Actual users but not usage, those are different things.
Yes, I agree, but I would have expected a growth in users and nodes in 2023 and 2024 due to the high fees. I wasn't trying to say that I'm trying to estimate payment volume

Normally now the time would be fine to set up Lightning nodes too, with the low fees, but it doesn't seem to happen. So at this moment I think it's not unreasonable to talk about a "stagnation". It's by far not "dying", though

For me, this topic includes both users and usage. My point is that assumptions are being made about the usage based off of the user data that is public, and this would scientifically be very false since it is only a correlation which will fluctuate at times. I also gave a personal example of usage and how a single user that is myself would be counted in those stats. Obviously it would be better if we had both user and channel growth too all the time, but that is just too optimistic way of thinking. It is probably going to move in cycles related to the market sentiment and then fluctuate also related to its own things. When it comes to the description of "stagnation", I can agree with that because in a way it also reflects what is happening currently with the Bitcoin market and even more with the crypto market.
The problem is also that service providers and merchants see theses stats and they will think it twice if they want to offer a LN payment option, above all a non-custodial one.
I think that impacts more bigger providers rather than ones that were or get personally on boarded, but I agree. But this comes back also to the issue that I mention a bit, imagine that we had data to prove that payment volume right now is very high then people are making the wrong decisions by basing them on conclusions made from data that can't be used to make those conclusions.
[...]since here we are talking about people who are already using custodial wallets then there is no negative effect from enabling lightning payments on the same custodial wallets. I think that is a good way to onboard people of this kind, so they don't have to install additional wallets or open channels. I would say that most people already use custodial solutions unfortunately, but the upside is that they can be adopted to lightning in ways that are not possible with non custodial solutions.
Fair enough. I think the effect of these solutions is a bit ambiguous. The interoperability advantage is indeed quite nice, and in addition you could connect later a non-custodial Lightning node if you familiarize yourself with LN. So if the user is willing to learn, then it can lead to some "true" (i.e. "NYKNYC") adoption. I'm not opposed to it. The reasons why I don't see Lightning as the only "general" scaling solution is more due to some design issues, like the well known liquidity problem, and also the potentially slow onboarding potential if there is really massive Bitcoin adoption (realistically, 20-50 millions per year, while in theory it could be 100 million this would require using all block space for LN).
I am not even trying to say that the benefit is huge from this, it is just a good way for me to give people who use custodial solutions an ability to experience and see lightning in action. I was also personally surprised when I used it the first time with my phone, it was a fantastic experience. Can you tell me more about the liquidity problems? Do you mean in general that liquidity is not that high or something specific because I have never once had an issue making a LN payments over the years? I understand if someone wants to make a big payment it may not work because of liquidity, but that is expected behavior because they are not supposed to use it that way.