dpreview's recent interview with Fujifilm executives literally started with a comment that should be paid attention to: "Even more impressive is the lens attachment rate, as we’ve sold so many lenses as well."
What is a lens attachment rate, and why is it important?
Basically, attachment rate is the number of lenses sold versus bodies. If you sold only one lens per body, then the attachment rate would be 1. Moreover, there'd be no reason why that camera should be an interchangeable lens camera ;~).
Historically, the overall attachment rate has been around 1.65 lenses per body sold, with a low of 1.5 and a high of 1.68 over the past 12 years. Here's the pertinent CIPA chart:
But this number is extremely deceptive. One of the things I discovered in surveying tens of thousands of enthusiast digital camera users is that they upgrade bodies regularly. So consider someone who started with a Nikon D70, then upgraded to a D90, a D7100, and recently to a D7500 (that's an every-other generation upgrade). If they followed the attachment rate, they'd now have 6.4 lenses for their camera on average. What my surveys showed is that this isn't quite right. The on-going updater actually has an average of 8-10 lenses, which would imply an attachment rate of something closer to 2.
At the other end of the extreme, the more casual shooters who just think they need an ILC and then buy a superzoom to use with it don't tend to buy another lens. The superzoom stays stuck on the camera, and they upgrade far less often. I don't have as good information about this group as the regular upgraders, obviously since my sites tend to appeal to someone who is more enthusiast and trying to keep up with their equipment, not the casual consumer. But the attachment rate I measured for this group is something between 1.1 and 1.2.
Now consider the person that buys into a new mount, which happens a lot with mirrorless. Yes, there's the "adapter route" that many think they're going to take, but I don't see many actually following that usage for long. Instead, what I measure is higher attachment rates. So Fujifilm shouldn't be surprised that they've seen high attachment rates for their new medium format mirrorless camera. The only person that's going to buy that is a serious enthusiast or a pro, and they not only have a higher attachment rate to start with, but they're starting from scratch when they buy a body.
Now why does all this make a difference in mirrorless?
Olympus and Panasonic made a mount switch in 2009 (to m4/3). They filled out a full lineup of lenses (and keep doing so), but I'll bet that they see higher initial attachment rates among switchers (people moving to their mount) than they now do among updaters (people upgrading bodies every couple of generations). Sony made sort of a double switch (E in 2010, FE in 2013). With E they took a mostly consumer approach, mimicking what Canikon did (wrong) with EF-S and DX, and I'll bet that they saw a very low attachment rate, which is probably one reason why we haven't seen a lot of E-only lenses lately. With FE, I'm pretty sure that Sony has a very high attachment rate, because the A7/A9 crowd is a more sophisticated enthusiast/pro and needs new lenses.
So what about Canon and Nikon?
As far as I'm concerned, Canon has made the same mistake with EF-M as Sony did with E. Target low and you get low results.
Full frame, on the other hand, is going to be interesting. If Canon really does use the EF mount for their full frame as my sources say will happen, that changes the dynamic entirely. Canon already has an installed base of over 100m EF lenses. Canon won't get an attachment rate bump if they use the same mount.
The same is true for Nikon, as well. Their crop sensor mirrorless prototyping has been all around a new mount (mimicking EF-M). But their full frame prototyping has been both new mount (Sony choice) and old mount (Canon apparent choice).
This has to be an agonizing decision for Canikon. All those legacy lenses in closets and in use are their strength, and one reason why they dominate DSLR sales (and overall ILC, as well). Keep the mount and they enable their entire user base to pick mirrorless or DSLR without much or any of a lens penalty. But their attachment rates will be low. The number of lenses they're selling now will be about the number of lenses they're selling later, so you won't see either of them proclaiming, as did Fujifilm, to be surprised by lens sales.
On the other hand, a new full frame mirrorless mount from Canikon would mean a higher attachment rate and more new lens sales. But it would also mean that they would be five years behind Sony in terms of lens choice and supply, which means they'd still be enabling people to switch.
So which will it be? We'll know before the end of the year, I'm pretty sure. If it had been up to me to make the decision, I would have voted to stick with the current mount and just deal with the snout that the empty mirror box would leave behind, and deal with the attachment rate problem a different way (e.g. marketing).
As you might have noticed, the mirrorless cameras are getting bigger. Another quote from that Fujifilm interview at dpreview: "One purpose of us doing the X-H1 is that some customers actually requested a bigger grip and better handling, especially together with bigger lenses like the 100-400mm." Yep. When you put two pound lenses out front, you're going to need a grab handle, not a little soap bar design camera body.
That bodes well for keeping the lens mount in the Canikon full frame mirrorless world. No way do I want to put my 400mm f/2.8 on an adapter on a small body.
