Is it possible you're underestimating how long the teams burn to get their line set and get the spike off? Of the two spikes I found for Brady this year (@BUF and @TEN), the clock ticked 15 and 14 seconds for plays of 4 yards and 10 yards, respectively. On the first play, you're suggesting they would have saved 6 seconds by spiking, which means it would take only 11 seconds from snap to set to spike. On the second play, you're saying it would only take 11-12 seconds from snap to set to spike. That seems unrealistic, unless you have something to support it (I'm no expert).
I think the stronger point is the ability to reset and decide on a play, but I think you're overstating the time savings.
I admit, I could be underestimating a bit, but I don't think it's by much at all. For example, as I noted, after the first first down yesterday, I thought the Rams did a really good job of getting to the line, getting a play set and getting the snap off. They took over at 1:08, threw a 12 yard pass for first down, and then ran a quick pass for 6 yards out of bounds that ended with 45 seconds on the clock, so it took them 23 seconds to run the first down play, get to the line and run the next play, which was a quick 6 yard out that went out of bounds.
I think the Tennessee example works better than the Bills example, because in both cases, the first play began with a stopped clock (don't get me going on the Bills either, because I think the Pats should have been spiking earlier in that drive too and it cost them a chance later in the drive). In that game, Brady snaps with 21 seconds left, hits Allen for a 10 yard catch and then spikes it with 7 seconds left. So, it took the Pats 14 seconds to run the first down play and spike it.
If you estimate the 6 yard out the Rams ran took 5 seconds (which is probably conservative on the high side), that means it took them 17 seconds to run the 12 yard play and then snap, which is 3 seconds more than it took the Pats to run a 10 yard play and spike it.
It works out to about the same 3-4 seconds on the 2nd first down (although in that situation, the Rams only got 5 yards on the previous play, so presumably, it should have been faster, which is why you probably had a bunch of us in the game thread pointing out all of the milling around they were doing at the time).
As I've tried to explain quite a bit though, it's not just about the time savings. It's about that next play. Yesterday, maybe it only cost the Rams 5-10 seconds in the long run, maybe not. Maybe instead of rushing to the line and throwing a quick pass for 6 yards or 8 yards, had they spiked it and set up a good play where Goff can read the defense at the line, take his time, etc., they hit a 15 yard route on the sideline instead? Fortunately, they didn't make any huge mistakes with the clock running like we see all the time, (ie. they didn't take a sack, or penalty or turnover) until Gurley forgot to run out of bounds (which doesn't really have any bearing on the spiking conversation), so in the grand scheme of things, you can't say it cost them the game by not spiking it yesterday after 1st downs, but it certainly cost them some time, albeit maybe less that I was estimating originally, and it may have also cost them the opportunity to make a better play than what they did.
This was really just the first example I noticed in real time, but there will be more (much more egregious ones) soon enough. It'll be interesting to watch it play out, especially if the situation comes up in the playoffs.