|
Frequently Asked Questions
General
General
A: Yes, but not in the traditional sense. ClickTracks is best described as a visitor behavior analysis tool. While ClickTracks does read web server log files and does present statistical data, we don't really consider it to be similar to most existing 'web stats' programs. Because ClickTracks is designed for marketers and online advertising professionals, it intentionally leaves out the superfluous technical information provided by standard web stats software.  Log files
A: The log file is generated and stored on the web server. Use an FTP tool to login to your webserver in the same way you would to upload new web pages. You'll find the log file in a directory close to the HTML files directory. Q: My log file is sliced into daily/weekly/monthly chunks by my ISP. Which do I need? A: When using ClickTracks for the first time. You'll need about 1 month of data. Download the appropriate files and simply import daily or weekly log files after that--ClickTracks will automatically append your imported files to your existing dataset. Q: How does ClickTracks handle importing log files that overlap with existing data? A: It's handled transparently. ClickTracks works out what it already has and only adds the newest data. Q: My ISP doesn't generate log files for my site. What can I do? A: Check out ClickTracks Hosted Q: What format log files can ClickTracks read? A: Apache 1.x & 2.x, Microsoft IIS 3.0, 4.0, 5.0, 6.0 Netscape / iPlanet 3 & 4 Any W3C Extended compatible log ClickTracks identifies the file format automatically--there is no complicated configuration. Q: My log files are huge. Do I need disk space on my desktop PC to store them all? A: No. You can download each log file as it's generated on the server and import it. ClickTracks converts the data to an proprietary format that speeds the analysis process and takes about 1% of the original disk storage space. Once imported your log files are no longer used by ClickTracks and you can delete them (though you should keep a copy of them somewhere.) Q: Can ClickTracks read compressed log files? A: Yes. Ideally you would compress the files on the web server to .gz format and simply download and import the compressed file into ClickTracks. Q: My site traffic is huge. Can ClickTracks scale? A: Yes. ClickTracks handles even very large sites with millions of hits per day. There is no limit to the size of log file you can read. A 1Gb file takes about 2 minutes to fully process on a 1Ghz PIII. 100Gb of log data can realistically be processed. Q: How do I configure IIS to generate detailed log files suitable for importing into ClickTracks ? A : This 2 minute tutorial will tell you all you ever wanted to know about IIS configuration. Analysis & reports
A: A page can contain several different link elements that lead to the same page. If the HREF in the link is identical in each case, there is no data from the webserver that could tell ClickTracks which was navigated by a user. A solution is to insert an unused parameter like '?ClickTracksPage=x' into each HREF on the page, and assign a different value of x to each link. ClickTracks will then see a different request. The parameter will be ignored by the web server/scripting language. Q: Why do some links on the page have no bar? A: Several possible reasons :
2. The link is a pop-up menu or similar that is generated through JavaScript after ClickTracks analyzes the page. You need to press F2 and click on the link. ClickTracks then analyses the JavaScript and locates the link on screen, mapping in the behavior data and displaying the bar. 3. The web site has been updated and the link is new and no log file data for it has yet been imported. Q: I get a lot of links all over the page and ClickTracks claims they're duplicates, but they're actually different. A: ClickTracks has incorrectly masked one or more parameters that makes these pages appear the same to the analysis engine. Hover the mouse over some links and note the parameter names shown in bottom status bar. Open the 'Dynamic Page Parameters' dialog from the 'Analysis' menu. Uncheck the mask option for each of the relevant parameters and reanalyze the data. Q: How can I count clicks on my site that lead to another site ? A: Read our mini-tutorial Labeling user populations
A: When you create a visitor label you're telling ClickTracks to separate the users who meet the specified criteria. Using the example of labeling 'users referred from a partner site' you'd create a label that identifies users with a certain referring domain. The Page Analysis report might show: Visitors that see this page:
Partner refers 78% 67 This reveals that 54% of the entire population of visitors sees this page, equaling 349 visitors during the selected period. By contrast, 78% of visitors referred from the partner see this page, representing a total of 67 visitors. Q: When I create a label, why do the values for 'all' not go down ? A: The values for 'all' represent all visitors, including any users labeled separately. Dynamic page web sites
A: ClickTracks compares pages seen in the log file with pages viewed in the built-in browser. More complex web sites use dynamic pages with parameters within the URL, like '?page=5' and '?sessionID=673456456'. Some of these parameters are used by the web site to generate different pages, while some parameters are used for keeping track of user information as they browse the site. Masking a parameter causes ClickTracks to ignore it during analysis, so in this example the '?page' parameter wouldn't masked because the value causes different pages to be generated and is important. The sessionID, by contrast, is masked out and subsequently ignored. Q: What does 'auto calc parameter masks' do? A: In the previous question we told how ClickTracks can mask certain parameters that don't produce meaningful statistical data, while retaining others that are useful. When importing the first log file ClickTracks will look at all parameters and determine if it should be masked based on statistical variance calculations.
Specific technical issues and solutions
|
