This question was the bane of my existence when I did B2C more regularly. Thankfully the solution is pretty straightforward—but first, a caveat.
How big is your database? Just how many emails are you sending?
I ask because the mechanism to make this work relies on a separate scoring mechanism, and if you're either sending to a lot of different people at one time (think 1 million+) or if you send multiple emails (think you're a spammer and you know it), you really should make the equivalent of what I'm about to describe on nightly batches. Do not pass go, do not collect $200, do not bog your instance down with this.
So, I worked with a company with multiple overlapping departments' (and often conflicting!) goals in terms of what email messaging should be sent to users. I also needed to respect the fact that users should never get more than 2 emails a week as to not be incredibly annoying. As a result, I had multiple score fields set up: a "7 Days Email" score field and a "30 Day Emails" score field. There were further splits per department (because Department A wanted to ensure they were not monopolized by Department B, etc.) but for the sake of this explanation we'll pretend we only need a 7 Day count and a Monthly count.
Score fields are the unsung heroes of Marketo because they allow us to do math, but only to a certain extent. As such, it is imperative that you set these score fields up first in your CRM and then let the mappings flow down to Marketo. Then and only then should you set it as a score field!
So, whenever an email was sent out, each record received a +1 to their 7 Day Count and Monthly Count scores. This rolling score was then used to determine if someone had received too many emails for the week to receive any more messaging (if 7 Day Count is higher than 2, exclude!) and kept a running tally of emails for the month. There were two twin Smart Campaigns to reset the 7 Day Count every Sunday and reset the 30 Day Count on the first Tuesday of the month (so reporting could be gathered on Monday.)
For figuring out mean/median emails sent per month, the CRM was utilized and the aggregated data made splitting up not only overall email frequency but email frequency among different types of records pretty easy.
.png)