Google Update June 2019 — Latest Research

Google’s Latest Core Algo Update started June 4th [2] and finished rolling out June 8th [3]. In terms of coverage, this was the same update as the Medical Update in August 2018.

At the same time (since June 4 till June 6, 2019) Google rolled out another technical update, which is an independent update and coincided with a wide kernel update (“June update”).

Danny Sullivan on Twitter said [4] that in most cases Google will not show more than two results for a single domain on one query.

Objective

The objectives of this study were discovering the difference between two groups of sites – Losers and Winners.

Methods

The mobile—first indexing will be enabled by default for all new, previously unknown to Google Search, websites starting July 1, 2019, — John Muller said [5].

Assumption – all of the most important Google algorithms were updated, including Panda and Penguin. The Mobile—first indexing update, in fact, more ambitious than that mentioned by John Muller.

According to these conditions when we are comparing the list of domains, the focus of the study is specifically on mobile issuance rates. Traditionally Google all of the system updates are mainly rolling out in the USA. So, the region for data collection is the USA.

Our sampling includes 200 sites.

Used data sources:

SEMrush Sensor - information about the degree of Google update influence on sites depending on their Topics (niches). Indicators of positions and keywords changes on the TOP100 Winners and Losers.

Majestic - dataset from the Bulk backlink checker tool.

Ahrefs - dataset from the Batch analysis tool.

Google PageSpeed Insights — Mobile Score, Mobile First Contentful Paint, Mobile First Meaningful Paint values for the mobile version of the site.

Used software:

Netpeak checker - scrapping Google PageSpeed Insights data for selected sites.

Orange - Data mining collected data.

AnswerMiner - Correlation Matrix and Forecast Tree.

Try Now Free

Data preparation

All non—numeric data are converted to numeric values.

The separator character comma is replaced with a dot. Otherwise, Orange does not perceive such data as numeric.

The target variable is Status (the values are Winner, Looser).

Results

The most affected sites by last Google Update topics (niches) can be divided into two groups:

News (9.10), Sport (7.90), Food and Drink (6.90), Games (6.60), Online Communities (6.60), Arts & Entertainment (6.10);
Internet and telephony (5.80), Animals (5.70), Health (5.50), Finance (5.30), Beauty and Fitness (5.20), People and Society (5.20).
In the first group, only News belongs to YMYL sites.
In the second all topics, except the first two, belongs to YMYL sites.

With Orange, sites from our sample were clustered based on data set numeric values.

Most of the Losers are in the one cluster C5 (painted in yellow color, see Figure 1).

Figure 1.

Fig-1

The Winners are less common and most of them are in two clusters — C3 and C6 (painted green and purple colors, respectively).

We conducted factor analysis to examine the structure of the factor loadings and intercorrelations for all the items and dimensions. We employed a correlation value, provided by AnswerMiner, for examining the correlations between each dimension of our dataset.

According to AnswerMiner data, the strongest correlation to the target variable “Status” has “New keywords” criterion (Very strong correlation — 79% - Figure 2.).

Figure 2.

Fig-2

In Figure 2 we can see that Winners have got more “New keywords” are ranked by Google.

Mobile Score Google PageSpeed Insights has a strong correlation to target variable “Status” (36% ).

The following indicators have a moderate correlation with the target variable “Status”:

OutLinksExternal — 15%, Majestic API term — the number of outgoing links for this domain;

NonUniqueLinkTypeHTTPS — 15%, Majestic API term — the number of non—unique links that use HTTPS as a protocol;

Trust Flow — 14%, Majestic metric — trust of the site;
Ref domains Dofollow — 13%, Ahrefs metric;
Total Keywords — 13%, Ahrefs metric;
Total Traffic — 11%, Ahrefs metric

Figure 3.

AnswerMiner Data Analysis Tool - Relation Matrix

Mobile Score has a strong correlation with the “Dynamics of positions” and “New keywords”.

The likelihood that a site will be assigned to the Losers category is 87%, if “New keywords” is set to zero in the dataset.

If “Trust Flow” less than 40 and at the same time “NonUniqueLinkTypeHTTPS” is less than 44, the probability that the domain will be assigned to the Looser category is almost 70 percent.

Many Winners with Trust Flow from 35 to 75 do not shine with Mobile Score. But the difference in Mobile Score between Winners and Losers is not statistically significant, i.e. if there is, it is more likely random.

Figure 4.

Fig-4

Conclusions

Under a simplified procedure, a statistically significant difference is when the difference is not accidental.

As our study shows, Winners are more trustworthy (have better Trust Flow value) than Losers and this difference is not accidental.

Figure 5.

Fig-5

The difference in the values of the “Total Backlinks” (Ahrefs metric) between these two groups of sites is also statistically significant.

Figure 6.

Fig-6

The Google Mobile—first indexing update affected Losers only. It can be assumed that if the trusted sites publish a lot of useful unique content and the number of keywords ranking up, Google can “closes its eyes” to a low Mobile Score.

If you buy links, try to get them from pages that use the https protocol.

Losers have fewer trust and backlinks, have few or no “New keywords” ranked by Google.

The problem with the new key may be related to blocking (termination of updating) the sitemap file in Google Search Console. This is most likely not a bug, but a method of punishment.

Sites that have had more than two results in a SERP can significantly lose traffic because Google will not show more than two results for a single domain on one query.

Check your backlinks in Google Search Console. It is noted that sites that have lost traffic in the Google June update 2019 have many links from .cf, .ga, .gq, .ml, and .tk domains. You need to check .info domains as well. Most likely this is spam. You may disavow them.