I tend to agree with Blurymind.
In my opinion, a software rental scheme (so-called 'subscription', which it is not!) only works well if
1) your sofware is the industry standard (people depend on the software for their living), and/or
2) it is the best in class, and/or
3) it is unique/fills a niche that no-one else offers, and/or
4) it offers functionality no competitor delivers.
Compare to Adobe:
1) industry standard? Check.
2) best in class? Mostly, yes. Certainly on a professional level. Check.
3) Unique? At a professional level, yes, mostly.
4) offers functionality no competitor can deliver? Yes, for a large part, in particular for professionals again.
1) Nope, C3 still has to proof itself. C2 is not the industry standard either - not by a long shot.
2) up to a point, as a visual editor, perhaps. But this is marred by its Achilles' heel: lack of native export, and other game dev environments just offer more features (animation timeline, for example!).
3) No, Construct is not the only visual game editor in town. With competitors improving this aspect (Fusion 3, Godot, Unity&external plugins, ...) it is not unique here. The web export is outstanding, though. Trouble is, the competitors also provide web export, aside from native export.
4) No, competitors deliver equal or more functionality at this point.
Here's the rub: as Blurymind mentioned, software rental works well for professionals. If Construct 3 would be aiming at that segment of the market, I believe it might do well. The trouble, though, is that Construct 2 isn't really part of that market. It is mainly small developers, freelancers, and hobbyists for whom Construct 2 is an attractive proposition, and that is how Scirra are marketing their tools.
Switch to a software rental business model, and I am pretty sure a large (if not the majority) of Construct users will leave for alternatives - or at the very least consider a switch. The market for game engines is just too open, and I can see commercial competitors such as Clickteam rubbing their hands together right now, and I assure you they will offer Construct 2 license holders a cross-update when Fusion 3 comes out later this year.
On the other side competitors like Unity, Godot, Unreal all offer excellent FREE options for the market Scirra is operating in currently. Why pay the rent for software that is out-classed in most departments (excepting perhaps the 'easy' visual scripting) when so many free escape routes exist? Speaking for myself, I am now teaching myself Godot, and will look into Fusion 3 when it comes out. (Godot is actually quite an amazing tool.)
Anyway, I just can't see this work out for Scirra. Perhaps I am a cynic.
Agreeable *slow clap clap clap*. I don't think you're cynic, the arguments are rational. The really question now is what will Construct 3 free version cover? If the free version can satisfy most of the users (like unity most casual devs are ok with unity free) and Scirra will only target cost saving features, specialized features or of sorts, then Construct 3 can really put them in a better position.