As it stands the report seems to be of very limited use. The attribute report just
shows the overall attributes that had the highest impact on your KPI (in our
case clicks) and our list just has the Target basics all the way out to
the top 10—time of day, browser type, etc, not much actionable that a business
could do with that information.
The Segment report shows the segments but no information why the system created different
segments (in our case there were about 216 segments made by the AI
engine). Is there some way to understand why the system created different
segments? We are curious how when we only get 9,444 clicks a month on the spot where we are using Sensei, how the system can
draw statistically significant distinctions between 216 segments for those
clicks? That’s only 44 clicks per segment on average.
The report doesn’t show any relation between the segments and their component variables on
one side, and the creative variations on the other, other than going line by
line and having to parse each group of segment variables for each creative
offer separately—an almost impossible task with 216 segments (you can’t see
anything about the creative offers in aggregate). What we would
strongly recommend as an enhancement to this report is a creative focused
report, that, for each piece of different creative, shows 1.) the attributes
that caused folks to convert on that creative, and 2.) what the top segments
that clicked on that creative were, etc. And then 3.) some sort of
explanation of what causes a new segment to be created by the AI engine.