Forum Moderators: LifeinAsia
I have noticed the 'traffic' that is being advertised at competitors sites vary. Some use 'hits', some use 'page views' while others use 'users'.
I would like to advertise using a STANDARD. But what IS the 'standard'? Hits, views or users?
Or is there a reason to use any of these in a different way?
The problem with that is, my site(s) are growing so fast, that I may need to modify that metric with a projected yearly rate,based on current growth.
Problem with that is that no analysis tool is really able to tell you how many unique you have. And you can expose one visitor to an advertisement more than once, which will add to its brand effect, so the number of page views is the important one.
Not quite true. It is often perceived that the numbers given to advertisers are uniques, but in most cases they are either estimates or opportunities to view. As when a TV ad is either viewed by the whole family watching the evening movie or not at all because one went to the bathroom, another to get more beer and so on. Same for billboards, how could they tell how many people are going to walk/drive next to it AND look at it.
As I said above, unique visitors can not be determined exactly, but the web has the opportunity to give advertisers data that is at least more correct, i.e. the page views. I also know that some views are registered although the page never loaded completely or that some page views come from the browser cache. But IMO it is still the most exact data.
Reach and frequency - how many people will be exposed, how many times. Those are the measurements advertisers are looking for.
How many times people will be exposed => page views.
if 2 websites each generated 100,000 page views a month, with one receiving 50,000 visitors and the other 10,000 visitors, which would you prefer to advertise on? comparing page views doesn't show this
My competitor advertised 'XX,000 page views in one month!'. Well...that may be true, but his site is a mess, thus the large quantity.
However, if I advertise my 'uniques' compared to his 'page views', I lose. Why? Because my customers don't know the difference...and if I try to explain it them, their eyes glaze over. If I have to give a primer to every customer, I have the potential to lose them. I'm making a sale, not a tech presentation. If the two get mixed too much, the customer will think it's too confusing. (Been there-done that.)
That's why I am trying to use the same metric-page views. That way I can at LEAST get onto rankings in SE's.
How many times people will be exposed => page views.
But what's the reach? My point is that advertisers work on the basis of reach and frequency. Say you have 100 pageviews. Are you reaching 100 people once or one person 100 times? Which do you think would be more valuable?
A frequency number without any reach information is of no value. However, a reach number - by itself - can provide useful information to an advertiser.
I still don't like the visitors because it is not possible to have any exact data on that. My webtrends report is configured to give me 'New vs. Returning Visitors' and tells me 98.76% unknown. So how can it tell how many unique visitors I have?
Can we agree on page views AND visits?
Sure :)
I think that without cookies it is almost impossible to extract meaningful information on uniques. You also bring up another meaningful statistic - new vs. returning visitors. That, too, is an interesting statistic which can be best tracked using cookies. That said, my background is marketing, not html - so don't ask me for hints on writing the code!
I think as advertisers get smarter about buying the web they will be looking for more information than the typical "log analysis" program can provide. Page views alone has such potential for being misleading that I think you should be looking to arm yourself with higher-quality information.
Don
new vs. returning visitors. That, too, is an interesting statistic which can be best tracked using cookies.
Cookies will not tell you that much either. Many people don't accept them and then there are people like me who just delete them every other day.
I think as advertisers get smarter...
Many people don't accept them
About 96% do accept cookies, according to statistics I saw recently. As for deleting them every day, the average user wouldn't know where to find cookies, much less delete them:).
As for advertisers getting smarter, I think that will slowly happen. Now, Media Buyers getting smarter, that's a whole different story ;) In the broadcast business, we're still trying to convince them that the quality of the audience matters more than cost per point...oh well. One day!
Take care.
What is defined as a unqiue visitor (a unqiue server? A unique server with visits more than 20 mins apart, or 24 hours apart etc?) This makes a big difference.
That you have not counted things like SE robots, rippers, email scrapers, and the heap of other automated hits that don't get counted. These can account for say 10% to 80% of page views!
What you are counting as page views (just htm or php, asp, etc) or counting 3 pages in a frame as 3, or includes, js, and even css, which i have known some people to count as "pageviews"!
Well at least thats what I ask people who quote traffic stats to me.
My preferred solution would be to transfer the access logs to a database, which allows for quite complex and rapid (relatively speaking) queries. You could - for example - *estimate* the number of unique (human) visitors by running a "remove duplicates" query on the IP numbers stored in the database, and by asking the database to eliminate bots, sitesnaggers and other 'automated' hits based on their known signatures.
Firewalls might give you identical IPs for different users, and you will never know exactly (with absolute certainty) how many unique visitors you've got during a specific period, but that's not really necessary either. Estimates and approximations (which can be supported by statistical analysis) can be very helpful in detecting and analyzing trends.
As Chiyo has already pointed out, it's also very important to clearly define the term "unique visitor" (or any other metric which are important to you). This term is not absolute, and your own definition might change later, so it's important to know the definition with regard to statistical extrapolation or estimation for comparative analysis. It's also important to know for those who're interested in not only your figures, but also *how* your data was calculated.
