| Web
Analytics - Murder by Numbers - Part 2
Not accessing
and reviewing your vital website statistics is like never looking
at your checking account activity and never knowing how much money
you have in it.
In Part
1 of this two-part series I explained how to crunch
relevant website statistical data to facilitate constant e-marketing
initiative improvements. I explained what types of data are important,
such as unique visits, click-thru numbers and percentages, lead
conversion rates, and how to process all these numbers. Here in
Part 2 I’ll explain how you obtain the data in the first place
and then provide a fool-proof method for website click-thru statistical
acquisition.
The first thing
you need to know is where your website lives. Every website sits
on a server, a computer with the purpose of waiting for requests
from clients (people’s personal computers by way of a browser).
Each server physically lives in one of two places. It is either
located at its website owner’s company, which is called in-house,
internal, or self hosting. If company A has an active website and
owns the server the website is on, and the server is physically
located at their company, then it falls in this first category.
The other place
a website server can physically live is at an Internet Service Provider
(ISP) or host company. There are a number of configurations the
server can fall under in this category which is beyond the scope
of this article. The main thing to keep in mind is you first need
to know where your website’s server is.
Once you know
this, you can begin to assemble all the relevant site statistics.
All servers automatically generate all the data you’ll ever
need on an ongoing basis. They are relentless in their stats recording.
They record all the data in what’s called server log files.
Manually parsing through these log files is a horrible job that
should only be wished on your worst enemy. They are huge laundry
lists of everything from every site visitor’s IP address,
browser type, site referral, time and date visited, and much more.
Fortunately,
there are software programs that can do this for you. One of the
most popular is WebTrends (http://www.netiq.com/webtrends/default.asp).
You feed your server log files to the WebTrends software, and it
produces for you an excellent presentation of all your relevant
(and some superfluous) website statistics.
If your website
sits on a server that your company has in-house, than you need to
purchase WebTrends or some similar software and locate your server
log files. The files often end in .log. In other words, it’s
up to you to get your website’s statistics, and you do this
by locating your server log files and running them through software
such as WebTrends.
If your website
sits on a server in an ISP then you can either request the server
log files from them and run them through your own software, or you
can ask them if they provide an interface for you to review your
site statistics online. Most do provide this service. It’s
often web based and all you have to do is log onto their site to
view them.
Now you’re
armed with a lot of good data. But if all your e-marketing initiatives
drive traffic to your homepage, how will you know which ones are
working and which ones aren’t? If you send out emails to rented
lists and the call to action are all links that point to your homepage,
then you’ll never know which emails are doing better than
others. You may get an idea by seeing if your overall traffic increased
the day you sent out the email or posted the banner (even to determine
this you’ll need your website stats), but to do it right,
you need exact data, and the web will provide it for you.
Some sites that
you place banners on will offer you click-thru counting services
to you. Most email brokers also offer similar services, at a price.
But what if they don’t offer tracking information for you?
Or worse, what if you don’t trust their reporting?
The solution:
Create, implement, utilize and manage your own unique tracking pages.
It’s relatively
simple. In every e-marketing campaign you conduct you create and
assign a unique html page to it. Then the initiative’s call
to action (hyperlink) points to its unique page. After the campaign
is done, you can then go to your website statistics obtained through
your website’s server log files, and see how many visits were
logged for each unique tracking page.
For example,
let’s say you send out an email to a list of 1000 email addresses.
In the body of the email there is a call to action link that says,
“Click Here to Buy Now”. This link points to a page
on your website. But not just any page. It points to a unique tracking
page you created earlier to track how many of the 1000 people clicked-thru
from the email. It’s important that no users can get to this
new page in any other way than through the email. Let’s say
you named the page email-campaign1.htm. After the email campaign
is done (I like to wait about 2-4 days), you go to your website
statistics (the result of parsing the server log files through WebTrends
or its equivalent) and search for the page called email-campaign1.htm.
Finally, you view the page visits number. Let’s say the visits
to this unique page totaled 200. That number is your click-thru
number.
Now you can
really start to fill in all the relevant data discussed in Part
1. This will enable you to determine how well each campaign
is doing and whether you need to make adjustments.
To help manage
all these unique pages, keep them all in one sub directory of your
site. If you don’t do the technical work for your site, you
ought to consider giving Part 1 and Part 2 of this series to your
technical web person so they can get a better handle on your website
vitals.
Until you know
how well your website and e-marketing campaigns are doing, measured
in visits, leads and sales, you can’t possibly maximize your
operation and increase your bottom line. Now you have the information
to make this happen.
Article
by Jason O'Connor
© 2004
|
