Phase gates are evil
April 17, 2018
The one single phrase I remember best from my first Scaled Agile Framework training was the quote of Dr Alan Ward: “Phase gates are evil”. I found this statement both very funny and incredibly liberating.
As a project manager, I have scrambled through project phase gates for several years. This was a particularly important rite of passage at clients who had their own project management methodology with clearly defined phase gates. It was not only a matter of project quality assurance, but a matter of pride in the company’s way of working, that phase gate criteria were fulfilled to the letter before the project was allowed to proceed.
You can probably read between my lines that I didn’t always appreciate this dedication. I have two specific issues with phase gates, and while I don’t personally take quite as strong a stand as Dr Ward, I do think we need to be very careful if we decide to require phase gate approvals in projects.
First, phase gate approvals create unnecessary idle time for a team. The last product missing for the approval is holding the entire team back, and calendar time from the decision makers can be difficult to obtain. A smart and resourceful team can certainly work around that to some extent, but that means we’re working through loopholes in the structure – rather than the structure supporting the work, which is the whole point of having a project methodology.
Second, quality assurance impact of phase gates is near negligible. Quality needs to be worked into the product ahead of time, not scrambled together at the last minute for a gate approval. Quality criteria and definitions of ready and done (DoR and DoD) that are specified clearly in advance are much more effective for quality assurance.
On the other hand, most projects have some sort of phases, at minimum the initiation stage and delivery stage. If your project has multiple stages of differing activities, it is sensible to recognize them and keep the team and management informed about the big picture. Management needs to be informed of the team’s progress, and needs to be able to inspect the quality of the work done. This need is better served by regular checkpoints that are in everyone’s calendars well ahead of the time, rather than formal gate approvals.
It’s necessary to recognize that the other extreme also exists: projects that are initiated and pushed forward without adequate control and structure, and fail to deliver their expected value. These situations are best avoided by having clear content and quality criteria for project deliverables, a solid understanding of dependencies, and continuous review of the business justification for the project. A “fail fast” mind-set needs to be applied: we need to have courage to stop a project if expected content and quality is not achieved. This can be done at any time, not only at phase gates.
What do you think? What are your experiences on project phase gates?