The announcement of the acquisition of Disciplined Agile (DA) by PMI is almost a month old so I thought I would share my thoughts on it.
There is no doubt that PMI has been flirting with agile progressively over the past decade. The launch of the PMI-ACP credential, the addition of adaptive life cycle considerations to the PMBOK® Guide, Sixth Edition and the release of the Agile Practice Guide were all signs of this growing interest.
However, PMI suffers from being perceived as a champion of bureaucratic, traditional approaches to business value delivery which has generated a fair bit of cynicism from the agile community. The partnership with the Agile Alliance which led to the development and publication of the Agile Practice Guide were viewed by some as a unhealthy dalliance or a marriage of convenience.
Correcting perceptions and developing sufficient intellectual property (IP) would have taken PMI many years to do organically so acquiring legitimate thought leadership, credibility and IP was the better strategic move. A key decision was to choose either a method-centric (e.g. SAFe, LeSS) or method-agnostic (e.g. Disciplined Agile, Modern Agile) organization. Given the need to address a global market with varied needs, agnosticism won out.
There are a number of potential advantages to PMI, DA and practitioners.
PMI now has the ability to incorporate the significant intellectual property of DA within their knowledge base and by doing so, enhance the value proposition of their standards and practice guides. While tailoring considerations were minimally explored in the PMBOK® Guide, Sixth Edition, they can now go much deeper by leveraging the DA process decision-making framework. PMI can also expand the breadth of their credentials and by doing so, will add credibility to the existing DA ones. Given the strategic relationships which PMI's senior leadership has forged with major global corporations, this acquisition will open doors for DA which might not have been possible otherwise which in turn might accelerate the evolution of DA. It is also an opportunity for the DA framework to go beyond technology delivery.
But there are some risks, including the dilution of thought leadership, the obsolescence of existing credentials and the risk of PMI actively competing with partners (e.g. Registered Education Providers). PMI might also make the mistake of not fully integrating DA into their offerings which would limit the benefits realized from this acquisition.
If there is a single piece of advice I'd like to pass along to PMI & DA, it is from Audrey Hepburn: “If I get married, I want to be very married.”
Completion of a project usually focuses on financial and administrative activities such as transitioning verified and accepted deliverables to customers, getting final sign off on the project, closing open contracts, recognizing team accomplishments, survey stakeholder satisfaction, archiving key project outputs and so on.
But with an adaptive life cycle, the approach taken when performing certain activities might vary. Here are two examples of this:
So is there any difference to how projects which were managed using an adaptive or agile life cycle would end compared to those following a traditional or predictive one? At a high level, no, but our approach to specific procedures should always be tailored to fit the context of the life cycle used to deliver it.
It is common to see teams new to those agile frameworks which use time boxes struggling to define product backlog items small enough to get completed within a sprint and still represent value. If they have been used to traditional, predictive delivery approaches, they would have been encouraged to decompose scope down to a manageable level of control, but these work packages would still usually be larger than what is expected of a sprint backlog item.
Some teams respond to this challenge by using long sprints.
While this gives them the opportunity to complete more work within a sprint, it has downsides. The team might be inclined to batch their work by completing one stage of their delivery process for a number of sprint backlog items rather than completing these items individually over the life of the sprint. Long sprints might also reduce the team's inclination to sufficiently refine and split stories.
Other teams try to avoid the challenge entirely by skipping sprints in favor of a continuous flow approach. A mature team can effectively use this delivery method, but new teams without coaching support might end up with an ever-growing work-in-progress backlog as work commences on a work item but as soon as it gets blocked another work item gets pulled.
Shorter sprints make it harder for a team to ignore the reasons for why their work items aren't getting done, including:
While there is no silver bullet to deal with these causes, if a team remains true to the pillars of inspection and adaptation, their ability to deliver consistently should improve over time.
Face-to-face communication around a shared modeling surface such as a whiteboard is considered to be a highly effective method of creating shared understanding. When team members are distributed, collaboration tools can help with reducing misunderstandings and miscommunications between team members but can't fully eliminate those.
But what about the impacts of distribution on other roles such as agile leads or senior stakeholders?
To an outsider, an agile lead's visible contribution in the early stages of team formation might be the facilitation of agile ceremonies but this is really just the tip of the iceberg. If the team is maturing, any team member should be able to run ceremonies. The greater value brought by an agile lead in such cases occurs outside of ceremonies, through actions such as suggesting opportunities for increased collaboration between team members, coaching individual team members and working with key stakeholders outside of the team.
With co-located teams, agile leads often identify coaching opportunities osmotically by just being present with the team. A casual comment might spur an agile lead to have a conversation which might avoid hours of unnecessary work. But with distributed teams, unless the agile lead is plugged into all the conversations taking place between team members and with external stakeholders, there is an increased likelihood of missing a coaching moment or being unable to help the team to address a blocker in a timely manner.
Persistent group chat tools such as Slack or Microsoft Teams can help the agile lead to be somewhat aware of what's going on, but it is rare that all members of a team would conduct all of their interactions through these platforms. Online work management tools such as JIRA or Rally can capture significant updates for the team's work items, but it is rare that an agile lead will review all changes to those work items even if team members have been diligent about regularly updating them.
In such cases, eliminating or reducing the level of multitasking will be critical as the impacts of context switching are compounded with distributed team members.
With senior stakeholders, even if good information radiators have been constructed and published, lacking in-person access to the team might increase the number of ad hoc requests for updates as there is less opportunity for them to observe how the work is being done. It might also put a greater emphasis on regular showcases or frequent product reviews as the primary method of soliciting product feedback rather than the preferred ability to drop by and see the latest build as the need arises. This might increase the time it takes these stakeholders to begin to trust the team's capabilities. Similarly, because team members won't have regular in-person contact with these senior stakeholders, they may also take more time to start to trust them. Bringing the team together with key senior stakeholders on a periodic basis is one way to address these concerns, but this can be an expensive proposition.
With ever increasing needs to tap into dispersed skill sets and to support flexible working arrangements for team members, distributed teaming is here to stay. Acknowledging that more effort will need to be spent overcoming distance challenges is an important step to managing expectations with this way of working.
One of the learners in a class I taught asked me how self-organizing teams would handle the situation where a single team member is not performing a fair share of the work.
In a traditional, push-based work assignment model, this issue can also occur. Usually whoever has pushed the work onto the low performer will follow up with them in a timely manner and take direct steps to ensure that performance improves, the work gets re-balanced or the team member is replaced.
But with a pull-based approach where individual team members sign up for work items as capacity frees up and where the focus is on how much is getting completed by the team as a whole, the concern is that someone could take advantage of this by letting their peers take on the more challenging work items leaving them with a relatively lighter work load. From the outside, it would appear that the work is getting done but the contribution imbalance will be less evident.
On a mature self-organized team this is not likely to be an ongoing concern. Team members recognize the importance of demonstrating courage and showing respect for the team and will exert the necessary social pressure on the low performer so that they will either feel embarrassed and start to improve or will be transferred out of the team. If the team uses agile ceremonies such as sprint planning, daily standups or retrospectives, these provide an opportunity to provide feedback with radical candor to the low performer.
But on those teams which are relatively new to working in this manner, the team members might not possess the confidence to openly discuss or challenge such dysfunctions. Passive aggressive behavior might occur such as remaining silent during a retrospective and complaining around the water cooler afterwards.
In such cases, the agile lead or Scrum Master might need to get involved to eliminate the impediment. This intervention could start in a subtle manner such as asking the individual at a daily standup if they feel comfortable with their workload for the day compared with their team members, or "seeding" the conversation during a retrospective when the topic turns to what could be improved. They might encourage other team members to ask the individual to give them a hand with their work items or to conduct a non-solo work experiment (e.g. pair programming).
If this soft approach doesn't work, the Scrum Master may have to more drastic steps such as confronting the individual one-on-one, speaking with their manager or some other type of escalation.
Deferring decision making to the Last Responsible Moment is a lean principle. While a Scrum Master shouldn't intervene if the rest of the team can address an issue for themselves, they should have the judgment to know when they will need to take a more direct approach.