Forum Moderators: DixonJones
I don't think I have ever used a forum before, but here goes.
I made a website for someone who sells dog coats.
They have won some design awards recently and got some media coverage. I am looking at the awstats (I think I have a few other stat programs too)
But I really don't know how to make use of them.
Perhaps a few people here would be interested in showing screen shots of there webstats - we could take turns - and people would give there ideas about them, etc. We might all be able to learn something, and hey, it might even be fun!
I can also offer people some SEO optimization ideas as we go along... as I do a bit of that, but I'm pretty much doing it by guesswork.. I know some theory, but I don't have feedback...
Thanks
Russ
[edited by: Receptional at 3:37 pm (utc) on July 22, 2005]
[edit reason] No specifics please See terms of service [/edit]
Thanks for posting in Webmasterworld. Please Take the time to review THIS POST [webmasterworld.com] as you will get much more out of our forums.
We avoid specifics in here, so posting up screenshots would be a recipe for spammer to add their links to webmasterworld I am afraid. Instead, I am much happier to answer you in more general terms. I have not used the package to which you refer, butu shouldn't need to to help you at least a little.
The problem with free log based stats is....
1) A "Hit" is not a visitor in the way you would think. A "Hit" is technically a request to your website for information, as called by another machine. This could be a request for an HTML page, a ,GIF file, a word document or any other file on your server. If your hime page is frames based, with a top and a bottom and a "main", then has four pictures on the page, then that will be seven "hits" just to download the page. But when a user goes to the next page, it is quite possible that this time he does not need to download all of the same files again, so it may be less... As you can quickly see, measuring hits is a very poor indicator of site use.
2) What about IP addresses? You could try to measure users based on their IP number - but again you get problems. Mosrt people live on a "dynamic IP" - so their IP number can change from session to session and - with many ISPs - can even change during a session. Worse... the ISP could reassign that IP to another user, who may then come to your site and be identified as the chap who just left!
3) What about just recording page views? Well - you could give up on recording unique visitors and just record page views - but STILL there are problems. ISPs will "Cache" pages, so that they do not have to call your server too often. Not likely on a quiet site, but surprisingly likely as your traffic builds. So it is extremely possible for people to see your pages without even touching your web server or your log files!
So where is all this leading? Well - my preference is not to use a LOGFILE analyser to measure trafiic, but to use a javascript system instead. This has many other problems, but you probably don't want to know them yet! [webmasterworld.com]
Disclaimer: I prefer to use web server logs because I have used them for many years. I also caution the people who use the information that is derived from the logs that it is a description of activity on the server only.
An argument for using web server logs is that you get the information for each time your site is touched,
That is actually a fallacy. In the event of cached pages, whether by the browser, a local web accelerator, or a proxy, those hits will not be recorded by the server logs because the server is not involved in delivering the page to the end user.
In the case of web sites with aggressive caching policies this will lead to significant underestimation of traffic. Of course, the sites with the most need for caching are also the ones with the most to lose from under reporting traffic.
The above can be demonstrated with absolute certainty by observing traffic using a sniffer.
"cached" site:webmasterworld.com
... youll get a shed load of hits. Read the ones that look relevant and youll be able to pick out the pros ans cons for each approach to capturing activity data.
After I had read everything, I started writing a report for my employer about reporting strategies. So far im on 20 pages and Ive only covered logfiles so far... still got js tagging and bespoke solutions to go!
In general, most packages offer the same kind of stats. The trick is though to make sure that they suit your site. My site is so bespoke that Im looking at developing a custom solution. The reason for this is that if I go for just log file analysis or just tagging, my stats will be pure lies, or not cover every bit of activity that I want to cover. Sure it wont be 100% accurate but at least ill know that what stats i do manage to capture are fairly good. The end result would be pretty graphs I guess!
As for the usage... well thats a specific thing to the owners of the site. For example, the content writers want to see how much people have read thier article... thats quite a common thing but not the same for every site out there. Also our stats show bandwidth used and other techie things, which we cross reference with stats out of BMC Patrol. From this I can see which bits of my site are causing problems. For example, one SQL search I had was performing ok from the front end, but causing a lot of disk activity. Turned out the index on that db table wasnt fantastic. What Im saying is that stats are useful for business and techical decisions as well as impressing the content owners.
S
1. My stats do tell me the number of UNIQUE
visitors ...
2. I can see how these people found my site.. keywords or through alot of the 2 way links I set up.
3.I can see some detailed info, like what pages and images were served.. how long they looked at the pages et.
WHY I posted this origional topic is to discuss HOW to use this information.
So, IF I have alot of information in front of me (I think the unique visitors gives me a good idea right? and the whole bunch of other stuff)
How do I re-optimise, How do I use this information...
ok, people are finding the site when they search for certain keywords..
how do I maximise on what I have with the info I have.
I have some page optimization skills (that I learnt doing boring work for a US company) but I dont know how to tally it up with the stats/server feedback I am getting...
If It was traffic, and I could see that one road was being clogged, or another never used.. I would have an idea about what to do right? I would use the statistics and maybe put a TOLL fee on the overused road or develope the underused road etc.ect
It just feel lost trying to analyse the stats I have right now...
-Russ
Yes, These IS a good example of how to use the statistics.. I have been doing that a bit.. and it is important.. seeing how a customer passes through your site, and where they stop (or get cuaght up)
I guess it is some of the statistics like keyword phrases that they found you under... that I am pondering over. what do I do with them, what do they tell me.
(I can see which search engines are working for us.. and can focus on particular ones.. )
-Russ
WHY I posted this origional topic is to discuss HOW to use this information.
Well, that's a pretty open-ended question. IMO, you need to first understand everything about the stats -- you should be able to associate every stat with some actionable event on your web server, or an HTTP client, or whatever other entities are being captured in your stats. Then, you need to understand what it is you are trying to do. Do you want to make more money? increase pageviews? increase ranking? decrease ad spend (assuming you're doing that)? You also need to keep track of measurements over a long period of time, observing trends, and correlating those with other aspects of the business (sales, costs, etc.).
