“Winners take time to relish their work, knowing that scaling the mountain is what makes the view from the top so exhilarating.” ― Denis Waitley
Before we jump into the topic, could I just take you a step back and remind you what SAFe is all about? SAFe is a way of taking any iterative Agile way of working (normally restricted to a team or few teams) and scaling it up at various levels of the organization, whilst applying a mindset of Lean manufacturing. It also deals with scalability at various levels. Beginning from Essential SAFe right up to Full SAFe, the framework caters to all organizational levels of scaling agility. As part of this, it broadens the core idea of agility mindset beyond just projects/development teams right up to executives/CXOs, who must prepare for enterprise level uncertainties. In a sense, it provides valuable enterprise level scaling insights helpful for the executives to tackle any uncertainties/risks associated with a project.
As you start applying SAFe in your organisation, it is important for you to understand how each level works in conjunction with the other, depending on how mature your SAFe enterprise is. The key link between these levels is the SAFe specific events which help with smooth value delivery facilitation. The events help with alignment across teams, ARTs etc thus helping in managing risk by providing a level based cadence and synchronization.
Essential SAFe - Your First Level of Scaling Using an Agile Release Train (ART). Courtesy © Scaled Agile, Inc. Source: Scaled Agile
While it is important to go through your team level events (like the 4 sprint events if you are doing scrum etc.) it is important to have the scaling events that help with bridging gaps and unblocking dependency between teams. The most important part of these SAFe specific events is for ‘Business Stakeholders’ to get a look (demo) at a proper incremental product and thus the value arising out of it. Makes sense? It did for me and let me tell you why.
I was once associated with 3 feature teams, who were working towards a common product goal. They all had the same business stakeholders but were working on individual features. Team A was working on developing a Login page, Team B was working on a landing dashboard while Team C was hopping along, trying to provide a search functionality for the user. All of them were applying the Scrum framework and were running their own events. Sprint demos were happening individually and were being represented by the Product owner separately along with his business analysts. All seemed fine but there was a nagging problem. The product owner was worried, because he couldn’t bring any business stakeholder to view the demos, as they were being run in silos and there was no visibility on the incremental product. Well technically there was, but they would have to sit through three or four-hour events individually to get bits and pieces of the product demo. In the real world, it's not a possibility simply because your business stakeholders will not have that much time to spend on multiple demos. It is not a good use of their time either. So, what’s the solution? Simple, it’s SAFe to the rescue! Let’s try and understand how the SAFe specific events help with this.
Prescribed PI Cadence for Various Levels of Scaling. Courtesy © Scaled Agile, Inc. Source: Scaled Agile
SAFe is very relevant and designed to thrive in situations where there are significant cross functional dependencies between agile teams and support / functional teams (infrastructure teams, architect community etc).
As you start to scale up one level up, you will be working with anywhere between 5-12 agile teams who will all be collectively working towards a common goal which is the program increment or PI. The anchoring catalyst that brings them all together is your ART (Agile release train). Before getting into the events, lets understand the various roles involved at this level because this is the common denominator across all levels of SAFe and across organizations. This is where you need to get it right without which there is not much use in scaling higher.
Prescribed events on a typical Agile release train (ART). Courtesy © Scaled Agile, Inc. Source: Scaled Agile
According to me, PI planning (hands down) is THE most significant aspect of executing this framework. This is where all the magic happens. It is sometimes referred to as the heart of the framework as it offers a clear vision of what the program increment needs to be, what the cross-team dependencies are and how they bring together the cultural sustainability much needed within the release trains. It is so important, that if carried out incorrectly it could lead to several ambiguities, development challenges and mostly a disastrous product increment. However, when it works well, the iterative cycle serves to flesh out the crucial elements of the plan and the processes ensure buy in from the stakeholders.
Duration: A normal PI planning is a 2-day activity, which is a face to face cultural get together of the various ART teams. However, a new 3-day distributed PI planning has been introduced to help with geographically distributed teams (across various time zones), very apt for the current pandemic situation.
“There is no magic in SAFe® except maybe for PI Planning”. – The authors of the SAFe framework.
In big organizations with multiple distributed teams across multiple vendors, work streams etc. it is almost impossible to run these teams independently, whilst still having to deliver an incremental program. SAFe via the PI planning exercise mentioned above, helps with sorting out these issues by recognising cross team dependencies upfront, constantly negotiating & visualising them. This doesn’t just stop with the PI planning but the framework also proposes a cadenced way of continuing this via the scrum of scrums. The Program Board is an ideal way to showcase the cross-team dependencies.
A sample SAFe Program board. Courtesy © Scaled Agile, Inc. Source: Scaled Agile
An inspect and adapt event is scheduled after every PI. This event is dedicated to aligning to the principles of Kaizen, which simply means to change for the better. The events contain self induced thought processes to revalidate your assumptions that everything is working OK. The I&A event consists of three sub-parts as below:
Agile release trains tend to apply a cadenced synchronization process to help manage the ability to focus on continuous value delivery. An ART sync will typically comprise of the below sub-events.
This event is for representatives from all the teams on a release train to come together in a regular cadenced manner, especially on large ARTs. This is normally facilitated by the release train engineer (RTE) and will involve scrum masters of the individual teams and a few selected team members (authorised by the team). The sole purpose of the SoS calls are to understand progress towards the common goal, validate cross team dependencies and unblock impediments that may arise out of them.
Duration: The length and frequency of the meeting will depend on a few factors like the size of the ART, the release frequency, type of features being worked on, ability to decouple releases etc. For e.g an ART which releases features into production every 4 weeks might want to have an SoS call every 2 weeks for about an hour. Again, if this doesn’t work for you, just inspect and adapt to what works well for your organizational needs.
Just make sure that the SoS is utilised for its sole purpose and not just status updates as depicted in the below comic representation.
Scrum of Scrums
This event is represented by the Product Owner, business analysts and the product management group. This is used mainly to level up the product backlog refinement and for clarifying PI (Program Increment) scope, reviewing roadmaps and grooming for the upcoming PIs.
Duration: Very similar in concept to the SoS, so just follow what works for the group.
As part of a common understanding towards delivering incremental software, shortly after each iteration in the PI, there is a system demo scheduled. Work completed across all teams from the release train are compiled in a stable environment before it is reviewed by the business stakeholders and other important sponsors who may have a keen interest in the product. This is on top of the individual team level demos that happen after each iteration.
Duration: Anywhere between 2-3 hours that will allow time for a demonstration of the program increment in a collative manner, on top of what has been delivered from the previous PIs as well.
In case your ART is pretty small, then you may want to have just have some of the events combined into a more generic ART sync, where all roles come together to collaborate towards the program increment. This can sometimes occur if the ART is focusing on a particular value stream, confined to limited business functionality, rather than elaborate features.
As you scale higher, the processes and events become much less prescriptive. There is a good reason for this because the focus at this level is not just on having repetitive demos that have already happened before but on building thought leadership around business outcomes and enhancing business agility. Which is why we will not be diving deep into that in this blog. But let us look at the events that occur at the macro level.
Well, the more I talk about this, the more excited I am. A PI planning event when carried out to its truest purpose, gets half the job done. Here is where most of the brainstorming occurs and business value gets determined and, in some cases, gets assigned in a quantifiable manner to user stories and helps with the prioritisation.
The events are a constant reminder that all teams are working towards delivering incremental value either on a particular value stream, or feature or program. An RTE and Product Management will help reiterating the need to focus on the larger goal whilst helping sorting out inter team dependencies.
As is the framework itself, SAFe events/ceremonies are less prescriptive. An SPC would recommend, apply the principles but inspect and adapt as to what works for your organization. As per the example I provided earlier w.r.t to the duration of the SAFe events, start with something reasonable and then validate its effectiveness. Then leave Kaizen to do the rest.
Opportunity for Business stakeholders and sponsors to have a look at the overall program increment every iteration, thus helping them evaluate the progress and provide timely feedback on market trends.
Things can go wrong if there is not enough representation in the product management group, say for e.g at the PO Sync event. This can lead to a blurred product vision with each team working out of sync. This may ultimately get detected too late, probably at the time of the system demo, and lead to a whole lot of unwanted rework.
The Scrum Of Scrum event should be used as an event to unblock cross team impediments or dependencies and not to just update what each team has been doing or is doing in its current sprint.
Given the scale at which these events will be conducted, it is critical that the associated events are facilitated in a timeboxed manner or else the participants could end up sitting and talking for hours. Roles like RTE, SPC Coaches etc will be critical in addressing this issue.
Lack of effective collaboration tools could lead to some disastrous situations whilst facilitating the SAFe events. Given that most teams are running virtual ceremonies/events at the moment, its crucial to establish a working distributed model. This will then ensure that the platform is set up for the most effective collaboration and cross-functional work to take place.
While you try to scale, as per the implementation roadmap, its essential that you solidify the process around which your ARTs will be functioning. It’s like setting the railway tracks with the correct track gauge matching the configurations of the wheelsets of the trains that will run on them. If not, they will just derail. As your ARTs pass through your set process, they will only benefit by sustaining focus and pace while moving towards a successful incremental product delivery.
Thanks for your patience and wish you all the very best in your Agile journey. In case you want me to write about any specific topic, please feel free to comment below and I’ll be more than happy to add them to my ‘Blog Backlog’. If you liked the article, please do share it among your agile community to help spread the word.
Hope to see you soon, with more such interesting topics.
Your email address will not be published. Required fields are marked *
85 percent of respondents say Scrum continues to... Read More