The SAMR model was not invented as a way to classify apps. This poster continues to be circulated widely among educators over the past few years. Unfortunately, it also continues to perpetuate a misunderstanding of the SAMR model.
I use this poster and others regularly in professional learning workshops as a provocation. Discussion about the possible merits of the poster and its accuracy are very instructional. However, in the end, the poster is misleading.
The SAMR model is one possible way to think about how technology is used within a given instructional design. There are ample resources, videos and presentations online about the SAMR model and how to use it, not the least of which are those of its creator, Ruben R. Puentedura, on his blog.
However, this graphic seems to suggest that the apps pictured can only be used at certain levels of the SAMR model. Nothing could be further from the truth. Also, there is no indication of an original task. You cannot classify the way an app (or any technology) is used within the SAMR model without a context and none of the apps pictured are contextualized in any way.
In fact, this graphic is one person’s arbitrary estimation of what app might be applicable at some given SAMR tier within some given context. Any one of these apps could be used at any level. It’s not which app that’s important; it’s how an app is used.
For example, I have seen the Educreations app used in a large number of different contexts and at all four levels of the SAMR model. It all depends on the original task and how the app was integrated within the instructional design. I would even go so far as to assert that each and every one of these apps could be used at any of the SAMR model tiers.
Making things... Learning Things 
Yes, yes, a million times yes.
Yesterday I held a professional development session for 4th grade teachers. We did an activity sort by SAMR level. Given 8 learning situations for students, at what level was technology used to support the learning? We then resorted those same situations by Depth of Knowledge (DOK) level. The conversation that commenced was the highlight of the day for most of the participants.
They learned that SAMR is not a target. It’s a classification system. Teachers need to know that there is a time and a place for S, A, M, and R in the classroom. None of us use technology at the Redefinition level all the time. Nor should we. Likewise, teachers discussed the fact that a project in the Redefinition zone does not automatically equal enagaged, higher level learning. Case in point: A teacher spent a month with students creating a movie about the solar system. Green screen, voice over, time lapse… you name it, and that video had it. However, the content was all DOK level 1, recall of information. Students spent a month on Google-able facts. A lot of valuable learning opportunities were replaced by a shiny object.
These are important conversations to have.
Thanks for your thoughtful comment, Laura! As you said, the critical use of models such as SAMR is imperative. I also try to model that stance, too. The redefinition level is cool so long as the learning tasks involved are what you really want in your instructional design. Is the “R” level really accessing and generating the learning outcomes you wanted? Is the “R” level task obliterating skills or knowledge that you need to assess? I like your inclusion of knowledge domain analysis which brings up another good point – SAMR is just another thinking/planning tool to be used alongside others in order to help educators create the most effective instructional designs.