Loading ...
Sorry, an error occurred while loading the content.
 

RE: [webanalytics] Persisting attribution in Site Catalyst

Expand Messages
  • Warren Sander
    Crossvisitparticipation plug-in to create a stack of the internal campaigns. Needs an evar for this You could also put the internal campaign into a prop and
    Message 1 of 6 , Feb 21, 2013
      Crossvisitparticipation plug-in to create a stack of the internal
      campaigns. Needs an evar for this

      You could also put the internal campaign into a prop and use participation
      events to get visit based attribution.

      Marketing channels for first touch and last touch attribution


      -warren
      Warren Sander
      http://www.linkedin.com/in/warrensander
      warrensander@...
      508 740 9316
      On Feb 21, 2013 1:23 PM, "Blakeley, Robert (Rob)" <rblakeley@...>
      wrote:

      > **
      >
      >
      > Another different approach would be to place the click in a prop, say in
      > the format of pageName_clickID where page name was the page the link was
      > clicked on.
      > So home_action1 and widgetPromo_action1 and cartPage_purchase.
      >
      > You could then use pathing to see the sequence.
      > Participation in this scenario would surface the actions most involved, an
      > importance ranking rather than a fall out inquiry.
      >
      > (We also use the above method in combination with evars and success events
      > by copying the prop values into an evar.)
      >
      > Robert Blakeley | Analytics Product Manager | BI
      > 212.624.3854 | rblakeley@...@...>
      >
      > From: webanalytics@yahoogroups.com [mailto:webanalytics@yahoogroups.com]
      > On Behalf Of Matt Curtis
      > Sent: Thursday, February 21, 2013 10:18 AM
      > To: webanalytics@yahoogroups.com
      > Subject: Re: [webanalytics] Persisting attribution in Site Catalyst
      >
      > If you're trying to use conversion metrics, there is no clean way to do
      > that. You could use participation but that still doesn't solve your last
      > click issue (because then you're just assigning 100% of the success metrics
      > to each eVar values).
      >
      > Another idea would be to stack a series of actions a user takes in an eVar.
      > So basically take, say, a max of the last five elements a user has clicked
      > on and concatenate them sequentially using some sort of delimiter.
      >
      > ie,
      >
      > Action1:Action2:Action3:Action4
      > Action1:Action:2
      > Action4:Action5:Action6:Action7:Action8
      >
      > As you can see above, the risk here is that you're not capturing a complete
      > "path" of the "internal campaign" actions taken, since you've capped the
      > number of actions stored at 5. (this choice to limit the number of actions
      > you capture, however, is still usually a wise one).
      >
      > If you're interested in the internal campaign stacking idea and haven't
      > done something like this before, you'll need to add the Cross Visit
      > Participation plugin to your scode.js file. This article would also be
      > useful:
      >
      > http://blogs.adobe.com/digitalmarketing/analytics/cross-visit-participation-inside-omniture-sitecatalyst/
      >
      > Hope this helps,
      >
      > Matt
      >
      > On Thu, Feb 21, 2013 at 8:53 AM, dwhaltaboy
      > david.harris@...>wrote:
      >
      > > **
      > >
      > >
      > > We're currently using query strings in URLs to track clicks on certain
      > > buttons/tabs/etc. through our site (essentially internal campaign
      > > tracking). How do you handle attribution in cases like this, when
      > clicking
      > > a second element during the visit simply overrides the previous values in
      > > that query? I understand we can use multiple eVars for this - but is
      > there
      > > really a clean way to do that effectively? Any suggestions on the best
      > way
      > > to attribute particular click actions to a later conversion would be
      > > greatly appreciated.
      > >
      > > Thanks!
      > > David
      > >
      > >
      > >
      >
      > [Non-text portions of this message have been removed]
      >
      > [Non-text portions of this message have been removed]
      >
      >
      >


      [Non-text portions of this message have been removed]
    • Craig
      Just because there are ways to stack a variety of values into an eVar, doesn t mean that it makes good sense. Part of the craftsmanship that goes into an
      Message 2 of 6 , Feb 22, 2013
        Just because there are ways to stack a variety of values into an eVar, doesn't mean that it makes good sense. Part of the craftsmanship that goes into an analytics implementation is knowing what to leave out. I once inherited an implementation where an eVar was being stuffed with everything that moved on the site, and the reports were so fragmented that nobody could even look at them.

        Start by thinking about the report you would want to see, one that would help you make clearly state something about website behavior, or one that would help you make an obvious change to your website. Then determine how to protect the variable from extraneous information once it gets the data you need.

        best regards,
        Craig
      • dwhaltaboy
        Thanks Craig (and everyone else as well)! Your point touches on what I think is our biggest implementation challenge: focusing on what s actually important in
        Message 3 of 6 , Feb 22, 2013
          Thanks Craig (and everyone else as well)! Your point touches on what I think is our biggest implementation challenge: focusing on what's actually important in decision-making for our business, and not cluttering reporting, time and resources with the unnecessary.

          Lots of great ideas, I'll definitely be digging into these options. Thanks again!

          David

          --- In webanalytics@yahoogroups.com, "Craig" <craigdscribner@...> wrote:
          >
          > Just because there are ways to stack a variety of values into an eVar, doesn't mean that it makes good sense. Part of the craftsmanship that goes into an analytics implementation is knowing what to leave out. I once inherited an implementation where an eVar was being stuffed with everything that moved on the site, and the reports were so fragmented that nobody could even look at them.
          >
          > Start by thinking about the report you would want to see, one that would help you make clearly state something about website behavior, or one that would help you make an obvious change to your website. Then determine how to protect the variable from extraneous information once it gets the data you need.
          >
          > best regards,
          > Craig
          >
        Your message has been successfully submitted and would be delivered to recipients shortly.