We know since three years ago, Google stop giving website referral information cause of "privacy issue", there will be lost lots of information while we trying to know about users. But we still having few information about keyword in referral cause there are not all "(Not Provided)".
But is this true while we trying to figure out the all truth even we have only few evidence to know about all the facts?
In last articles, we know there are strongly relation with landings distribution and keywords distribution to SEO effort, if we want to earn more about traffic, we must know about what users read.
So, we can only read few data to figure out all the facts, but there is a very big question: Is there any bias in all data from "provided data"? This is a very good question, but I will post my experimental and tools next time to solved this problem (How will tools tell us after "not provided" age).
We know since Oct, 2011, "Not provided" become most visits keyword, and how many percentage so far and evolution? And another question is: How many keywords we lost in this situation? We must setup these dimension to watch the status:
1. Organic Search Visits
2. Search Percentage
3. Distinct Keywords Count
4. Distinct Landing Count
5. Keyword vs Landing Ratio
6. "Not Provided" visits
7. "Not Provided" Percentage
This is hard to see all the trends and relation, cause each website has its growth curve, but we just show the numbers today, I will draw a graph to show the evolution next time.
There are four kind of charts show lots of differences:
We can see the percentage of "not provided" varies from 40% to 80%, and hard to see the relationship to organic search visits and percentage, but it do growth a lot indeed.
But we can focus in "Keywords Landing Ratio", that is the number of rows of landing pages and landing keyword comparison, we can see the ratio decrease to 2/3 or 1/2 in these three years since "Google not provided", that is, it do loss information cause this situation. But we still have more than half rows of keywords to get information even we lost 80% rows.
There is very interesting thing, why the percentage varies a lot? Is this the evidence of devices or users behavior?
These is keyword distribution curves mapping the charts with same order. If we can take a closer look, we can find there is a very interesting things that if the curve grow slower, the percentage of "not provided" is more than others.
Can you guess what is the reason why will this happen?