 1, 2, 3, 4, 5 with a total value of X rubles, the costs in management accounting are distributed accordingly. In this case, willy-nilly, you have to keep yourself within the framework and really require only really relevant information, without a doubt throwing out unnecessary information.