Top critical review
This book is worth reading, but is highly over-rated.
Reviewed in the United States 🇺🇸 on July 28, 2013
Every now and then I'll find a book such as this one, that has stellar reviews but that in no way lives up to its hype. This is not even close to a 5 star book.
My first criticism - the first 120 pages are about overcoming organizational resistance to Scrum.... no exaggeration, 1-2-0 pages on this. I didn't find any of this useful. This material could have been easily summarized in 10 pages at most.
The next 200 pages get a lot better. Cohn discusses roles unique to Scrum, good technical practices, good team practices etc. I liked most of this - especially recommendation that teams should be feature driven vs component driven. Meaning that teams should be organized by functionality devilvered to the end-user vs. by layers of platform or application they are designing. This creates a more integrated product.
I also like his discussion on automating testing - which I 100% agree with. Automate early and often.
I didn't quite agree with his analogy that the product owner = traditional product manager, and ScrumMaster = traditional project manager. I've seen other interpretations of scrum where this is not the case.
My more serious criticism is that Cohn does not go into enough detail regarding flushing out user requirements. He doesn't really go further than to say the Product Owner is external facing and should figure out what the user wants. I think getting the user requirements right is one of the most critical jobs of a product development team, and I think discussion of this was sorely lacking in the book given all of the other topics he touched on.
One other thing I don't agree with , the use of user stories in place of requirements. I have not read his other book on User stories, but I have read several other books on requirements and I do not think it's a good idea for development teams to basically talk through what should be in a product. Important features should be written requirements that can be tested against and have only one interpretation. This is especially true for products have both hardware and software components.
The examples he gave where written statements are mis-interpreted are laughable. No one would write requirements like the ones he uses in examples. To be clear, I'm not advocating waterfall and I do agree though that defining requirements at the onset of a project is impossible. I just don't agree with the user story approach.
The last 100 pages discuss mostly the issues of implementing Scrum across the entire organization. Some of this deals with distributed teams, HR, facilities, etc. I didn't find most of this useful, aside from a point Cohn made regarding annual reviews being based off of team work rather than an individual assessment. He's definitely right about that.
Overall, I felt Cohn had a somewhat dry and long-winded manner of writing. This book (~450 pages!) could have easily been about 200 pages shorter. He writes from the perspective of someone that coaches and helps organizations adopt Scrum - and this comes across very much in his writing.
He does NOT write from the perspective of someone on a team actually developing a product, and the book suffers for this.
I'd give this probably 3.5 stars. Not a bad book, some useful sections in the middle, but not a great book either. I may look into Cohn's book on planning, because his brief section on this was good.
If you'd like to learn more about the nuts and bolts of implementing Scrum, I much preferred Exploring Scrum by Rawsthorne and Shimp and would point agilists to that book instead.