Forum Moderators: DixonJones
That is quite reassuring for me - it has been a right pain finding a tool that will do it without charging more than the profit margins. Had to write our own in the end. In the process I have paid for and tried most of the ones mentioned and still do, since our code is a bit constricting.
I too have tried a lot of different tools, and still use many different ones. For very large sites (+100 million page views a month) NPS still seems to be the best way to go but for most other sites I have converted to Indextools.
Some of the things I like the most about it is:
- The script (you insert on your pages) is very stable. You can insert it anywhere on your page (not just in the header) and it works even if you have other JavaScripts running on your site. Customization is also very easy (for campaign and action tracking + grouping of pages)
- I can track all kinds of campaigns: PPC engines, Banner ads, e-mail marketing etc. - not just PPC engines. Some of the other (good) ROI tools on the market only measure PPC-engines but I (as most people in here) work with multiple methods including natural search and I want to compare how well each of them perform. In other words, I want to know not just if my Google Adwords work but also how well they work compared to natural search.
- I can set up actions very easy for just about anything - even including downloads (using an onClick event on the download link that triggers the script). Then I can segment actions based on campaigns, language, country (using the build in geo targeting DB) etc.
- The tool is tracking exit links! I haven't seen that before. I guess they must be using the unOnload event to that. Very neat.
- I can dig into single users and follow each step and action they have taken on the site.
On top of that the company has proven to be very responsive to suggestions for improvements. I had a long list of European based engines I wanted to add when I started using it and within 2 days they where all added! For another client we needed some changes done in order to track a special file extension as page views and that was also done within days.
I should probably add that I am not an affiliate of the company - I just like the tool very much.
Receptional, NetTracker's Initial Referrer summary gives you that.
I use NetTracker and Webtrends. Most clients are familiar with Webtrends and that's the only reason I stick with it. NetTracker blows it out of the water!
Anyone using Indextools on a US based site. Their domain seems to be registered in Hungary and wondering is this going to cause a problem if the servers are overseas.
Hitslink provides excellent statistical analysis and is very easy to use as it has a web interface. Provides useful info on visitors' screen resolutions, browser types, navigation paths etc. I have it on 3 months free trial
Webalizer came free with my web hosting deal. Seems to count everything including visits by robots etc. It is a bit trickier to configure as you have to edit a Unix file. You can customise it to show useful info such as what search strings people have used to find your site.
Both seem to do the job fine, I like having 2 sets of stats as a cross check of each other.
And occasionally a c script for more complex queries. But then again, I generate 2 gigs of logs a day, so processing that with standard stuff is pretty much not an option.
I am really surprised to see how few people have put something in that properly measures where converted traffic came from (ie... A guy buys a widget, after three visits to the site, which search phrase and engine did he originally come from).
One reason might be because it's impossible. ;)
Based on standard logs it's certainly not possible to infer that kind of information with any reasonable degree of accuracy. It's not even at the "educated guess" level. With active visitor tracking (using cookies etc) you can do somewhat better, although you end up with a sample that's reasonably accurate but probably not representative.
Log analysis beyond the most basic level unfortunately involves a fair amount of guesswork. The danger of leaving to much guessing to the software is that it's very hard to know if the results can be trusted.
For the record, I base some of my guessing on analog output. It's a good program when properly configured.
Bob
Now I switched to WebCEO's Hitlens (great traffic amalysis and visitor behaviour for $6/m) and ConversionRuler(ROI calculation and conversions for approx. $30/m).
like it in general.
