Optimize the Time Parting Plugin to get More Detail and Use Less Variables

The Time Parting Plug-in is one of the more popular SiteCatalyst plug-ins available. A standard implementation of the Time Parting plug-in will consume 3 variables. One for Time of Day, one for the Day of Week, and one for Weekday/Weekend. How can we improve this to get more information, and more importantly use less variables? Here is how I have been doing it.

I use a combination of stacking the variables and SAINT uploads. For those of you who are not familiar with SAINT, Omniture describes it as, “…an acronym for SiteCatalyst Attribute Importing and Naming Tool. This tool enables you to download the classifications template, apply attributes to it, and then upload the data, thereby enhancing your SiteCatalyst reports with the new attributes.” This will allow us to upload a lot of detail about any variable you record.

Here’s how I’m doing it on this site. First I am using the 2.0 version of the plug-in and not the 1.4 version that I describe in a previous post. The 2.0 version includes support for Daylight Savings time and globalizes the year. You can find the 2.0 version from the SiteCatalyst Knowledge Base. If you prefer to use the 1.4 version, you can find it on this site.

/* Set Time Parting Variables */
s_hour=s.getTimeParting('h','-5');
s_day=s.getTimeParting('d','-5');
s_timepart=s_day+"|"+s_hour;
s.prop16=s_timepart.toLowerCase();
if (s.visEvent) s.eVar16=s.prop16;

Ok, let me explain whats going on here. As I said before the Time Parting plug-in captures 3 variables. If you notice in my code I am only using two of them. I don’t need to capture Weekday/Weekend anymore. I will take care of that later. The other two, I capture in two blank variables I created, s_day and s_hour. Next I combine the two of them in a single variable I call s_timepart, separated by a pipe. Then to ensure everything is consistent I copy the variable in all lower case to the prop that I am going to use. This next part is a little different. In the eVar I only want capture this value once per visit. Typically a simple getValOnce will be enough to get it done. Well then what happens when the visit extends from one time part into another? In that situation the Time Parting value will be different and therefore getValOnce will capture this as a new value since it has changed. I don’t want that to happen, I only want it once per visit. So this is when I tie in using the get Visit Start plug-in. This guarantees I will only capture the value only one time per visit.

This will return a report that looks like this:

We now have a total of 672 possible options in this report. The next thing we want to do is to classify these using SAINT. I set up 5 different categories to use. Weekday/Weekend (this is why we don’t need to capture it in the code, Day of Week, Hour of Day, Hour Part and AM/PM.

I then created the template to use that contains all of these values.

You can download a copy of the template that I use here.
Upload the template and that’s all there is to it. Do you have more conversions in the bottom of the hour or the top of the hour? How about morning vs afternoon? Which whole hour is the most profitable? Now you have an easy way to break down your time parting with finer granularity, at the sime time saving your self a couple of variables.

Enjoy!

Popularity: 6% [?]