Problems of volumetric test tasks when choosing a job

Looking at the history of my test tasks, I see patterns that I would like to warn colleagues about, because such cases occur regularly and, most likely, regardless of the specialization of the developments. For example, I came to such conclusions, having on my account more than a dozen tasks performed in different years with a volume of 2-5 working days each. And the conclusions are so paradoxical that, I think, they will cause controversy and surprise among those who did not succeed. I will formulate the main conclusion for a start, and then I will show the rationale by examples from practice.

Conclusion : (sophism) a test task with a volume of more than a day with the main goal (getting into work) is unprofitable.

Sometimes in the job market (5-10% of cases) there are offers from companies to do a “home” test task. Usually it takes at least several hours: they try not to give simple tasks. It turns out that it is of great importance for the applicant when, in what phase of communication this task is offered and how much it is. It does not always mean that the task is reasonable or profitable to complete. Very often the tasks contain flaws in the statement, and their implementation is less profitable than the strategy of an inseparable job search among companies that do not offer tests or offer less complex and faster tests (for example, directly at an interview or via Skype online).

In the course of several years of work on the web frontend, there were periodic tasks that were not always profitable to perform. I will even say this: not a single case of hiring with a test task was successful for any side - neither for me nor for the employer. At the same time, recruitment, built on slightly different principles, was successful. In total, I’m used to seeing tasks that are not the goals that, in theory, the other side sees in them, the employer. In fact, they do not fulfill their primary function. The main function of tasks (volume of 8 hours or more) in most cases does not work.

1) do not guarantee the justification of the time spent even in the case of error-free execution;
2) create disproportionately large efforts towards this vacancy;
3) may show irresponsibility from the position of the employer: failure to inform HR about the test results.

And since the number of job vacancies is no more than 10% of the total, even the strategy of not fulfilling them makes sense! However, I do not call for this: sometimes it is necessary to carry out tasks out of “love of art”, but at the same time (paradox!) One can not count on a successful reception, that is, they should not be done for the sake of hiring.

Decoding of sophism

Let's look at a couple of such beautiful examples that have been carried out in different years and already with an understanding of this paradox. Those. I carried out these specific tasks, not for employment, but for the 2nd or 3rd goals: to create a portfolio for myself and to write useful utilities for myself. Why experience in completing assignments led to such conclusions - let's talk in the review process.

Experience is limited to the development field for Javascript / CSS / HTML, but I think that the conclusions will be true for other areas of development and programming.

As in any case, the use of test tasks is a great art, and not everyone owns it. I want to illustrate a couple of striking examples from practice when tasks, sometimes even potentially paid, did not lead to a positive result, although they themselves did not detect the abilities or inability to work.

The paradox of long tasks was noticed a long time ago: all the vacancies where the tasks were were not mine, regardless of the success of their implementation (or rather not, there was one company, but I spent 4 times more time on the task there than I needed, because CSS is the concept of browser compatibility, and many tests are needed for a quality result). In the end, I think, not because I performed them relatively poorly, but someone “successful” performed each time the next task “good”. The reasons are deeper - in psychology, in the system, in the art of writing tasks.

Sometime in 2011, at one of the interviews, we did not agree with the employer about the rate (after the interview, during the discussion of the issue of salary), and in the presence of HR from a third-party organization, he made an unexpected decision: to formulate a test task and pay it. Volume - for about 3 days. They wrote a contract, I started to do, even did. It turned out such a beautiful widget with graphs, exactly according to the task and a screenshot of the sample. But, as it is not surprising, pitfalls of psychology formed here.

Task - graph widget

Link to the demo.
The task formulation required support for various browsers. And the rounded corners of a beautiful widget weren’t supported, as you know, by IE6-8 browsers. But the PIE library could not be used - it heavily loads the engine.

On this and played a cunning employer. More precisely, not even he, but the worker who supervised me - I don’t know why he personally needed to turn the contract into obviously disadvantageous. They recalled browser compatibility and the line in the contract “The widget should work the same in all browsers.” It formally meant that to support the rounded corners of IE6-8 (2011, he was still alive!) I need to increase the amount of work by about 40%, to cut images in all rounded corners and to do another 2 days of obviously rough work, which certainly no one needed, only the widget’s consumer (you can’t use PIE). She would not show anything from the skills, and assumed that the customer would really need the widget, and he did not want to transfer this part of the work to the regular job (after receiving).

Let us recall the moment that the final amount of the offer was still unknown, therefore, even having done the work, I might not have received a satisfactory offer. The solution came about by itself. Even having promised to finish the work, I could not find any reason to finish it. After 2 days, a job offer appeared from a large company after 1 interview and without test tasks, and on the 3rd day I worked in it. Finishing an officially completed test task, losing almost twice in his salary level, there was no point. At the same time, the customer received a free part of the test task and could be satisfied with it, but I received a test page, which I can show as an example from my work. Let this be the time spent, but according to its results there were:

1) a designed test page with 3 instances of widgets and settings;
2) experience in applying inheritance of DOM objects and HTML template engine;
3) use of the library for graphing;
4) an example of working with JSON and XML as input.

The potential reward for the test task turned out to have no value at all, for the sake of which the task should be performed. Thus, the task did not fulfill such functions:

1) it was not materially interested in performing it;
2) the interpretation of the contract by the customer transferred it to the category of financially disadvantageous;
3) distracted from the main task - search and device to work;
4) performed in conditions of uncertainty of the conditions of the subsequent offer (offer).

The fact that the work turned out to be personally useful to me was the result of the fact that I tried to solve this problem not for the sole purpose, but with the goal of reusing the written code. If you set a double goal for a test task, you can often agree to its free execution.

After this incident, there were at least a dozen test tasks of various kinds in subsequent years. As before, it was not necessary to take a job with the passage of the assignment - there were no more than 10% of those who wanted to give an assignment. For a lazy job seeker, you can even cynically advise never to complete tasks - and in this he will not lose much. But, nevertheless, they sometimes play a useful organizing role, as an open source in the world of commercial software.

If the tasks are aimed at the self-education of the applicant and solve a problem useful to the developer, then to perform such a task can sometimes be considered as a “matter of honor” and as a path of personal development. It is possible to carry out the task without hesitation, if time permits, and the task comes from the environment of an authoritative organization, which will be useful to talk with representatives of in any case. So it was in 2013 with mail.ru.

Tasks - Text Filters

Link to demo 1.
Link to demo 2 (highlighting words).

Before the first meeting, they sent a beautifully formulated test task of 2 tasks, the solution time of which seemed to be about a working day or a little longer. It was immediately clear what its second use could be. In addition to an example of your code for a not entirely trivial task (two), you would get an example of a fast JS filter over large text data arrays. To illustrate the speed of work, I generated a random set of words, about 100 thousand, and then words were searched for in it according to the sample of the word entered by the user. In the first task they were highlighted, in the second - only lines with the found fragments were shown. Out of 1000 lines, 5, 2, or not at all, zero lines instantly remained on the screen.

The main practical use of the task also strangely vanishes. Yes, I was invited after this for an interview, it was successful, everyone liked everything, but it turned out that mail.ru had several interviews, and the next day I was offered a job in another company, after an absentee interview on Skype, and 2 times closer to home, and with definitely no worse conditions. Again, it was not possible to practically realize the ephemeral potential of the test tasks, again they remained to fulfill their non-core goal in the future.

Unreasonable Tasks

In the end, I would like to recall such tasks (test tasks), a short analysis of which (20-30 minutes) led to the conclusion that it is absolutely unprofitable to perform them. A letter to the employer was usually written about this, justifying why the implementation of such a task would require an unreasonably long time and would give me as an applicant an unreasonably low chance of choice. Another 30 minutes of writing is spent consciously on this, but we get an almost correct separation, except for the fact that the employer has to be put in an awkward position, justifying that he is wrong. Usually, employers either canceled the task without new ideas, or did not respond.

Why is it unreasonable to carry out such tasks? Usually, these are tasks that require more than 2 days to solve (more than 16 working hours), for which there is no time or I do not know anything about the employer. That is, he could not even talk about the paycheck, about working conditions, but immediately give the task, which is often called "small", but "not for 5 minutes." Already by such phrases you can understand that we are talking about 2-5 days of leisurely work. Very often, such tasks are accompanied by a fuzzy statement, and you can practice your leisure time in your search for uncertainties in the statement of the problem. For example, nothing is said about the behavior of a web form in the case of boundary conditions (there is zero or a lot of data), the conditions for validating some cases of filling out forms are not specified, you need to “use Angular” there, where it does not need to be used in terms of meaning (because the page is reloading). In some of these puzzles, it is guessed that the master himself never did it.

In such cases, it is necessary to recognize the “rot” (dampness) of the task in time and “not to buy low-quality goods”. At the same time, a diplomatic, polite approach helps to maintain the right relationship with the employer, but, most likely, this also indicates a poorly thought-out campaign to find candidates, and it’s impossible and pointless to teach the employer on the other side of the vacancy.

There are two practical cases of such ill-considered assignments over the past year; in both cases, before the interview they offered to “work” for about a week, it is not known for what purpose. More precisely, in order to obtain the right to speak further with the employer.

Here, of course, there is no need to explain that 95% of other employers do not pose such obstacles, so saving a week or two on rejecting such offers will help to save a lot in living expenses. And it is more profitable to spend the free time on your own projects or on a normal job search and other self-promotion of yourself in a resume.

What tasks would be welcome

You can understand employers who want to see live examples of job applicants before they recruit them. As an option for a mutually acceptable dialogue, there is one such as completing tasks online via Skype / Jabber, with immediate comment on the train of thought.

In such tasks, a huge plus is that the applicant sees that the interviewee does not care how much he spends time. The interviewer himself will not waste his time if in the course of work the insufficient abilities of the applicant become clear. Tasks can be solved expeditiously, conditionally, or replace one another. They talk about 2-hour telephone interviews, but in practice I have come across such a 20-minute interview, which then continued at an in-person meeting.

Another option for an acceptable task may be if it involves a waste of time of the order of 2-5 hours. Maybe with an immediate phone call at the end. To create such tasks, you need a certain skill in choosing the balance of the complexity of the solution. You cannot make the task too difficult. You can’t come up with tasks with an unobvious solution - you need to even suggest a solution, because the main goal is to check the knowledge, and not waste the time of the applicant. I myself tried to create examples of such tasks, being in the place of the employer, on the other side of the vacancy. What happens? Even simply assigned tasks, not everyone is capable or willing to perform. And here there is a qualification of the time that the applicant is willing to spend, and, possibly, unclearly formulated conditions for completing this task.

Recall that it is very important for the applicant to know on what condition he will agree to spend his time on the task. It makes no sense to give a task before clarifying the conditions of employment, except, perhaps, the simplest tests for knowledge of arithmetic.

As a result, we can say that test tasks for the applicant are not a very safe selection tool. It is very easy for them to do harm in a hiring campaign, acting without taking into account the psychology of applicants and indiscriminately. This is the same as if putting an unjustified price on a product that has not yet been advertised. Such a vacancy will simply not be bought.

Tips for applicants:
*) Having heard about the test task, try to evaluate its scope of implementation on the basis of the statement of the problem, evaluate errors and uncertainties in its formulation in order to clarify them before execution;
*) evaluate whether it will be profitable for you to spend time on implementation, weighing the probability of a successful reception; keep in mind that promises to accept such salary in case of successful implementation, most likely, will not be;
*) evaluate whether it is profitable for you to do just this task right now, in the process of looking for work and other interviews, and whether you will end up with a product that is useful to you (in a portfolio, in an open-source project);
*) check with the employer if he will object to the publication of a solution to his problem by you for your own purposes.
*) if for all questions - the answer is “yes” - proceed to the test task!