Reflections on the topic of user rating

In many cases, to evaluate something on the Internet, we are used to clicking "like" or "dislike". The most obvious example is social networks. But sometimes the separation of likes and dislikes is simply not enough for a more accurate rating system.

Take, for example, the task of evaluating users of a new restaurant that has recently appeared in the city. Such a problem is usually solved by a “school” scale of 5 points. These can be asterisks or more and more often emerging emoticons with various emotions. It is easy, fast and convenient. But, as usual, does not do without "but". Putting your rating, the user describes the overall impression of the restaurant. This is an extensive assessment, which implies a combination of everything: kitchen, interior, music, service, etc. And what, in essence, does the first and fifth stars in this rating mean? The first is worse than in the cheapest tavern, and the fifth is the level of the best European establishments? No one answers these questions, and this rating adapts to the user's perception, which, by the way, is a very interesting phenomenon (but let's not talk about it). This is how I imagine what the stars actually mean:

For me, the scale of 5 points narrows to 2.5. We have to make extra efforts to be objective. Conclusion: like the likes, such a system copes with its tasks, but it is not suitable for detailed “debriefing”.

Facts about man

It is necessary to raise the question of objective assessments in advance. Public morality teaches us to soften our opinion. It is easier to say “Your restaurant is not as good as it could be” than “This is the worst place in the whole universe.” Even though the web is not one-on-one communication, we are faced with the same rules as in the real world. Some deliberately overestimate the rating so as not to offend. Well, it happens that they underestimate.
')

Where problems begin

I will try to formulate a new task. Suppose that the management of a new restaurant wants to improve the quality of service and understand what visitors like or dislike in their establishment. They want honest feedback to improve or change.

A reasonable solution in this case would be to divide the assessment into parameters: kitchen, interior, service, atmosphere. This will provide an opportunity to understand in which area the most problems. So let's do it, but here we are in for a “fork”.

Option 1. I like how it was

We do not want to go away from the usual, well, okay. So much better already. This we have often seen, for example in Skype when assessing the quality of communication. But we again give the user to reflect on the meaning of the stars. And it works this way: every time, moving to the next assessment, we subconsciously create our own scale with our own values ​​(bad - good, tasty - tasteless, beautiful - ugly). This causes additional difficulties when the estimated parameters belong to different spheres (music and service). And if there are many such parameters, the difficulties turn into user discontent.

Option 2. More specifically

In this form, it does not matter what is used: sliders or asterisks. The bottom line is that the extreme values ​​are the pillar of the scale. It helps to understand what the worst and best score means. Now our 1 and 5 start to work when the user agrees with the extreme values. He does not waste time creating his own rating scale, and this is a plus. However, the success of such a rating is directly dependent on how accurately the extreme values ​​are suitable for determining the evaluated entities, which is sometimes difficult to imagine. So in one restaurant the goal of the chef may be the refinement of dishes and their presentation, and in the other - the taste and price.

Option 3. Tag cloud?

A rather unusual decision. We again came to the conclusion that instead of four parameters, we have one, but a composite one. Such a decision is somewhat reminiscent of a psychological test, the result of which will be the answers to the questions that a person can ask himself. With a cursory reading of the keywords, the user will stop at what he considers most appropriate. At the same time, we do not show exactly what grade each word carries. This helps to avoid user doubts when making categorical ratings (very bad and very good). In theory (I admit honestly, in practice I have never met such a person) such a rating system can give very plausible results. Another plus is imitating more “human” user interaction with the site. And the same "human" characteristic as a result. No need to scroll through the comments to understand what people think about the place. According to this characteristic, restaurant owners will be able to understand what users do not like and whether their positioning corresponds to reality.

The disadvantages of this decision can be easily attributed to what you need to read, that it can be cumbersome and unusual. And of course, that such rating needs to be tested. However, the concept seems to me very much like the right path.

Conclusion

Each assessment system has its own tasks and scope. The choice of an infinite number of possible solutions depends only on your goals. But still the question of a convenient and versatile tool for forming a rating of a large number of heterogeneous parameters for me (I think not for me alone) is open. And it is interesting enough to express your thoughts.