How to understand keyword data? Some interesting statistics

Keyword selection and analysis in the work of an SEO specialist is the daily bread. When we think of keywords we probably think of the data by which we analyze them. Metrics such as:

• average monthly number of searches,

.

• cost per click (CPC) in AdWords,

.

In today’s post, we’d like to introduce you to the topic of these two metrics and show you some interesting statistics we’ve created based on our 101 million keywords.

Are 0 searches really 0 searches?

.
We mostly get our keyword data from the Google Keyword Planner tool. This is a tool that, by design, is meant to make the process of building an AdWords campaign easier. One of the metrics that Keyword Planner presents is the average number of monthly searches for a word. It tells you how many times a keyword is typed into a search engine per month (this is the average of the last 12 months).

This is where the first problem arises. The tool, in the case of many keywords, shows 0 as the average monthly number of searches. Many people consider that an SEO campaign for such words does not matter, while this is not the case. Google, in order to organize its entire database, created ranges for its data. What this means is that the average monthly number of searches is not the real number of searches, but the number matched to the closest lying interval. So 0 searches means that the keyword falls within the range of 0-5 searches per month..

Currently, for Polish results, Google uses 70 search ranges. The ranges start from 0 to 68 million searches (the phrase “Facebook”). Below we have prepared a table that tells you exactly which ranges Google uses and how many keywords are in each range.

Individual columns stand for:.

• Average Monthly Searches: this is the range we are in. 0 means that these are words that have 0 searches according to Google Keyword Planner

.

• Word count: this is information on what number of words out of 101 million were ranked in a given range

.

• Potential of all words: this is information on how many searches all words in a given range have

.

• % of all words: this is information on what % of 101 million words make up a given range.
• % of total potential: this is information on what % of all searches the sum of searches for a given range represents

.

Analyzing the results from the table, several conclusions emerge:

• The long tail rules: 93% of all keywords fall within the range of 0-10 searches, they also generate almost 10% of all searches (if one were to take into account that 0 does not always mean 0% this % would certainly be higher)

.

• The intervals between the 0 – 100 range (the highest concentration of keywords) are small. However, if we look at the average interval (for example, between 60,500 and 74,000 we have 13,500 searches). Thus, having a phrase assigned to a given range the data overestimation can be as high as 6250 searches! Therefore, if you are estimating the traffic your site will get with the campaign, you should take a large reserve for keywords in the higher ranges.

So we can see that we can’t quite trust the data that Google shows us, especially when it comes to popular keywords – the misrepresentations can be quite large.

Some interesting statistics

.

Potential of words by number of words

.
The following table shows the breakdown of phrases by the number of words in the phrase.

CTR by position

.
Individual columns mean:.

• Number of words: the number of words in a phrase

.

• Word count: the number of words representing a given number of words in a phrase

.

• Sum of average monthly searches: sum of average monthly searches for all words in a given range

.

Some conclusions:

• The majority of phrases consist of 2,3 or 4 words – they account for 85% of all phrases.
• The majority of phrases are made up of 2,3 or 4 words.
• The greatest potential lies in 2-two word phrases – all words are searched 640 million times per month.

Divide words by features

.
At Senuto, we also study the characteristics of keywords. The features we recognize include:

• Brand: information whether the keyword is not associated with any company or product name

.

• City: whether the keyword includes a city name

.

• Name: whether the keyword contains any name

.

• E-commerce: whether the keyword is related to a product, e.g. “refrigerator”, “TV”, etc.

.

More information

.
On 14.01.2016, we had the pleasure of conducting a webinar on keyword research for SEO campaigns, in which we discussed some of the above statistics. We encourage you to watch the recording and learn more about our presentation.

I hope the above post has given you an idea of what the distribution of keywords in our country looks like and how to understand keyword data.

Read also: Keyword Search – a step-by-step guide.