Most SEOs I know hate sites with separate mobile URLs. They served their purpose back in the early days of mobile browsers, but over time the maintenance overhead combined with the SEO uncertainty brought on by having two different sets of near duplicate URLs has caused many of us to advocate for migrating to a single set of URLs for both mobile and desktop. But maybe we shouldn’t be the ones advocating…
I have had the pleasure of working on several large M-dot URL migrations. In theory, these projects should be pretty straightforward – 301 redirect all the M-dot URLs to their desktop versions, maintain the redirects forever, and all is well. Too bad theory often runs into a reality brick wall.
When one argues for a big project like an M-dot migration, odds are someone with a nicer office than yours is going to ask for some proof that it’s worth doing. Here’s a handy tweet to whip out when the “Why would we want to do that?” question comes up:
Re:m-dot vs responsive/dynamic serving: I strongly recommend not using m-dot URLs if you can, especially with new sites. It makes everything so much easier. Re:PWA — they can do cool stuff quickly, but they’re also hard to get right, especially for SEO (not impossible though).— 🍌 John 🍌 (@JohnMu) September 27, 2019
Now I love John Mueller’s advice as much as the next guy, but SEO types often tend to think of themselves within the “scientist” Knowledge Graph, so we also like to see independent corroboration of our crackpot theories. Fortunately Sistrix has a handy tool to estimate organic traffic to subdomains. Let’s look at how some big e-commerce sites have fared with M-dot migrations:
If you believe the upticks were even partially caused by the M-dot migrations (and if you believe Sistrix’ data is directionally correct), then pulling the trigger on a migration seems like a no-brainer.
But as we SEOs know all too well, not everything always goes smoothly.
This is what it looks like when a large e-commerce site with an expensive IT consulting firm spends a year redesigning their site and kills their M-dot URLs without redirecting them because “it wasn’t in the original scope and would be too complicated”.
And this is what it looks like (on SEMRush) when another large e-commerce site does their M-dot migration in phases and neglects to redirect some of the old m-dot URLs. Six months to get back to where they started from.
Now I know what you’re thinking – “I run the tightest SEO ship on the planet. Those kinds of screw-ups would never happen to us.” If so, good for you, but you might do well to consider the following conversation I had with the head of SEO at one of the largest retailer sites in the U.S.
SCENE: A networking event. Everyone standing around in business casual drinking corporate craft beers.
ME: (oddly proud) I’m working on a big M-dot migration…
EXEC: (stopped cold) Dude…I pitched a big M-dot migration last year. We estimated a big increase in organic revenue post-migration. After we pushed it live, we saw basically zero lift.