Attribution IQ Custom Lookback window

scotterthomas 21-02-2019

We recently upgraded to Adobe Analytics Prime and I have been experimenting with Attribution IQ (a very cool feature)

However, when running a standard YTD time range I had a large % of conversion (15-20) show up as none.  In looking at documentation and conferring with support, it seems the issue is the Attribution IQ lookback window, which is exactly limited to the first day of the month to the first day of the time range selected.

So in my case, the lookback window ended at Jan 1.  If I altered my time range to 12/31/18 to date the lookback window extends back to 12/1/18 and the Nones are greatly reduced, which is great except my business doesn't want YTD to be from 12/31 to present because that really doesn't make sense and adds a number (though small) of conversions that actually tool place LAST year.

Note: This is NOT the logic used by the legacy attribution metrics that still exist in Ad Hoc.

So my ask/suggestion (and a potentially pretty valuable change to the tool) would allow me to assign the lookback window along with the allocation type (first, last, linear, etc) in the Attribution IQ settings – especially since per Customer Care “the (Adobe Analytics) product is moving towards the new Attribution IQ (Analysis Workspace)” and since Ad Hoc is being phased outMarketers hate “None” and it seems quite arbitrary to cut-off the lookback window like that when clearly the data is there, as evidenced by the legacy Ad Hoc reports.

3 Comments (3 New)
3 Comments
admoseremic
Employee
25-02-2019

Hey Scott, this is definitely something we're aware of and we're currently working on a better method for more granular control over the lookback window. As a short term workaround for your use case set the reporting date range to start on Dec 31 (as you mentioned), but use a segment to filter the results to just Jan 1 - Jan X (using the orange "Day" dimension as a filter). That'll force the report to use Dec in its lookback window, while still only reporting values that happened in Jan. Perhaps not ideal, but works for now.

FWIW, we're also working on a better way to handle "None" values in reporting (allowing users to include or exclude it from their reports altogether).

Thanks for the post!

scotterthomas 26-02-2019

Thanks for the suggestion.  That seemed to help when my model was Last Touch but for any other Attribution model adding the Orange time variable the recorded conversions dropped significantly (first, linear, U).  Before adding the orange time dimension as a filter, every attribution model had the same grand total for conversions (with totals varying per channel).  After the orange time dimension filter was added, Last Touch showed true YTD total (no Dec 31) but First, Linear, and U totals all dropped significantly.  In my case, First was 30% lower, Linear 10% lower, and U 13% lower.

miikkal3149476 31-03-2019

Hi,

This is a real problem, and hopefully will be "fixed" in the future. In our case, there is f.ex. 1500 conversion events in a month, but when I use any attribution model except Last Touch, +5% of the conversions are reported as none (when using marketing channels dimension). After reading this thread I understand why it is, but we'd still need a way to attribute the exact amount of conversion to the selected time period, not so that it would be divided to couple of different months.