Rating. What is it and how to use it in gamification? The question seems simple, even rhetorical, but in fact such obvious mechanics have many nuances, including those due to human evolution.
This article is the first of my series of articles about components, mechanics and interesting examples of gamification. Therefore, I will give brief definitions to some general terms. What is “gamification (gamification)”? Wikipedia gives a definition: “the use for application software and websites of approaches specific to computer games in non-game processes in order to attract users and consumers, increase their involvement in solving applied problems, using products, services.”
I prefer another option: “gamification is the management of the behavior of users of the system using game mechanics.” The difference between these definitions is that a system can be either a website or software, or a public park or transportation network. Gamification is applicable not only in the IT field. Further, some game mechanics are used to increase user engagement, some are used to attract users, but this is combined into a generalized concept of “behavior management”. To implement gamification, it is important to know what users in the system are doing (can do if the system is not already in use), and what users should be doing from the point of view of the owners of the system. Gamification is useful for moving from “do” to “should do”.
Rating is a simple and popular game mechanic used in gamification. There is no exact definition of the term “game mechanics”, sometimes it means anything from badges and achievements to behavioral impulses. Bringing order to the terminology used in gamification is a topic for a separate article, but here I will limit myself to a brief explanation of what I mean by game mechanics. This is the lowest (most specific) level of designing a gamified system, conventional Lego bricks. Game mechanics are selected and applied when the upper, more abstract levels of system gamification have already been thought of. Therefore, rating, badges, levels are game mechanics, but virality or group work is not.
Rating is a numerical or ordinal indicator that displays the importance or significance of a particular object or phenomenon (definition from Wikipedia). The rating mechanic is related to the score mechanic and often to the user's level mechanic. Rating without points is not possible - the system will not understand in what order to display users in the rating, rating without levels is possible.
Let's try to classify the ratings by their value for the users of the system.
- Competitive - encourages users to rank higher than other users. The most common rating.
- Definition of a losing situation - the system imposes a penalty if the specified number of rating points is not scored. Possible penalty options: transfer to the previous rating group, rank reduction, defeat in the competition, write-off of some amount of game currency, moral penalty (board of shame). Used less often than its win-win counterpart, requires careful thought before implementation and analysis of user behavior, because penalties have too negative an impact on the user and can greatly reduce motivation.
- Determination of a winning situation - gives the right to a reward for reaching a given number of rating points. For the first places in the ranking, for the intermediate stages. As a reward, the same options are used as in the penalties in a losing situation, but with a plus sign. Rewards for milestones in the rankings is an interesting but rare practice that allows the user to lose enthusiasm more slowly in the process of moving from level to level. An example is the rating of the old version of Chefmarket. This is a home delivery service with recipes for self-cooking. Each client has a status displayed in their personal account, points are awarded for cooked dishes, levels are given for points, but to reach the next level, you need to cook dozens of dishes, and this can be demotivating. Gifts for every X points help mitigate the demotivating effect (the number of points depends on the current level of the client).
Chefmarket user rating. Pay attention to how other game mechanics are organically used: badges, progress bar, titles, packed into a nice looking interface. - Status - increases the authority of a user with a high rating in the eyes of other users. Used, for example, in online question projects (StackOverflow, [email protected]). MMR systems (matchmaking ratings) in MOBA games can also be attributed to the status rating.
- Trusted - increases the credibility of a user with a high rating in the eyes of other users. Became the standard for online auctions. Habr user karma is another example of a trust rating. Confidence rating is used in systems based on user interaction with each other, especially if this interaction is offline or involves the exchange of services and goods.
An example of an online auction rating with badges issued when a certain rating level is reached.
The ratings from the list above are combined in various ways within the system. In theory, a competitive rating of users is possible, with intermediate winning situations, with a penalty for rating outsiders and a high level of status and trust for the rating leaders.
Another option for classifying ratings: by who changes the user's rating - only the system, only other users, or the system and users. The option when only the system changes the user's rating is the most common. It is often used in online games. The player performs various actions (kills monsters, completes quests), for which the system awards experience points (ratings). Other users do not affect the player's rating in such a system. The option when the user's rating is changed not by the system, but by other users of the system, is usually used together with the confidence rating. Examples: increasing or decreasing karma, positive and negative feedback after transactions on trading floors. A combined option is also possible, for example, in online questions. For answering a question, the user automatically receives rating points from the system, and if other users recognize the answer as the best, then the user receives additional points.
The next method is based on the positive and negative changes in the user's rating. I single out the conditional “rating plus”, “rating plus or minus positive”, “rating plus or minus negative” and “rating minus”. The first option, "rating plus", implies only an increase in the user's rating. This option is used, for example, for buyers on eBay. As a result of the transaction, the seller leaves the buyer only positive feedback or does not leave it at all. Yes, a fraudulent buyer can be blocked by the administration, but his rating cannot decrease (until he sells badly himself).
A plus or minus positive rating implies both an increase and a decrease in the user's rating, while the rating does not fall below zero. Such a rating will not allow the user to fall too deep in case of unsuccessful actions (and to know the power of an angry Habr). But at the same time, a new user and a user whose rating constantly fluctuates around zero due to systematic “bad” actions will visually look the same, which has a bad effect on the credibility of the entire system.
Rating plus or minus negative means that the user rating can both rise and fall to any values. In practice, there is no point in a large negative rating and it is recommended to enter a threshold negative value in the system, after which it is worth applying punitive measures for such a user, up to account blocking. At the same time, it is important to think about the situation of deliberate “draining” of the rating by other users, to exclude this possibility or make it difficult to implement.
Rating minus is a rarely used mechanic in which the user's initial rating can either remain unchanged or decrease. I don’t remember right away the projects that use such mechanics, but theoretically it is possible. For example, for projects or elimination games, or “last heroes”.
When using the rating mechanics, one should not make an important mistake: the gaps in the number of points scored between users of the system (or between user levels) should not be demotivating, unattainable. This difference is especially demotivating for new users who see that they have zero points, while the leader of the rating has millions. Why is this happening, why would a new user in such a situation think that it is impossible to catch up with the leader? First, the new users of the system have not yet spent enough time to understand the dynamics of scoring. Two or three million points of the leader of the rating may not be so unattainable if the system awards thousands of points for each user action. The problem is that a demotivated new user stops using the system before they realize it. Secondly, the problem is in our natural logarithmic perception of the number series.
We are accustomed to living among the linear ordering of numbers. House numbering, roulettes and rulers, graphs and clocks - everywhere the numbers are located along the number line at equal intervals. It is obvious to us that the difference between 1 and 5 and between 5 and 10 is the same. The same difference between 1 and 500. In fact, the linear ordering of numbers is the result of exposure to our culture, and not an ability from birth. Our distant ancestors, who lived tens of thousands of years ago, did not have a modern mathematical apparatus, and numbers were perceived logarithmically. That is, they were located on the number line more and more closely as they increased. They perceived numbers not in terms of exact values, but in terms of estimates. This was necessary in their way of life. When meeting with enemies, it was necessary to quickly, therefore, approximately, evaluate who was more - our own or others. The choice of which tree to collect fruits from was also made on the basis of a rough estimate. Our ancestors did not calculate exact values. The logarithmic scale also takes into account the laws of perspective and our perception of distance. For example, if we look at a tree 000 meters away and another tree 1 meters behind the first, the second 500 meters appear shorter.
The white player in this picture doesn't need to know the exact number of black pieces to know he's in trouble.
For more information about the logarithmic perception of numbers, about the studies that confirm this theory, and about other interesting facts from the world of mathematics, you can read in the popular science book by Alex Bellos “Alex in the land of numbers. An extraordinary journey into the magical world of mathematics.”
The logarithmic perception of numbers on an intuitive level has been inherited by us. Hidden under the cultural layer, it manifests itself, for example, in the sense of time (in childhood, the years passed slowly, but now they just fly by). We still, despite all education, get lost at very large numbers and instinctively switch to their logarithmic perception. We understand the difference between a liter and two liters of beer, but ten billion and one hundred billion liters of beer seem to us about the same numbers that fit into the concept of “very, very much beer”. Therefore, the problem of feeling out of reach of the rating arises if the gap between the current position and the leader is “very, very many” points. The user's brain will not intuitively analyze the situation, study the dynamics of points accumulation, calculate the time to reach the top of the rating. He will simply issue a verdict - "this is a lot, it's not worth wasting energy."
To avoid the situations described above, it is necessary to apply a floating dynamics of rating points accrual, in which the user receives rewards and gain rating points at the beginning of the intended life cycle of using the system faster than in the middle and end. An example is World of Warcraft and similar MMORPGs with a “European” (not “Korean”) character leveling system. The conditional European leveling system implies a quick passage of the initial levels of the game, followed by a gradual slowdown. The system used in typical Korean (and other Asian) games involves a drastic slowdown in the rate of gaining the last levels of a character.
For example, in Lineage 2, to reach level 74, you need to gain 500 experience, for level 000 - 75, for level 560 - 000, for level 76 there are already much more - 623, and to go from level 000 to the maximum level 77 you will need to gain 1 million experience, while the speed of gaining experience almost does not change (the entire table of experience and levels in Lineage 175 is available at ). Such a slowdown is seen as redundant in gamification, as it demotivates users too much.
Another point worth remembering is that it is easier for a user to abandon a game or a gamified system at the beginning, and more difficult when he has spent a lot of time in the system, after which the user will be sorry to abandon the accumulated points, levels, items. Therefore, give new users a temporary bonus to their points, for example, + 50% for a month. The bonus will serve as an additional incentive to use the system; during the bonus period, the user will appreciate the speed of scoring points, get used to it and are more likely to continue using the system.
An example of a demotivating rating gap error is the Gett Taxi app. Before the last update, there were twenty levels in the loyalty program, the maximum required 6000 points (an average of 20-30 points was given per trip). All twenty levels were evenly distributed on a scale from 0 to 6000, roughly in line with the European leveling system in online games. After the update, three more levels were added to the application, at 10, 000 and 20 points, respectively, which is closer to the Korean system (given that the number of points received per trip has not changed). I don't have a representative sample of app users' opinions about this update, but eighteen of my friends and colleagues who use Gett Taxi noted the demotivating effect of the new rating levels. None of them have received a single new level since the update (more than a year).
The gap between the three new and previous levels in the Gett Taxi loyalty program is unreasonably large and demotivating.
In order to avoid a demotivating gap in the rating, in addition to the global rating, it is necessary to add local ratings to the system, for which the gaps between positions will not be so large.
Possible ways of dividing the global rating into local ones:
- Between friends. Shows a rating consisting only of the user's friends. People like to compete not with an unknown opponent, about which only a nickname is known (such an opponent is not much different from a bot), but with friends and acquaintances.
- By time. Rating gained for a certain amount of time (day, week, month, year). Good at resetting and replayability. I didn’t manage to win this week - I’ll try next, and the gap between users is regularly reset to zero and does not grow to cosmic values.
- By geotargeting. A rating that shows only users from a certain area (district, city, country, continent). Just in such a situation, Gaius Julius Caesar said, passing by a poor barbarian town: "it is better to be the first here than the second in Rome."
- By gender. Then compare the results of men and women, playing on hype feminist and chauvinistic motives (use carefully, there may be flows of hatred and feces from both sides).
- By age groups. For example, in the gamification of near-sports systems and systems that require skills that change in a person with age. For example, projects that motivate people to go in for sports, allow you to upload your results and see the results of other users. It is clear that it will be much more difficult for a 65-year-old person to run as much as a twenty-year-old has run, and it will be much more interesting to compete with their peers. On the other hand, an example is online chess and other complex intellectual games in which an experienced grandmaster will be unattainable for a fourteen-year-old teenager.
- According to other data about users that are available in the system (rating only for Mercedes drivers, only for plumbers, only for the legal department, only for elves of level 120).
As you like, combine the methods listed above with each other, feel free to experiment with them.
During the operation of a gamified system, keep track of how the rating meets the goals set during the design. For example, if the purpose of a rating was to increase other users' trust in users with high ratings, pay attention to finding and limiting possible honest and dishonest ways to quickly increase the rating. The basis of a trust rating is the difficulty of obtaining it and the possibility of losing it very quickly. If there are loopholes in the system for an unreasonably rapid increase in rating, users' confidence in it will drop sharply. For example, if an online auction has the ability to increase the seller's rating for each transaction made with each user, then two users can maintain their rating at a high level simply by buying penny goods (ideally digital) from each other. At the same time, possible negative reviews about low-quality service or fraud will be clogged with a mass of fake positive reviews, as a result, there is a risk of a massive loss of confidence in the system.
To wrap up, here are three more tips for using ratings and levels:
- Do not show the user the number of points required for subsequent levels. This is demoralizing for beginners who do not yet know the speed of scoring in the system and the possibilities for scoring. When a user sees that the first level is reached for 10 points, the second for 20, and the twentieth one is already for a hundred thousand, this is demotivating. One hundred thousand seems like an unattainable number.
- The number of points required to the next level, show taking into account the scores. The user scored 10 points, moved to the second level, and 20 points remained before reaching the third level. Don't show the user's progress as 0 out of 20, better show it as 10 out of 30. Create the illusion of an unfinished task, our brain does not like unfinished tasks and tends to finish them. This is how the mechanics of progress bars work, this principle is also relevant in our case. Logarithmic thinking is also useful here. When we see that we have gained 450 out of 500 experience points, it seems to us that this task is almost completed.
- Remind the user about successes in various ratings of the system (after all, the user himself may not even guess that this week he is in the top three among men in his area).
In this article, I do not pretend to be a comprehensive analysis of possible options for using the rating mechanics, so I probably did not mention some cases and use cases. If you have an interesting experience using ratings in games and gamified systems, please share it with me and other readers.
Source: habr.com