Median function in calculated metrics

[matthewp2989885]

Hi,

I'm trying to create a metric which shows the median purchase value. The range of sums can be quite large, so the mean can be skewed by large orders reducing its usefulness.

I have tried using the median function but it isn't behaving as I would expect. What I would ideally like to do is have a median value metric and be able to apply this to a number of different dimensions. So for example have a table of products (A,B,C) and the value to be the median value spent on each. Then break this down by client and see the median value each client placed into products A B or C.

At the moment the median metric I have created seems to be working for time ranges - so it'll give me the correct figure for a particular day, but is not working when other dimensions are introduced (the same median value is given to all the elements in the dimension). I have even tried putting segments within the calculated metric, but this then (bizarrely) it just returns the total sum of the values!

I'm guessing my issues are because Adobe is looking at it on a column basis, but I'm hoping someone has a neat work around or suggestion.

Thanks

Matt

[IvanAndrus · Accepted solution]

This isn't currently possible, but there is an open idea, so you should vote for that (if you haven't already).

[fsakjlfdsakljfd]

Have you considered instead of evaluating a median, evaluating a % of orders over a certain threshold?

For example you could export your revenue at transaction id level for a given month via data warehouse, use excel to derive a median, and then use that as your benchmark.

(Visit: Revenue > 200 & Orders Exists [visits]) / (Visit: Orders Exists [visits]).

Although not exactly what you are hoping to find it will give you similar directional value.