What will be the help systems in the future?
I am the creator of the business directory YPAG.RU , I would like to highlight the industry, tell them the advantages and disadvantages, in what direction they will develop.
It is no secret to anyone that the classic paper business directories have gone down in history, it was difficult to work with them, it was difficult to find the necessary information, they were cumbersome. They were replaced by electronic directories such as COMPASS and Internet directories. The main disadvantage of these directories is access to current data. Any large help system requires constantly checking the relevance of the data. To do this, you need to have a large staff of workers, ring up all the companies regularly, request correct data. This is a very expensive pleasure, the business of reference systems does not have a high profitability; therefore, almost all reference systems have about 30% of outdated information.
We at YPAG.RU solved this problem in the following way: we passed the test of the relevance of the data to the system users. More than 60,000 people visit us daily, each looking for his own information. Users are in those parts of the directory, which is unlikely to reach the hand of the moderator. About 50 applications are submitted per day for correction of data. We check all this information and then adjust it.
The next problem is the relevance of companies to this section.
Previously, we trusted the owners of registered firms. As practice has shown, this is a very inefficient way, they indicated almost everything with at least some side concerning and sections relating to them, as a result, the reference book could turn into a simple dump, where you can find anything. Moderators are people, too, and it is also not very effective, and requires large resources to shift everything on their shoulders.
We built a neural network to define sections, it was trained on the existing base, after the learning process and adjustments somewhere in a month, the error in determining the section was about 3%. He designates not only the section, but also the company's place in the section.
Nowadays, reference books in a classic form are of little interest to users; they want to enter, press one button and get an exhaustive information about a product or service: cost, delivery time, warranty, etc.
Each help system goes its own way, some plant a pool of operators who are ready to advise users on a paid or free basis.
We tried to implement in the form of a tender center.
Any user can declare a tender, request prices, find dealers, and much more by filling out one form.
After that, the information is sent to the target audience. Companies that come to the tender page can bargain among themselves for a client.
So what will the reference systems be in the future?
The more technologies develop, the lazier the user becomes.
Perhaps in the near future, we will learn to identify user needs on the basis of indirect signs, behavioral targeting, etc.
Then the user can simply go to the directory, press one button and the system will give him 100% target information ...
It is no secret to anyone that the classic paper business directories have gone down in history, it was difficult to work with them, it was difficult to find the necessary information, they were cumbersome. They were replaced by electronic directories such as COMPASS and Internet directories. The main disadvantage of these directories is access to current data. Any large help system requires constantly checking the relevance of the data. To do this, you need to have a large staff of workers, ring up all the companies regularly, request correct data. This is a very expensive pleasure, the business of reference systems does not have a high profitability; therefore, almost all reference systems have about 30% of outdated information.
We at YPAG.RU solved this problem in the following way: we passed the test of the relevance of the data to the system users. More than 60,000 people visit us daily, each looking for his own information. Users are in those parts of the directory, which is unlikely to reach the hand of the moderator. About 50 applications are submitted per day for correction of data. We check all this information and then adjust it.
The next problem is the relevance of companies to this section.
Previously, we trusted the owners of registered firms. As practice has shown, this is a very inefficient way, they indicated almost everything with at least some side concerning and sections relating to them, as a result, the reference book could turn into a simple dump, where you can find anything. Moderators are people, too, and it is also not very effective, and requires large resources to shift everything on their shoulders.
We built a neural network to define sections, it was trained on the existing base, after the learning process and adjustments somewhere in a month, the error in determining the section was about 3%. He designates not only the section, but also the company's place in the section.
Nowadays, reference books in a classic form are of little interest to users; they want to enter, press one button and get an exhaustive information about a product or service: cost, delivery time, warranty, etc.
Each help system goes its own way, some plant a pool of operators who are ready to advise users on a paid or free basis.
We tried to implement in the form of a tender center.
Any user can declare a tender, request prices, find dealers, and much more by filling out one form.
After that, the information is sent to the target audience. Companies that come to the tender page can bargain among themselves for a client.
So what will the reference systems be in the future?
The more technologies develop, the lazier the user becomes.
Perhaps in the near future, we will learn to identify user needs on the basis of indirect signs, behavioral targeting, etc.
Then the user can simply go to the directory, press one button and the system will give him 100% target information ...
')
Source: https://habr.com/ru/post/93758/
All Articles