I've always benefited in this area of bike riding by living in hilly places. My regular bike commute for years featured a mile long 12% grade near the end and a couple other lesser grades along the route (17-18 miles). Where I live now, there are roads up to 20% (short and steep, those).

Back in the "good old days" of a bike with two chain rings and a 5-speed freewheel, the lowest gear I could get was 27". We used that low gear to do many, many tours in the mountains of the western US (the long 7% grades of the Icefield Parkway in the Canadian Rockies, several times, for example).

We have modern bikes, now. We are also several decades older (I'm closing in on 60). With a 44-32-22 crank set and a 9-speed cassette of 11-32, that gives a low gear of 19" (or, so).

On recent tours (again, to places with mountains, the French Pyrenees, up some of those cols made famous by the Tour de France), I have usually not gone lower than the 25" gear. I still have 21" and 19" in reserve. My wife has make happy use of the 19" gear. I get to do a LOT more riding than she does, so have a much higher level of fitness.

Honestly, though, those (21" 19") are such low gears that you'd only be riding at about 3-4 mph, max.

Your fitness level, excess weight you might be carrying, how much "stuff" you plan to haul along, etc., will have a bearing on your low gears. But, really, a low of something in the 25-27" range should work for most roads. If you KNOW you'll face something much steeper than 8%, you might want something a bit lower.

A lot of hills have the percent of the grade marked on a road sign, especially for trucks on the downhill. By now, I know what it feels like to drive and ride a 7-8% grade. That is a common max on highways. That gradient max is out of considering of truck traffic. Smaller roads in the mountains of the western US are often at a higher gradient. But, you kind of have to hunt for those.

Why two bikes? Why not two cassettes, and change the chain, if necessary if there is a big difference in the max/min tooth count between the two?

Back in my "good old days" I would change out the freewheel (the old term for what we now call a cassette). Actually, what I would do was change the cogs. For touring, with only five cogs, I dumped the high gears and favored cogs that gave me a good spread of mid and low gears. When I am touring, I am not going to pedal down a hill fast enough to have any use for high gears over 85". If I am on a down hill, I will enjoy the coast.

From the criteria you are dithering over, I would suggest you want good shifting. I would also suggest not obsessing about it. You can probably get by with a low of 27-25". If you are afraid of finding grades steeper than 8%, or don't feel you are in very good shape, (etc), opt for a lower low. Dump higher gears, if you have to. They are basically useless for loaded touring.

One thing to mention about inclinometers. There is a bubble type, that I thought I'd try out. Seemed like a good idea. I found that is had a "flaw" in use. It gets bitten by physics. If one is climbing a hill, one is accelerating versus gravity. With a bubble inclinometer, when you accelerate the bubble/fluid, the bubble rises, being lighter than the oil. So, the bubble inclinometer will measure a higher gradient than is real -- unless you are standing still. I first realized this on a test ride as I was pedalling up a, oh maybe, 2% rise (I had accurately leveled the inclinometer before setting out) and it said the gradient was 4% -- as steep as another hill I ride, where I has used a tape measure and a carpenter's level to find that was 4%. This measured hill is visibly steeper than the little rise. The bubble inclinometer was measuring inaccurately because I was pushing up the ride. If I stopped pedalling and coasted, no longer accelerating, the bubble settled down to the real gradient.

Then riding up a steep hill, hear our house, I found that with each pedal stroke the bubble would ride and read a higher gradient. The gradient of that hill is too steep for me to try coasting, very much, but when I did, the bubble sank back toward the true gradient of that climb. The variability of the reading has nothing to do with steering (as I once saw someone suggest).

There is a German made cyclometer I am thinking about that has an altimeter and inclinometer feature. It may work more accurately than a bubble type inclinometer on a moving bike. Pricey, though.

Cheers