When I had the privilege to work with David Jonassen a number of years ago, his thoughts on problem solving models made an impression. He spoke, among other things, of how generic problem solving models have been shown to be too generic to actually help people solve real world problems. So, for example, telling someone to first consider possible causes isn’t really helpful.
I’ve seen ADDIE upheld as a useful basic framework for thinking about instruction. It certainly lays out some of the basics like the importance of evaluation, and it has an “industry standard” feel that helps credibility when talking to non-instructional designers (and novice instructional designers).
I wonder if ADDIE has the same issue as generic problem solving models–too generic to be useful. I saw someone recently present ADDIE by using an analogy of how he solved a home appliance problem, which I thought was unfortunate because ADDIE is definitely too generic to be a useful general problem solving model. But in terms of instructional design, it is probably best to think of it as a framework rather than a process. The focus should be on the processes one uses for analysis, design, development, implementation, and evaluation.