Trying to set up an alert but my only option is "summary number" which is useless for what I'm trying to achieve. I have a really simple table that has:
Customer name % change for 7 day engagement % change for 28 day engagement
What I'm trying to do is have an alert sent out if either of those numbers changes by x% for a particular customer.
For example, if Acme Inc has a % change for 7 day engagement of >-15% then send an alert.
Can you attach a screenshot of your current alert configuration? I would expect you to have the ability to execute an alert on a row level, but it sounds like you've already exhausted that option so want to see exactly what configuration is failing you.
Just noticed the time stamp on your original post (6:00pm) so I am guessing you may be signed off for the day. My Dojo day shift is coming to a close so I wanted to circle back.
Depending on the number of clients you want to track, the first and easiest option would be to have separate cards for each client and set the summary number change discreetely for each card.
Alternatively, another option is to define a beast mode summary number as the max % change in the table. That would allow you to trigger the alert when any row changes by X%. (Credit: @DanB).
Lastly, you could move the calculation upstream into the dataflow and then only reveal on the card clients whose delta's met your criteria. Then your alert on the card would be when "anything changes".
Thanks for your reply! It's 500+ customers so creating individual cards won't work unfortunately.
I don't fully understand that beast mode approach that you mentioned. You're saying create a calculation like: max(`7d % Change`)?
That turns the summary % into 10 which doesn't really make sense to me...
I then set up my alert as when summary number is less than <0.15.
I don't see the option to trigger anything based on row though...
When I first started as a client, we created 30 identical pages on which the individual cards had different filters. This was back in the pre-pdp days! So no, creating separate cards is not an option.
My suggestion assumes the dataset is already at the customer level with customer name, and %change fields (or the numerator and denominator fields to calculate those %changes) populated. That is to say, my suggestion assumes you're NOT aggregating up to the customer level in a Beast Mode Calculation**. If that assumption is true, then you could create a Beast Mode calculation like this:
max([current amount]/[28 days ago amount])
Then, you could set an alert for any time that max value is larger than some threshold. Now, to be clear, this alert method will only notify you if/when at least one customer is beyond the threshold; it won't show you which one customer(s). But at least this way, the alert would direct you to consult the card and then you could identify which customer is beyond the threshold.
Hope that points you in the right direction.
**Note: the source dataset for the card would need to look something like this:
customer | current | 28 days ago | % change ---------|---------|-------------|----------- ACME Co | 50 | 40 | 0.25
Bob's Co | 90 | 100 | -0.10
That's because we can only do one level of aggregation in a Beast Mode, and that one level needs to be the max (to identify the max % change, so we can alert based on that value).