Fun with Data: Ranking mailbox providers in .edu

Time for fun with a smaller data set! In my most recent snapshot of the top ten million domains list, just over 4200 of them are domains ending in the ".edu" TLD. The .edu top level domain is a small one, with nowhere near the many millions of domains registered under the .com TLD. (Wikipedia suggests that between 7000-8000 domains exist under .edu at any given time, but the linked data supporting that point is from 2015.)

Domains registrants for .edu domain names are required to be United States-affiliated institutions of higher education, though there's a good chance that domains registered before this rule was put into place in 2001 are likely still allowed to exist, grandfathered in. As far as I can tell, the entities in my domain data snapshot all seem to be US institutions of higher education.

So who hosts email inboxes for the faculty and students of these educational domains? When I roll the data up based on where a domain's MX records point to, here's what I find:

• Microsoft: 41.62%
• Google: 21.68%
• Proofpoint: 5.58%
• Barracuda: 5.04%
• Cisco: 2.11%
• Mimecast: 1.87%
• Sophos: 0.30%
• Other: 21.80%

Were you as surprised as I was to find that Microsoft is more represented in this space than Google? I was. The data suggests that Microsoft likely leads the way when it comes to hosting end user mailboxes for institutions of higher learning here in the US. Of course, that gets a bit fuzzy given that services like Proofpoint, Mimecast and Barracuda can be "in front of" mailboxes ultimately hosted by Microsoft or Google, but even if all ~12% of those domains had mailboxes ultimately hosted by Google, Microsoft would still lead the way by number of domains.

When it gets down into the "other," what's interesting to me there is that just about all of the "other" are domains where the institution hosts their own mail. Or at least, the MX records are configured in a way to point at servers under their own domain. This suggests that perhaps outsourcing of hosted user mailboxes is something that is not something that all schools engage in. "Self hosted" effectively becomes the third-highest ranked mailbox provider in the educational space, making this likely a bit of a wild west when it comes to deliverability and reputation management. Lots of individual endpoints, each governed individually, nothing obviously showing them as coordinating processes or even platforms.

I feel a little bad for Sophos. While they are represented (lightly) in the data, Google Sheets didn't even deign to label that tiny slice of the pie.