I dive in Web Analytics and figure out some questions. Please, help me to find answers.
All my questions will be in terms of classical tools for web analytics such as Google Analytics.
Can we replace them?
As I know, when we look at web log, it contain information such as IP, time, browser, action and etc., but for powerful analytics we need much more.
1. I saw presentation about Splunk Web Analytics, and there was advantage of Splunk - Splunk doesn’t use tags and you don’t need to adjust your web site. But in this case how can I get information about following events:
For example, my company has external banners, how can I get information about people who just saw my banner? (impressions) And maybe some more information than just impression, such as customer interest.
Can I get information about customer behaviour? For example, he doesn’t make any clicks but he moves and scrolling mouse?
2. Funnel chart is very popular in order to visualise conversion, can I build it in Splunk?
3. How can I analyse visitor journey and visualise it?
4. How can I compare two periods in Splunk? For example Yesterdaty with Day before Yestrday.
A lot of your questions can be answered by going through Splunk's documentation. It sounds like you have yet to download and play with Splunk. The best I advice I have for you as of now is to go through the Search Tutorial documentation from start to finish and once you have your basic questions answered, then post more specific use case questions you have here on Answers.
So, the answer to your first few questions lies in what source data you have available to you. Splunk is an amazing tool to index and correlate basically any text based event log. So, for example, you could use Splunk to calculate impressions of your banner AS LONG AS you had some way to access the data.
As for being able to create funnel charts, visualize your visitor journey, or compare two periods -- you can do all those things in Splunk but they are not "out of the box" features in the way you might find them in Google Analytics.