There is an old saying in marketing: if you cannot measure it, you cannot manage it.
There is another lesson that is just as important:
If you do not understand what you are measuring, the number itself may be almost meaningless.
This matters particularly when somebody is trying to sell you online advertising.
You may be told that a website gets hundreds of thousands, or even millions, of “hits” every month. It sounds impressive. It may even be accompanied by a report from Google showing millions of impressions.
Before reaching for your wallet, there is one very simple question to ask:
What exactly does that number measure?
The answer can make an enormous difference.
A “hit” is not necessarily a visitor
The word “hit” has been used for years as a loose description of website traffic, but it is not a particularly useful modern measure of an audience.
A hit can refer to an individual request made to a web server. A single webpage can generate numerous such requests for its HTML, images, scripts, fonts and other resources.
Consequently, one person looking at one webpage could potentially generate many hits.
That makes “five million hits” a particularly poor way of describing how many people actually use a website.
A potential advertiser should instead be asking for meaningful measures of audience and activity, such as users, sessions and page views.
And even those figures need to be understood properly.
Then there is Google Search Console
This is where things can become particularly confusing.
Google Search Console is an extremely useful tool. It tells website owners how their content is performing within Google’s ecosystem, particularly in Google Search.
Among its principal measurements are impressions and clicks.
Google defines an impression as the occasion when a link to a website is shown to a user in Google Search, Discover or News, subject to the rules for the particular type of result. A click is recorded when the user clicks a link from Google to the website. Google calculates click-through rate by dividing clicks by impressions.
Those are valuable measurements.
But they are not website visitor figures.
Imagine that a newspaper website has an article about driving in Spain.
Someone searches Google for information about driving in Spain and the website’s article appears in the results.
That can produce an impression.
The person may click it.
That produces a click.
Or they may not click it at all.
The impression still happened.
Therefore:
Impression ≠ visitor
and:
Impression ≠ page view
and:
Impression ≠ reader
It is an indication that Google presented the site’s content as a possible answer to someone’s search.
That is useful information for an SEO manager.
It is not, by itself, evidence of the size of the website’s audience.
The numbers can become surprisingly different
Consider a completely hypothetical example.
A website appears in Google Search 5 million times during a month.
That sounds enormous.
But suppose those appearances generate 60,000 clicks.
The site’s Google Search click-through rate would be approximately 1.2 per cent.
That means roughly 98.8 per cent of those search impressions did not result in a click.
That does not mean the website is unsuccessful.
Far from it.
It may have a large and valuable audience, and thousands of people may still be finding its content through Google.
The point is simply that 5 million Search impressions and 60,000 Search clicks describe two very different things.
And neither figure tells us how many people actually visited the website during the month.
Search Console and Analytics answer different questions
This is why it is important not to treat Google Search Console and Google Analytics as interchangeable.
Very broadly:
Google Search Console asks:
How is my website performing in Google’s Search ecosystem?
Google Analytics asks:
What are people doing when they visit my website?
Search Console can tell you how often your content appeared in Google Search and how often people clicked through.
Analytics can provide information about users, sessions, page views, traffic sources, engagement and other activity occurring on the website itself. Google defines a session as a period during which a user interacts with a website or app, while an engaged session is one lasting more than 10 seconds, containing a key event, or involving at least two page or screen views.
Those are much more useful questions when somebody is trying to sell you advertising on the website itself.
Why this matters to an advertiser
Suppose someone offers to put your banner advertisement on a website.
You are not buying its Google Search visibility.
You are buying exposure to people who visit the website.
Therefore, the important questions become things such as:
- How many users does the website have?
- How many sessions does it receive?
- How many page views are generated?
- How many of those pages contain my advertisement?
- Where will my banner appear?
- How frequently will it be displayed?
- How much of the audience is in my target geographical or demographic market?
- How long do visitors typically remain engaged with the site?
- Can the publisher provide actual advertising impression figures?
Those questions are much more relevant than simply asking how many times the website appeared in Google.
There is another trap: “people” are not the same as “visits”
Even when you obtain proper website analytics, you still need to be careful.
A website can report users, sessions and page views, and those are all useful measurements, but they are not interchangeable.
One person can visit a website several times.
One visit can contain several page views.
One person can also use more than one device.
Consequently:
Users ≠ sessions ≠ page views
This is not a flaw in the statistics. They are measuring different things.
The mistake is treating them as though they are the same thing.
And there may be something else hiding in the numbers
There is another issue that is increasingly important when looking at website statistics:
Not all website traffic is human.
Websites are constantly being accessed by automated systems.
Search-engine crawlers, monitoring systems, content scanners, social-media crawlers, security tools, bots and other automated processes can all interact with websites.
Some automation is entirely legitimate and necessary. Google needs to crawl websites, for example, in order to understand and index their content.
Other automated traffic may have little or nothing to do with a human visitor consuming the site’s content.
This is another reason why a headline traffic number should never be accepted without understanding how it was measured and what has been included or excluded.
Modern analytics systems attempt to deal with various forms of automated and abnormal activity, but no measurement system should be treated as a perfect electronic census of human beings.
No statistic is perfect
This is perhaps the most important part of the whole discussion.
The answer is not:
“Google Analytics is always right.”
It isn’t.
Analytics systems have limitations.
People block tracking.
People use privacy tools.
Cookies can be restricted or deleted.
Different browsers and devices behave differently.
Consent requirements can affect what data can be collected.
Some traffic can be difficult to classify.
Different analytics systems can produce different figures.
Consequently, there is no magic number that represents the absolute, perfectly accurate number of human beings who visited a website.
The objective is not to find a perfect number.
The objective is to find the most appropriate and transparent measurements for the question you are trying to answer.
The question should come before the number
This is the principle that is worth remembering.
If you are an SEO specialist, Google Search Console is enormously useful.
You may want to know:
How often is my content appearing in Google?
Which search terms are generating visibility?
Which pages are attracting clicks?
How well are search impressions converting into visits?
Search Console is designed to help answer those questions.
If you are selling advertising on your website, however, the prospective advertiser is likely to have a different question:
How many people might actually see my advertisement?
That requires a different set of measurements.
And if somebody is asking you to pay for advertising based upon a headline number, it is entirely reasonable to ask them to explain exactly what that number represents.
Do not be impressed by the number. Be impressed by the explanation.
A publisher who says:
“Our website receives 250,000 sessions per month, generates 600,000 page views, and your banner will appear on pages accounting for approximately 300,000 of those views.”
has given you something that can actually be assessed.
You can then ask whether the price represents good value.
A publisher who says:
“We get five million hits a month.”
has given you a number.
But they have not necessarily given you useful information.
Ask:
Five million what?
Search impressions?
Server requests?
Page views?
Sessions?
Users?
Advertising impressions?
And how were they measured?
Those questions are not being difficult.
They are simply quantifying the claim before you buy it.
Because the most misleading number is not necessarily a false number.
Sometimes it is a completely genuine number that is being used to answer a question it was never designed to answer.
And that is why, whenever somebody gives you a very impressive statistic, the smartest response may be the simplest one:
“That’s interesting. What exactly does it mean?”
Other Articles
The Summer Ahead
