Welcome to Processland and its inhabitants—the Evil Phase Gates! ??

Are Phase Gates Evil? ??

The thought-provoking expression, 'Phase gates are evil,' attributed to Dr. Allen C. Ward, has lingered with me since my time as a former trainer on the Scaled Agile Framework (SAFe), where I first came across the quote.

While I don't believe phase gates, stage gates, or checkpoints are genuinely evil, they can quickly tip into a scenario reminiscent of a tractor-pulling event, rather than increasing the likelihood of successful implementation and elevating the quality of an idea as it enters the Approval Process.

After nearly two decades in diverse architectural roles and participating in authorisation and accreditation processes, I must confess that I have unintentionally played a part in what Ward describes as 'evil Phase Gates.'

Regrettably, I fear that I am in good company regarding my past overtrust in the approval process, its phase gates, and their benefits. The objective of this 'article' is to use two hypothetical stories and the role of a Scent Architect to illustrate potential pitfalls and, ideally, guide others in steering clear of replicating my past mistakes and misconceptions.

Why We Have Phase Gates

Phase gates are frequently warranted by a range of factors, such as compliance with bureaucratic requirements, architectural considerations, the assurance of cost-benefit analysis, prevention of duplicate efforts, and alignment with the strategies and perspectives of stakeholders or owners.

Story 1 - The Arrival of the Scent Architect

Following new phenomena or incidents, there's a tendency for additional gates to enter the process. Now, let me delve into the backstory. Enter the realm of Stringency Co., a fictional company, where an unexpected revelation comes to light: a specific code combination emits an undesirable smell wherever it's applied. This discovery triggers an incident, prompting crisis management to gather like flies around the malodorous code, or at the very least, the incident reports. Their objective: to formulate a strategic action plan.

Concurrently, in various corners of the world, curious ethical hackers swiftly pool their intellectual resources, comprehending the issue and formulating strategies to avert analogous unintended smelly consequences. This newly formed community of interested individuals creates ripples, eventually giving rise to a new cohort of 'experts' in anti-smelly code—let's dub them ‘Scent architects.’

Given the scarcity of this resource, companies fight and overbid each other to secure one of these Scent architects. In a remarkable twist, not told by this story, our Stringency Co. succeeds in recruiting one of these rare Scent architects at a salary unimaginable for most of its other employees. This newfound Binary Odour Maestro becomes the saviour, pledging to mitigate the risk of smelly code with one modest demand—namely, the implementation of its own control station in the process. And thus, another ‘evil’ phase gate is born.

Story 2 – Organisational Learning and Knowledge Diffusion

In an alternative scenario, we choose to leverage our internal resources instead of seeking external expertise. The management supports the team in their pursuit of knowledge and additional resources to orient itself around the new smelly phenomenon. The problem is analysed and solved at the actual place and by the team where it appeared (Genba). While this approach may initially demand more time, it not only resolves the issue within the company but, more importantly, demonstrates confidence in our internal capabilities. This represents a departure from constant reliance on external consultancies and experts for all tasks involving analytical or reflective practices.

Fostering a culture where our internal teams are empowered and feel committed to take ownership, steering clear of any perception of infantilisation.

The organisation actively addresses the potential challenge of dealing with smelly code, and the actual place with the affected team becomes the hub of the development process. Instead of relying on distant experts, the problem is solved internally with inspiration drawn from available knowledge to create best practices for avoiding smelly code. These practices, implemented and consistently updated, foster shared experiences across teams. A proactive aim to prevent future 'smelly' code design becomes integral to the process and culture, regardless of phase gates.

Dantar P. Oosterwal, Product Development & Product Planning Leader at Harley-Davidson Motor Company, provides another empirical indication of the ineffectiveness or potential harm of phase gates:

However, when we evaluated five years of product development data, it exposed the fact that, in spite of our best efforts to standardize our methods and to create cadence and flow, there was, in fact, no correlation between exiting phase gates on time and the success of a project. The data even suggested that the inverse may be true [The Lean Machine, p. 148-149].

Beyond the Gates: Emancipating Knowledge Diffusion from the Paradoxical Risk-Averse Ouroboros Logic

Considering phase gates as a solution to mitigate incidents, as depicted in Story 1 above, there is a risk of isolating knowledge and information that could be explicit and diffused throughout the organisation. The potential consequences of such knowledge isolation are far-reaching and include at least three significant impacts.

1. Lack of Ownership

Embracing responsibility for the entire product or service, rather than just one development stage, leads to transformative impacts on both time and quality. Toyota serves as a compelling empirical proof of this approach. Rather than merely aiming to pass the phase gate, the goal becomes delivering a product or service that, due to you or your team, will not malfunction further down the delivery/lifecycle line. It's crucial to note that when signing off at a phase gate, the psychological responsibility shifts to the individuals at the gate, and the psychological power of ownership, as well as commitment to the higher goal, is lost or at least severely weakened.

2. Non-productive waiting

Humans struggle to intuitively understand exponential patterns, much like our difficulty comprehending the delay implications of a phase gate. Teams are pressured to work more efficiently, exert more effort, and even resort to overtime, despite the fact that the effective production time is typically a very small fraction of the waiting time associated with various phase gates. I repeat, a very small fraction of the waiting time. This is where my phase gate colleagues and I may have incurred costs beyond what I dare to contemplate.

3. Costly Overhead

Efforts to avoid misconducts, such as producing smelly code and others, render our carefully designed process, aimed at addressing all potentialities, administratively burdensome, lengthy, and costly to pass. The unintended consequence is that no small intervention can afford to enter this red carpet of a process, which is only rolled out for those with big audacious ideas, goals, and budgets.

The Paradoxical Risk-Averse Ouroboros Logic

The Greek myth of the tail-eating snake, Ouroboros, serves as a metaphor to illustrate how our intentions to minimise risk can have the opposite effect. Since small interventions can't justify the overhead cost induced by a comprehensive process and its phase gates, you need to accumulate several initiatives to be able to bear that overhead cost. As we learned in primary school, when we multiply the likelihoods of success, connecting them in a dependent manner, the risk of failure experiences exponential growth. Consider a simple scenario with three initiatives, each having an independent likelihood of success at 70% (0.7).

If we bundle those three into one initiative to share the overhead cost, the probability for success will be: 0.7 * 0.7 * 0.7 = 34%, representing a doubled risk from the initial 70%.

This is where our thinking turns into a Paradoxical Risk-Averse Ouroboros Logic. In our attempt to design a watertight process to avoid failures, we unintentionally create constraints that drive behaviour to accumulate assumptions. Accumulated assumptions create a foundation akin to a house of cards for the bundled initiative/project, and the risk of failures, or at the very least, delays and costly overruns, is the most likely outcome.

Knowledge diffusion

Instead of fixating on the pursuit of a perfect process to mitigate all potential risks, which tends to lag,

consider the possibilities of viewing incidents and phase gates as potential indicators of knowledge that needs to be created and/or shared (diffused) throughout the organisation.

With this framing, we can separate the connotation of 'evil' from phase gates. As a result, we encourage internal problem-solving, information sharing, and ownership, while minimising non-productive wait time. This, in turn, is likely to lead to shortened process time and reduced costs, enhanced quality, and, not least, work satisfaction. We transform continuous improvement from a theoretical concept into a practical reality, rather than a mere lip service constrained by burdensome and costly processes.

To conclude

It all boils down to Question Zero: Do my actions magnify the threat we pose to a potential adversary?

If you're the 'owner' of, or involved in, a phase gate, take a moment to reflect on its added value. Consider whether you might be withholding information/knowledge that could better benefit the organisation if shared. Open the gates to everyone and co-create knowledge for better quality, cost-effectiveness, and faster development processes. Embrace the psychological power of ownership and eliminate any signs of infantilisation of your fellow adult colleagues within your organisation from the process.

Assisting in the removal of a phase gate could be the most significant contribution you've ever made to your organisation, ultimately enhancing its competitiveness or, in a military context, posing a greater threat, cost, or threshold to an adversary!

?? Further reading

I highly recommend Allen Ward's book Lean Product and Process Development, which inspired me to test and develop the concept of A Set-based Approach for addressing military problems. However, it is a dense read, so starting with Dantar Oosterwal's book The Lean Machine could be a good progression.