An example of the impact of the mass effect on the number of downloads
In the footsteps of the previous article , which was a little over a month ago, about the number of downloads of user scripts for the site habr.ru, another interesting phenomenon appeared that deserves consideration. If we assume that there is no cheating, then it can be explained only by the mass effect - the influence of a large number - the download rate on an even faster growth of this number.
Surprising thing - the script has not been updated. The article could not affect the increase in the number of downloads, because it says that this script does not work, it is outdated and is not supported by the site since about October last year. On the contrary, the new script , which the author of Inversion wrote to replace him, is efficient, and they should be used (except for Chrome). But the number of downloads (330) increased slightly, by about 200 units. The old script, judging by the indicators, for the same time downloaded 650 times.
(It’s time to make a reservation that the calculation on the website of the last few months is not going well - for one download 2 units are added to the counter, but for simplicity we will be considered traditional.)
So, 200 downloads of the correct new script versus 650 downloads of the old idle. What ideas can be, why so?
')
In addition, I will say that right away, 2-3 days after the article, there was such a picture that the new script was downloaded immediately 70-90 times, and the old one was downloaded very little, about 30-40 (I don’t remember the exact values). That is, the effect of the article was observed in the first days after its publication. This means that subsequently the difference in downloads is even more noticeable - 120-150 downloads of the correct script versus 600 are incorrect.
The only assumption can be made like this: it affects the amount of downloads. Habr's user comes to the scripts site and searches for a list of scripts for his site. On the first page looks descriptions, the number of downloads and dates. From the fact that all other scripts do not grow fast, at about 50-100 downloads per month, and scripts with large numbers of downloads - faster, we can conclude that the user focuses on this number, although it is disorienting. Probably not everyone pays attention to the fact that scripts written before the last major site updates will not work. And the latest updates were in October 2011 and February 2012 (or March).
It is possible that the considerable simplicity of the most popular script makes it easy to fix it, so some of the users are able to see the results of the work even of an outdated script. This can only be prompted by the users themselves in the comments - what did they do with the 17 most popular “Highlight Topic Author Comments” downloaded most popular script? Did you fix it or just ignore it?
In the meantime, as a fact, we have the “magic of large numbers”: a non-working script leader is 5 times more popular than the working correct replacement script from the author Inversion . As far as I remember from the numbers of another popular script, from crea7or , “HabraFormatHelper”, he scored downloads of about 300, if not more, that is, it also has a mass effect (or a snowball effect).
The conclusions can be drawn as follows: if you want unfair competition for your product - become the leaders according to a popular and understandable indicator, the “snowball” effect will be provided. And if you are in the ranking of equal services, programs, products, you want to observe objectivity - on the contrary, do not submit deceptive, but very speaking indicators. In this case, the merit of the script over the past 4 years cannot say anything about its quality over the past 3 months, but the table of search results on the site cannot say anything about it. The situation could be corrected by feedback from visitors on the site, and there is everything for feedback - reviews, comments and a forum, but for some reason, visitors use them very little. From the available reviews, it is even more difficult to compile an opinion on the quality of scripts than on speaking indicators. In general, by any means, for objective presentation of a product in any lists and ratings, objective information about quality is needed, otherwise we see such paradoxes of the mass effect, as in this modest and illustrative example.
Surprising thing - the script has not been updated. The article could not affect the increase in the number of downloads, because it says that this script does not work, it is outdated and is not supported by the site since about October last year. On the contrary, the new script , which the author of Inversion wrote to replace him, is efficient, and they should be used (except for Chrome). But the number of downloads (330) increased slightly, by about 200 units. The old script, judging by the indicators, for the same time downloaded 650 times.
(It’s time to make a reservation that the calculation on the website of the last few months is not going well - for one download 2 units are added to the counter, but for simplicity we will be considered traditional.)
So, 200 downloads of the correct new script versus 650 downloads of the old idle. What ideas can be, why so?
')
In addition, I will say that right away, 2-3 days after the article, there was such a picture that the new script was downloaded immediately 70-90 times, and the old one was downloaded very little, about 30-40 (I don’t remember the exact values). That is, the effect of the article was observed in the first days after its publication. This means that subsequently the difference in downloads is even more noticeable - 120-150 downloads of the correct script versus 600 are incorrect.
The only assumption can be made like this: it affects the amount of downloads. Habr's user comes to the scripts site and searches for a list of scripts for his site. On the first page looks descriptions, the number of downloads and dates. From the fact that all other scripts do not grow fast, at about 50-100 downloads per month, and scripts with large numbers of downloads - faster, we can conclude that the user focuses on this number, although it is disorienting. Probably not everyone pays attention to the fact that scripts written before the last major site updates will not work. And the latest updates were in October 2011 and February 2012 (or March).
It is possible that the considerable simplicity of the most popular script makes it easy to fix it, so some of the users are able to see the results of the work even of an outdated script. This can only be prompted by the users themselves in the comments - what did they do with the 17 most popular “Highlight Topic Author Comments” downloaded most popular script? Did you fix it or just ignore it?
In the meantime, as a fact, we have the “magic of large numbers”: a non-working script leader is 5 times more popular than the working correct replacement script from the author Inversion . As far as I remember from the numbers of another popular script, from crea7or , “HabraFormatHelper”, he scored downloads of about 300, if not more, that is, it also has a mass effect (or a snowball effect).
The conclusions can be drawn as follows: if you want unfair competition for your product - become the leaders according to a popular and understandable indicator, the “snowball” effect will be provided. And if you are in the ranking of equal services, programs, products, you want to observe objectivity - on the contrary, do not submit deceptive, but very speaking indicators. In this case, the merit of the script over the past 4 years cannot say anything about its quality over the past 3 months, but the table of search results on the site cannot say anything about it. The situation could be corrected by feedback from visitors on the site, and there is everything for feedback - reviews, comments and a forum, but for some reason, visitors use them very little. From the available reviews, it is even more difficult to compile an opinion on the quality of scripts than on speaking indicators. In general, by any means, for objective presentation of a product in any lists and ratings, objective information about quality is needed, otherwise we see such paradoxes of the mass effect, as in this modest and illustrative example.
Source: https://habr.com/ru/post/144112/
All Articles