I wanted to share an analogy with anyone who might be trying to more firmly grasp some of the elementary concepts of Omniture’s SiteCatalyst analytics software. I’ll outline the meaning and intended use of:
- eVars – (s.eVarN) aka “custom commerce variables”
- s.props – (s.propN) aka “custom insight variables”
- custom events – (s.events) – “custom events”
Imagine there’s a bustling marketplace. You have many entrances, many shoppers, many browsers and in all likelihood a few shoplifters. If you were trying to better measure various metrics such as the ratio of those who enter the market to those who purchase, or even how many times people pass through a certain entrance you’d have quite a job to do. You’d have to be so many places at once with your eyes on more then you can possibly keep track of.
Cast a wide net and tag every entrance with java script tags! Custom insight variables or s.props will count whatever pass through them. Perhaps you’d like to find out the use frequency of the N entrances to the market. Let’s use
s.prop1=”entrance 1″ each time someone passes through entrance 1, and use s.prop1=”entrance 2″ each time someone goes into entrance 2. Keep in mind that although the values we’re passing through the custom insight variable vary, the actual tag (s.prop1) will stay the same for purposes of continuity. We can go into the SiteCatalyst admin and name s.prop1 “entrance frequency.”
So now you know where people are coming in…that’s nice. Say you want to know how many people go into store X in one month. Every time someone passes through the entrance point of store X you trigger a custom event, say s.events=”event1″ Now each time a shopper passes into the store the tally of event1 which will probably be set to “counter” if you’re not aggregating a monetary value the event count will increment upward by 1.
An astute reader will now say, “Jeff, what is this? You can just use a custom insight variable, why are you being all fancy using this custom event stuff eh??” Excellent point I’ll reply! I could use an s.prop and each time someone passes through the store entrance pass through the store name. So what’s the fundamental difference?
Well, when someone leaves the store it will also count them when you use a prop, whereas with the custom event you can serialize the event count so that you only increment the event once per “session” or visit. Is it better to know the number of times something passed through a gate or the # of unique people who went into the store that day? Both have a certain value (and that was a rhetorical question) but as we’ll see next, the custom event’s utility reaches further beyond that of the s.prop in the case of measuring commercial activities.
Okay, so we have the bustling marketplace, lots of people entering here and there, going into stores and we know from where they enter and which stores are most popular. The whole point of going there is to get them buying!!!! Spend your money with meeeeee says the marketplace’s collective wisdom.
Which entrance leads to the most revenue? Which store generates the most revenue? Let’s think about this…if we pass revenue into the s.prop, wait – you can’t do that it just counts the frequency of what gets placed into it, and how would you know what entrance they came into? If we set the custom event to “currency” (an advanced setting in the sitecatalyst administrative console) we can pass the revenue of each store into a custom event, but you’d have to use multiple custom events for each store. How can you get the data in 1 report.
Well first, to save face with any knowledgeable readers Omniture SiteCatalyst uses the s.products=”” tag to count revenue and product data so the custom event set to currency isn’t necessarily applicable here.
With that said, we can use the “custom commerce variable” aka the eVar (s.eVarN=”something”). A key feature of the eVar is it’s use of persistent cookies. Each eVar has its own unique attribution settings and you can control:
- How long the persistent cookie will persist for.
- Whether credit will be given to the first or last touched (first or last what, right? Whatever you put into the tag s.eVarN=”whatever is in here”).
I don’t think any cookie expiration length is typical, but attribution to the last item touched is. There is a great debate regarding the accuracy of this but in my view (without getting into great detail) I think it’s the best measurement.
So let’s set our eVar to persist for 1 day. When someone comes into the mall we have an entrance eVar, when someone enters a store we have a store eVar.
Someone comes into the mall –
s.eVar3=”west gate entrance”
Someone comes into Macy’s –
At day’s end you can run an eVar report for each custom commerce variable against revenue and you’ll see which entrance generated the most revenue and which store generated the most revenue.
In all fairness in this strange example we could use the store name as the products variable in the s.products string and eliminate the need for the eVar4 report but that’s too tangential to this discussion.
Another cool thing about the custom insight variable is you can correlate them with other traffic variables. So if I wanted to know which articles were most popular in my international news section of my news content site, each time a page was viewed in that section, the article name would be passed through the prop. Check it out:
s.propN=”9 out of 10 politicians at home and abroad are corrupt: public easily forgets”
You could then correlate channel and prop and see how popular articles were within the section.
If you need consulting shoot me an email at JeffJames.email@example.com or catch me on LinkedIn: