Functions in calcualted metrics - same daily number each day

[Meghan_Powers]

Hello,
I've created a standard deviation calculated metric for content velocity, which is pageviews/visits.  When I pull that into workspace and look at it daily, it has the same value every day.  I'm guessing it's because I have a function for a metric divided by another metric and so it's just taking the average for each day.
I know I'm missing something obvious, but how can I see that calculated metric by day?

[Jennifer_Dungan]

I haven't used the Standard Deviation formula much in Adobe, but I know that this calculation is about the expected deviation for a group of numbers.. so I would expect the entire range of data (the set) to share the same STDEV.

Let's look at a small set of data:

4.37, 4.19, 3.85, 4.01, 3.97, 4.02, 4.12

There are 7 numbers (1 week's worth for this example).

The Sum is 28.53

The Mean is 4.0757142857143

Now, we take each number, subtrack the mean and square the value, and sum all the values using this: 0.17117142857143

Now, divide that by 7 (for the 7 unique values): 0.02445306122449

Then find the square root of that: 0.15637474612126

So, rounded to 2 decimal places, the standard deviation is 0.16, which is what Adobe shows me:

Now, of course, this doesn't add any "Margin of Error" (Confidence Interval). If you need a 95% confidence rate, you would need to create a calculated metric to convert your STDEV into a STDEV95 or whatever confidence you require). But based on this alone, for the 7 day period, the standard deviation is ±0.16 from the mean.

So here, I show the mean, and the "High" and "Low" thresholds, I can see that Nov first and third are outside of the standard deviation. I could probably do another column using nested IF statements to check the PVs / Visits against the high and low.

If the PV/Vs is higher than the high, then return 1, if not, check if it's less than the low, return -1, otherwise return 0... then hide the resulting numbers and use conditional formatting to show red, green and yellow (green is higher than the deviation, red is lower than the deviation, and yellow is in the range) - or something like that... (I did not do this cause it required a lot of nested logic)

I hope this helps.

[Meghan_Powers]

Thank you, Jen, as always, for your insightful commentary!  I'll take a look through this!

[Jennifer_Dungan]

Yeah, STDEV is a little complex... and I think to really make use of it, you have to use some other logic around it (like to find your upper and lower thresholds)... though I'm sure more "mathy" people probably have some amazing additional uses for this

Good luck.