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

Re: Conversion Funnel Measurement Methodology

Expand Messages
  • jwb96
    There s really no use or sense in putting events on your checkout pages. Events are only measured in instances, which is a poor metric for what you re trying
    Message 1 of 4 , Feb 2, 2006
      There's really no use or sense in putting events on your checkout
      pages. Events are only measured in instances, which is a poor
      metric for what you're trying to do. They're also success metrics,
      and really the only success is the order.

      The simplest thing to do, assuming that your checkout process
      contains specific pages, each of which has one, single unique
      pagename, is to use the fall-out report and select each page in your
      process. This report is meant for this type of reporting.

      Now if you have some pages that are optional, just ignore them.
      Only choose the required pages in your flow. For instance, some of
      our customers have to go to a ship-to page if shipping is different
      than billing, but since customers don't have to go through this page
      we ignore it in any reporting. We can look at the exits from this
      detour in the checkout process separately if needed. The fallout
      report does not require each page to be in uninterupted sequence - a
      person can leave the funnel and re-enter and not be double counted.

      Another option, if your pages don't have a single unique pagename
      for each step, is to assign a second s_prop to those pages and code
      it so that each step in the flow gets only one unique value. You
      can then enable pathing on this s_prop and do as I described above.

      Also, again assuming your checkout process is linear, you don't need
      to use the fallout report. Just knowing the visits to each step
      will allow you to calculate the success or abandonment from each
      step.

      Good luck,
      Jim

      --- In webanalytics@yahoogroups.com, "juliusodian" <JODIAN@...>
      wrote:
      >
      > Hi, I'm new to this group, I manage the e-commerce marketing group
      for
      > http://www.falabella.cl We are the largest department store
      retailer
      > in South America with a big presence in Chile.
      >
      > Anyway, we are in the process of finishing the installation
      Omniture
      > Site Catalyst on our site and I'm having some challenges making
      sense
      > of the conversion funnel numbers.
      >
      > Our inicial installation used the s.events on each of the checkout
      > pages, but I found that this overstated abandonment because people
      > move in and out of the shopping cart. Furthermore, the default
      setting
      > is to measure pageviews. So I asked my implementation manager to
      use
      > the getValOnce function so that we would count only one visit per
      > session. This has helped quite a bit, but for reasons that I still
      > fail to understand, the number of people who arrive at our order
      > thankyou page is higher than the order confirmation page. There is
      no
      > way to reach this page without going to the confirmation page
      first.
      > Second of all, we are using are passing the order ID to prevent
      > duplicates.
      >
      > I would apreciate any suggestions/wisdom as to why this might be
      > happing and how we should track abandonment in a more accurate
      way. At
      > this point, I don't know what I can believe.
      >
      > Thanks,
      >
      > Julius
      >
    • juliusodian
      Jim, Thanks for your response. I have a couple followup questions: Suppose our checkout process is: 1. Cross Sell Page 2. Shopping Cart 3. Login 4. Shipping 5.
      Message 2 of 4 , Feb 2, 2006
        Jim, Thanks for your response. I have a couple followup questions:

        Suppose our checkout process is:

        1. Cross Sell Page
        2. Shopping Cart
        3. Login
        4. Shipping
        5. Payment
        6. Confirmation
        7. Order Thankyou

        During a certain period of time range we found the following:

        560 orders.

        During that same time range, using the Fall-out report, we see:

        1. Cross Sell Page
        2. Shopping Cart
        3. Login
        4. Shipping
        5. Payment -------- 560 visits
        6. Confirmation --- 383 visits
        7. Order Thankyou - 371 visits

        My first question is why there is such a big difference between the
        number of orders we have and the visits on page 7. To place an order,
        its necesarry to traverse each an every page listed above. It is
        possible to start from the Shopping cart page in a second session
        given that we save the contents of the shopping cart between sessions
        if the user has logged on. I don't understand why I get distinct
        numbers when I add the Shopping Cart page to the analysis. (For
        technical reasons, the shipping page is not tracked.)

        Shopping Cart page added to Fallout Report
        1. Cross Sell Page
        2. Shopping Cart -- 1407 visits
        3. Login
        4. Shipping
        5. Payment -------- 544 visits
        6. Confirmation --- 372 visits
        7. Order Thankyou - 360 visits

        Also, if the visits don't add to 560, I'm not working with complete
        data, for that reason, I had chosen to use success events. I'd love to
        know where I'm going wrong.

        Thanks,

        Julius


        --- In webanalytics@yahoogroups.com, "jwb96" <jwb96@...> wrote:
        >
        > There's really no use or sense in putting events on your checkout
        > pages. Events are only measured in instances, which is a poor
        > metric for what you're trying to do. They're also success metrics,
        > and really the only success is the order.
        >
        > The simplest thing to do, assuming that your checkout process
        > contains specific pages, each of which has one, single unique
        > pagename, is to use the fall-out report and select each page in your
        > process. This report is meant for this type of reporting.
        >
        > Now if you have some pages that are optional, just ignore them.
        > Only choose the required pages in your flow. For instance, some of
        > our customers have to go to a ship-to page if shipping is different
        > than billing, but since customers don't have to go through this page
        > we ignore it in any reporting. We can look at the exits from this
        > detour in the checkout process separately if needed. The fallout
        > report does not require each page to be in uninterupted sequence - a
        > person can leave the funnel and re-enter and not be double counted.
        >
        > Another option, if your pages don't have a single unique pagename
        > for each step, is to assign a second s_prop to those pages and code
        > it so that each step in the flow gets only one unique value. You
        > can then enable pathing on this s_prop and do as I described above.
        >
        > Also, again assuming your checkout process is linear, you don't need
        > to use the fallout report. Just knowing the visits to each step
        > will allow you to calculate the success or abandonment from each
        > step.
        >
        > Good luck,
        > Jim
        >
        > --- In webanalytics@yahoogroups.com, "juliusodian" <JODIAN@>
        > wrote:
        > >
        > > Hi, I'm new to this group, I manage the e-commerce marketing group
        > for
        > > http://www.falabella.cl We are the largest department store
        > retailer
        > > in South America with a big presence in Chile.
        > >
        > > Anyway, we are in the process of finishing the installation
        > Omniture
        > > Site Catalyst on our site and I'm having some challenges making
        > sense
        > > of the conversion funnel numbers.
        > >
        > > Our inicial installation used the s.events on each of the checkout
        > > pages, but I found that this overstated abandonment because people
        > > move in and out of the shopping cart. Furthermore, the default
        > setting
        > > is to measure pageviews. So I asked my implementation manager to
        > use
        > > the getValOnce function so that we would count only one visit per
        > > session. This has helped quite a bit, but for reasons that I still
        > > fail to understand, the number of people who arrive at our order
        > > thankyou page is higher than the order confirmation page. There is
        > no
        > > way to reach this page without going to the confirmation page
        > first.
        > > Second of all, we are using are passing the order ID to prevent
        > > duplicates.
        > >
        > > I would apreciate any suggestions/wisdom as to why this might be
        > > happing and how we should track abandonment in a more accurate
        > way. At
        > > this point, I don't know what I can believe.
        > >
        > > Thanks,
        > >
        > > Julius
        > >
        >
      • jwb96
        If you don t use the fallout report and just use visits from most popular pages, how many visits do you record for the order thankyou page? 560? That would
        Message 3 of 4 , Feb 3, 2006
          If you don't use the fallout report and just use visits from most
          popular pages, how many visits do you record for the order thankyou
          page? 560? That would be my first sanity check. And is the 560
          orders number from your backend system or from Omniture? If you can
          confirm that simple visits to that page is similar to the number of
          orders recorded, then I'd talk to your account rep and explain the
          issue to see what could be going wrong. However, if just visits to
          that page don't match the order count you'd expect, then I'd be
          suspicious of whether the pixel is sending info, is too low on the
          page, etc.

          Another way to say it is, if you just get visits to each of the
          pages in the process from the most popular pages report, does that
          look more reasonable? Since your flow is predetermined, you can
          calculate the abandonment and success from each step yourself.

          Jim

          --- In webanalytics@yahoogroups.com, "juliusodian" <JODIAN@...>
          wrote:
          >
          > Jim, Thanks for your response. I have a couple followup questions:
          >
          > Suppose our checkout process is:
          >
          > 1. Cross Sell Page
          > 2. Shopping Cart
          > 3. Login
          > 4. Shipping
          > 5. Payment
          > 6. Confirmation
          > 7. Order Thankyou
          >
          > During a certain period of time range we found the following:
          >
          > 560 orders.
          >
          > During that same time range, using the Fall-out report, we see:
          >
          > 1. Cross Sell Page
          > 2. Shopping Cart
          > 3. Login
          > 4. Shipping
          > 5. Payment -------- 560 visits
          > 6. Confirmation --- 383 visits
          > 7. Order Thankyou - 371 visits
          >
          > My first question is why there is such a big difference between the
          > number of orders we have and the visits on page 7. To place an
          order,
          > its necesarry to traverse each an every page listed above. It is
          > possible to start from the Shopping cart page in a second session
          > given that we save the contents of the shopping cart between
          sessions
          > if the user has logged on. I don't understand why I get distinct
          > numbers when I add the Shopping Cart page to the analysis. (For
          > technical reasons, the shipping page is not tracked.)
          >
          > Shopping Cart page added to Fallout Report
          > 1. Cross Sell Page
          > 2. Shopping Cart -- 1407 visits
          > 3. Login
          > 4. Shipping
          > 5. Payment -------- 544 visits
          > 6. Confirmation --- 372 visits
          > 7. Order Thankyou - 360 visits
          >
          > Also, if the visits don't add to 560, I'm not working with complete
          > data, for that reason, I had chosen to use success events. I'd
          love to
          > know where I'm going wrong.
          >
          > Thanks,
          >
          > Julius
          >
          >
          > --- In webanalytics@yahoogroups.com, "jwb96" <jwb96@> wrote:
          > >
          > > There's really no use or sense in putting events on your
          checkout
          > > pages. Events are only measured in instances, which is a poor
          > > metric for what you're trying to do. They're also success
          metrics,
          > > and really the only success is the order.
          > >
          > > The simplest thing to do, assuming that your checkout process
          > > contains specific pages, each of which has one, single unique
          > > pagename, is to use the fall-out report and select each page in
          your
          > > process. This report is meant for this type of reporting.
          > >
          > > Now if you have some pages that are optional, just ignore them.
          > > Only choose the required pages in your flow. For instance, some
          of
          > > our customers have to go to a ship-to page if shipping is
          different
          > > than billing, but since customers don't have to go through this
          page
          > > we ignore it in any reporting. We can look at the exits from
          this
          > > detour in the checkout process separately if needed. The
          fallout
          > > report does not require each page to be in uninterupted
          sequence - a
          > > person can leave the funnel and re-enter and not be double
          counted.
          > >
          > > Another option, if your pages don't have a single unique
          pagename
          > > for each step, is to assign a second s_prop to those pages and
          code
          > > it so that each step in the flow gets only one unique value.
          You
          > > can then enable pathing on this s_prop and do as I described
          above.
          > >
          > > Also, again assuming your checkout process is linear, you don't
          need
          > > to use the fallout report. Just knowing the visits to each step
          > > will allow you to calculate the success or abandonment from each
          > > step.
          > >
          > > Good luck,
          > > Jim
          > >
          > > --- In webanalytics@yahoogroups.com, "juliusodian" <JODIAN@>
          > > wrote:
          > > >
          > > > Hi, I'm new to this group, I manage the e-commerce marketing
          group
          > > for
          > > > http://www.falabella.cl We are the largest department store
          > > retailer
          > > > in South America with a big presence in Chile.
          > > >
          > > > Anyway, we are in the process of finishing the installation
          > > Omniture
          > > > Site Catalyst on our site and I'm having some challenges
          making
          > > sense
          > > > of the conversion funnel numbers.
          > > >
          > > > Our inicial installation used the s.events on each of the
          checkout
          > > > pages, but I found that this overstated abandonment because
          people
          > > > move in and out of the shopping cart. Furthermore, the default
          > > setting
          > > > is to measure pageviews. So I asked my implementation manager
          to
          > > use
          > > > the getValOnce function so that we would count only one visit
          per
          > > > session. This has helped quite a bit, but for reasons that I
          still
          > > > fail to understand, the number of people who arrive at our
          order
          > > > thankyou page is higher than the order confirmation page.
          There is
          > > no
          > > > way to reach this page without going to the confirmation page
          > > first.
          > > > Second of all, we are using are passing the order ID to prevent
          > > > duplicates.
          > > >
          > > > I would apreciate any suggestions/wisdom as to why this might
          be
          > > > happing and how we should track abandonment in a more accurate
          > > way. At
          > > > this point, I don't know what I can believe.
          > > >
          > > > Thanks,
          > > >
          > > > Julius
          > > >
          > >
          >
        Your message has been successfully submitted and would be delivered to recipients shortly.