The second breath of reference promotion

Increasingly, it is possible to find information that links no longer work and that by purchasing them, you are more likely to get into the ban than in the TOP. Is it really? Is there any chance for external links? A study conducted by Rookee experts confirms that there is.

About what links will accelerate the output in the TOP and where is the neural network – read further.

The hypothesis and the beginning of the work

Rookee technology analysts suggested that links located on the pages of donor sites as close as possible in the sense of being promoted can be indexed by search engines better and this will speed up site promotion.

Machine learning was used to analyze content on pages and compose semantic pairs. We have trained a neural network that analyzes more than 1,000,000 pages of accessible donor sites and selects exactly the same as the promoted page *. Then they have placed links. Example: the promotion page offers car toning services. The neural network analyzes a lot of sites of automotive subjects and finds pages about toning. Then the link is automatically placed on the selected pages.

To confirm the hypothesis, we conducted a study.

* It is not necessary to use a neural network to match semantic pairs. You can select them manually. However, this technology significantly speeds up the process: for a couple of seconds the neural network will analyze as many pages of sites as people can not study for a whole day.

The course of the experiment

The study was conducted on a random sample of projects.

In total, the experiment involved 1294 domains, divided into three groups:

1. Experimental group number 1 (745 domain): projects that have been moving for more than three months, during which both lease and eternal references were already purchased. Most of them already were in the index of search engines. The group was divided into several subgroups: for each purchased a certain proportion of links on the new algorithm (from 5 to 50% of the total volume of rental links).

2. Experimental group number 2 (509 domains): projects that have not been promoted before and do not have a significant external reference mass.

Work began on the purchase of rental and perpetual links. In this case, the group was divided into several subgroups: for each purchased a certain proportion of links on the new algorithm (also from 5 to 50%).

3. The control group (40 domains): projects that have been moving for more than three months and have a significant reference mass. Links for the new algorithm were not purchased for them.

The study was conducted from September to November 2017.

The results of the study

After 2 months, the first results were announced. The rate of access to the TOP by low-frequency queries increased by 30%: search engines indexed links placed on related pages, much better.

The experiment showed better results for all experimental subgroups compared to the control group. At the same time, the most significant results were achieved on the projects from the experimental group No. 1, for which the share of links placed under the new algorithm was 30-35%, and projects from the experimental group No. 2, for which 35-40% of the links were purchased using the new algorithm.

Table 1

Experimental group number 1

% of links on the new algorithm

Dynamics

concerning the previous period

% of requests in the TOP 10 at the beginning of the experiment

% of requests in the TOP 10 at the end of the experiment

5%

-4

1%

0%

10%

+3

0%

0%

15%

+2

0%

0%

20%

0

0%

1%

25%

+8

0%

8%

30%

+9

1%

24%

35%

+26

3%

29%

40%

+18

0%

26%

45%

+24

0%

27%

50%

+19

0%

31%

Table 1 shows that there is a significant difference between the average position on requests at the beginning of the experiment and at the end. For example, when a 30% share of links from the total mass was placed in the new algorithm at the end of the experiment, the site was in the TOP for 24% of the requests, compared to 1% at the beginning of the experiment. We recall that similar results were achieved within two months.

Table 2

Experimental group number 2

% of links on the new algorithm

Dynamics

concerning the previous period

% of requests in the TOP 10 at the beginning of the experiment

% of requests in the TOP 10 at the end of the experiment

5%

+1

0%

0%

10%

0

1%

0%

15%

+4

0%

0%

20%

0

0%

1%

25%

-1

0%

0%

30%

+5

0%

3%

35%

+8

1%

14%

40%

+27

3%

27%

45%

+23

7%

29%

50%

+19

1%

33%

In table 2 we can see that the use of smart links will accelerate the progress in TOP and new projects.

Additional tests once again showed that domains for which 30% of links were purchased using the new algorithm showed better growth dynamics than other experimental subgroups.

On the graph, you can see the dependence of the percentage of requests in the TOP on the proportion of links from thematic pages in the index of search engines.

At the same time, placing more than 30-40% of such links does not make sense – there is no qualitative improvement in dynamics, and it is also important to remember the diversity of the reference mass.

Conclusion

Placement of links on pages that are as close as possible to the promoted ones increases the speed of access to the TOP by low-frequency queries. Just 30% of such links are enough to speed up the progress. They look as natural to search engines as possible and improve the site’s link profile because diversity is one of the most important criteria for the quality of reference mass.

Thus, it is too early to say that the reference promotion does not work. Just need to test new methods of placing links, use modern technology and monitor the change in search engine ranking algorithms.

Be the first to comment

Leave a Reply

Your email address will not be published.


*