Footcloths

Programmers love to draw portraits. If you need a report on sales - they will dump the entire sales table, with contractors, nomenclature, organizations, contracts, amounts and quantities.

All anything, only with the help of such a report is difficult to manage. Analyze - you can, if there is a bunch of free time. And who has a lot of free time? The analyst has, for example. Well, if he is an analyst by position. There is in fact the vocation of the soul of analytics. His position, for example, is a sales manager, but he doesn’t want to sell or is not able to, but in numbers it’s nice to pick one.

The head of time for picking in the report, alas, no. At least in the framework of regular management. He needs a short, capacious information that answers a simple question: how are things going? Or in another way: are we all good?
')
How to answer this question with the help of footcloth? Yes, nothing. The footcloth, as it were, says to the manager: do you want information? Well, here it is. ALL! Come on, sort it out, and look for the answer to your question.

If the manager digs up the programmer with his task, he tries to "help." For example, it makes a sales plan / fact report. Beautiful, comfortable, customizable. There you can not just compare the plan with the fact, but also the fact with the fact for different periods, and in general - any number of tables. What is the result? Another footcloth, only more complex.

If you get to the bottom of the programmer again, he will say: damn it, take a mood and report to yourself, and save the setting. Filter the data, select a period so that only the data you need is displayed. Well, as an option, come down. But only if the manager is able to customize. And if you can not?

Ok, the programmer, reluctantly and with crushing bones, will set up the report himself. And he will say: everything, on, choke. Aloud will not say, of course, but think. All happiness came?
Not. First, the first time it does not set up - will have to run several times. Or write a good, high-quality teh.zadanie. And it is almost always impossible, because the manager on the bank does not know in what form he needs a report.

As long as the manager has a foot wrapper, it seems to him that it is enough to set up the groups, and he will be able to manage it using a report. Having received groupings, he will want the filter. Having received the filter, he will understand that sorting is necessary. And so on.

Secondly, one report is not enough. The manager, after all, is not only interested in sales? Even the movement of money needed. And the work of the staff. And overdue orders - both buyers and suppliers. There will be several reports, and for each - one or more settings.

Thirdly, in most cases the information system is desktop. To find out how things are going, you need to turn on the computer. Even before it is necessary to reach. If a manager is such that he is sitting at computers all day, then there is no problem. And a lot of these leaders?

The head will want to watch reports remotely, via the Internet. It is desirable - from the smartphone. It's so convenient - even at lunch, even in traffic, at least in a boring meeting. The programmer will have to try again.

Fourth, there were several reports. Everyone answers their own question. Everyone needs to open and watch separately. One opened, the second closes. Maximum - made a split-screen and saw two at once.

In desktop systems, they can not even be looked through, like on a smartphone - well, so that all reports fall out into one long footcloth, and you can run your fingers from top to bottom.
The manager will want to see all the reports on one screen. This is a normal human desire. But for a programmer, this is hell.

A regular large report, even tuned and filtered, looks good only on full screen. If you narrow it in half, scroll bars immediately appear that spoil the whole impression - even if there is nothing behind the visible area of ​​the screen, the hand itself stretches down or sideways to make sure of it.

And if there are three reports? Or four? It turns out ass-s. There is nothing left for the manager, except for the traditional viewing of reports - one by one, in full screen.

Even the presentation of information in the form of a diagram does not save, if it is displayed, as in 1C, in a tabular document. The same scroll bars appear, the scale of the chart cannot be reasonably controlled, emotions are heating up.

And what's the problem? And is she there?

Yes, and very simple: the programmer loves footcloths too much. He was taught that way, such a system of values ​​was set: they put everything in the report and more. And provide a smart setting mechanism.

This approach is good for developers of boxed solutions, because they have no choice. They do not know what information and in what form the particular manager will want to see, therefore they make footwear + customization themselves.

But field programmers, i.e. for specific customers, for some reason, blindly follow this paradigm. They behave as if they themselves are the developers of boxed solutions. And they make all new footcloths.

Friends programmers. I am also a programmer. I also love to make footcloths. It's simple, interesting, even beautiful in its own way. But footcloths are not suitable for control.

I am not saying that something is wrong with you or with us. I just propose to solve a new, interesting, engineering task: to draw a small report.

Not a big footcloth of a thousand lines, but a small sign, with two columns and three lines. Not a hefty chart that doesn’t fit on the screen, but a tiny one, with three bars that clearly answer a specific question. Do not throw out a bunch of numbers so that the person on their basis answers the question, but to give an immediate answer to the question is short, concise, understandable.

This is incredibly interesting. And this is all easy to algorithmize. The principle is very simple - the funnel "what if?".

Here we have a plan and the fact of sales - separate tables with data. We draw two reports - the plan and the fact of sales.

But what if you combine them, and immediately draw both the plan and the fact in one report? Well, not bad already.
And what if to turn to groups of the nomenclature? Then the report will be small, at least in height.
OK, we do.

And what if the same report to draw in the form of a diagram? The usual two-bar histogram. The first will be the plan, the second fact. You look, and you immediately understand where the plan is being carried out, and where the failure goes.

And what if recalculate the plan for the current day? After all, we threw out the entire monthly plan in the first column, and the fact will catch up only by the end of the month. During the whole month, looking at the diagram, the manager will think that everything is bad. Or he will have to figure out in his mind what part of the month has already passed. No, let's not derive the entire monthly plan, but part of it. The tenth is one third, the fifteenth is half, and so on. Then the diagram will become clearer and more interesting.

But what if we color the bars depending on whether everything is good or not? After all, we know if the plan is well executed, or not - we need only two numbers to compare. Make the bar red if everything is bad. Yellow - if dangerous. Green - if good.

And what if you do not display the bars for which everything is fine? For example, if we have a diagram in the context of groups of nomenclature. We will display only those that are yellow or red. Well, the head say that if there is no column, then everything is fine. He will understand? Does he need information to react and make decisions? Tasks to put, accents, on the carpet call someone. So let the system speak only about what to react to.

And what if I remove the diagram at all? We almost all the posts have already been hidden, what is the meaning of the diagram, how is the way of presenting information then? We'll print a short table, in which there will be only the problem groups of the nomenclature - yellow and red. Well, the numbers, the percentage of the plan.

But what if you don’t display a report at all? Why make him stare every day? Let's do this: if there is a problem, we will give him a signal. Not us, but the system we are developing. There was a yellow or red color - we write a letter, or in a vibe, or SMS, it does not matter. The head will know: if there are no signals, everything is fine. You can not waste time on the analysis, it is better to just work, something useful to do. Of course, the first time, he will still go in and watch - he doesn’t trust programmers. You never know, suddenly sending letters to the stake rose. Well, let - poke, and get used to it.

And what if he does not need information on some nomenclature groups? Well, he knows that they are not for sale, and will always be either red or yellow. Why should a person be jerked with unnecessary signals? We remove.

But what if we are not reporting a list of bad groups, but one line - is everything good or is everything bad? Knowing the importance of specific nomenclature groups, and the overall sales plan, we introduce weights and a simple function that will calculate the answer to the question: are we all right? Uninteresting groups will give a coefficient of 0.1, interesting - 0.5 or 0.7, and that's all, it turns out we have one big, beautiful figure. If the main group sells well, then we are fine.

And what if you give the head two different numbers? After all, it is clear that tasks for different groups can be fundamentally different. For example, the main group is a stupidly shaft of sales. But a group with a new nomenclature, for example, a new equipment, does not give a shaft and cannot give it by definition. There, one or two sales per month is considered a boon. Do you understand? Not even the amount is important, but the quantity. Sold a new piece of equipment - it's a holiday! If you work only with amounts, then this unit of equipment will always hide among the main nomenclature. Why do we need this? Let there be two numbers: how things are going on the main line, and how things are going on the new equipment. Different evaluation criteria, different scale.

But what if we still put these two numbers in one? We rewrite our function with weights, replacing the absolute numbers with the relative ones. Let the main nomenclature be considered by the sum, and the new one - by quantity. And all this is converted into percentages. Then the weights can be corrected - to increase in the new nomenclature, since it is important for us.

And what if two factors are taken into account: the plan / fact at a particular point, and the dynamics of the last days? It is as it happens. On the first day of the month we made a large shipment - right away 20% of the monthly plan. And all our charts for six days will show that all is well. But in fact, our sales were staked - not a single shipment for almost a week. From a formal point of view, everything is not bad, but are we not here for formalities? The head should know that managers kick in the noodle without holding the sales dynamic. There is nothing complicated for us - we simply add another variable to the function, the dynamics of sales, with its own weighting factor. And that's beauty.

But what if we in our function take into account not only the plan / fact of sales, but also the plan / fact of money movement? We will derive the head an infinitely beautiful number, or a human-readable line - “everything is good”, “sales are good, money is so-so”, “sales are 120%, money is 90%”, etc.

What if…

So you can continue indefinitely. The main thing is to stop in time, and give the manager to use the tool, get used to it, formulate feedback. I, as a programmer, understand that when solving problems of the “what if?” Funnel, an irrepressible desire arises to make it easier, better, more informative, more useful. We must be able to stop yourself.

The main thing is to perceive this task as an engineering one. Because she is. This is not beautiful or bows, it is - building a control system.

The leader himself with such a task can not cope, alas. He does not know all the possibilities, all the data, all the relationships. And the programmer knows. But silent.

Let's get out of your dark corner. You, programmers, can bring an incredible benefit to company management. You are engineers. Control system, management automation is an engineering job. And the manager will be its user. It will manage with your system.