How disputes grow into projects

The idea of ​​the project began with a colleague and I argued that if we collect feedback from car owners on tires and compare them with the test results of magazines published in Russia, such as “Auto Review” or “Behind the Wheel” and others, then the results will not match .

Moreover, this will only happen if we take reviews in which there will be relatively objective information, that is, do not take statistics of the type “very bad tires” or “I like everything” into the statistics.

To test this hypothesis, we collected reviews that mention two tire models and make comparisons. For comparison, we chose several criteria that can be generally distinguished from the reviews and which are understandable to everyone (controllability, aquaplaning, behavior on ice, etc.). So we got the initial statistics from which our project grew as a resulttyresaddict.ru

To begin with a small update on the subject of Habr

The fact is that reviews about tires are usually poorly structured and there is not much useful information in its pure form when recalling a _one_ model. Adding ratings to reviews for the same _one_ model also doesn’t show really important differences between them.

The method of leaving reviews in which two models are compared, in our opinion, turned out to be more relevant in a particular niche. This has been developed in the site’s interfaces.

What happened as a result

After the collection, we got a base with quite interesting results. For example, on the existing data it turned out that the model that was consistently in the top of our magazines (Nokian Hakkapeliitta 7) turned out to be very noisy and, according to some car owners, had problems with handling, and according to the results of the top of winter tires, for some reason, the non-studded Hakka R first came forward .

that is why, having not exactly the results we expected, we decided to continue our business and turn it into a project. There are actually two reasons: the number of reviews is enough to get a preliminary analysis, but not enough for full-fledged large-scale statistics, and it would also be foolish to just drop what we already received.

What happened

Since Habr is a technological, and not a car site, I’ll tell you exactly what technological things we did at home:

- Firstly, the reviews, as it turned out, are very subjective, so we decided to give users only comparative reviews. And since most resources, on which parameters are indicated on a scale of 1-10, people usually give ratings of 1 or 10 :) and the result is a mess, we abandoned this in favor of simple ratings worse-better.



- Secondly, since we have data on the use of tires on different machines, then we have the opportunity to build application analytics and tops for different brands of cars.

- We decided to add tests to our site, but unlike many resources we made it possible to compare tests of different magazines and resources and different years of release for a particular bus model. By the way, for example, it turned out that such tires as the Pirelli Winter Carving Edge got similar points in ZaRule tests exclusively chaotically. Which in my opinion is very strange.



- In order to monetize, we have invited and continue to invite tire shops to us so that we can not only compare tires but also see offers on them. Now we already have on the site about 10 thousand positions from stores from Moscow and St. Petersburg, but this is far from the limit.

“Well, finally, we decided to enable users to copy tire tread images to forums and blogs.” It turned out that this banal "feature" is needed by very many.

Criticism?

Reasonable criticism is welcome, it is very interesting for us to hear an opinion from you about our methodology in the first place, of course, comments on the site and of course go and use the resource.

PS We have connected reformal for suggestions on improving the project and reporting problems.