Murray Robinson的动态

"Mutant Agile" The Rise and Fall of Agile in Government - Part One ? Back in 2007, the Department of Education and Training saw a ground breaking shift. Myself and a team of top-tier ICT professionals rolled out one of the first major Agile implementations in government. With the skeptical but ultimately supportive backing of executives like Terry and Adam, we embarked on a journey that transformed our delivery process. ? We turned a department struggling to deliver software into a powerhouse of delivery, shipping more software and delivering greater business value. This transformation wasn't without its hurdles. From Business Analyst strikes to change management roadblocks, the path was anything but smooth. ? Our success hinged on a steadfast commitment to Agile principles. We didn't sacrifice good governance; we embraced it. We maintained project steering committees, conducted formal status reports, and yes, we adhered to schedules and budgets. This disciplined approach allowed us to deliver faster and cheaper than traditional methodologies, while maintaining a high standard of quality. Our department operated lean, with tight budgets that left no room for wasteful activities. ? The biggest lesson? Agile isn't a buffet where you pick what you like. It demands high discipline from everyone involved. The core principles are non-negotiable. The policies and procedures we developed were tailored to our environment but always stayed true to Agile's integrity. ? Fast forward to today, and what I see in government is a disturbing trend of "Mutant Agile." In many cases, it's worse than the old "Waterfall" approach. ? The root of the problem? A flawed funding process. The ridiculous first and second pass business case procedures are draining millions and encouraging bloated programs and budgets. ? The empirical evidence is clear: larger projects have a higher failure rate. This ties back to Agile. Smaller projects and budgets mean less waste, higher productivity, and faster iterations to achieve business outcomes. Massive budgets, by their nature, are anti-Agile. ? We need a major overhaul in government procurement to unlock significant value. This reform will help return Agile to its core principles—stretch targets and disciplined execution at the most fundamental level.

Another day, another round of 'The End of Scrum' and 'The Death of Agile' headlines flooding our social media feeds. It's like a broken record, isn't it? And those recent mass layoffs of Agile Practitioners from major companies only add fuel to the fire. But amidst all this chaos, a post by Maarten Dalmijn caught my eye today, ?and it made me think. [To think I hoped for a day off from social media ??] While I agree that Scrum as outlined in the Scrum Guide may slowly be disappearing, I'm not entirely convinced it's heading for extinction. Maarten Dalmijn ??, let's have a chat about this! My take on it is that many companies fell into the trap of creating a bubble of Scrum Masters and Agile Coaches. Similar to the dot-com bubble of the 2000s. They believed that flooding their teams with these Agile People (as I will refer to Scrum Masters and Agile Coaches) would automatically make them Agile. As if doing Scrum by-the-book will elevate the efficiency and productivity of the companies. But as we dig deeper, we find cracks in this approach. Firstly, hiring a wave of Agile practitioners at once doesn't guarantee top-notch performance from each one. Secondly, achieving true agility requires more than just following Scrum rituals—it demands a holistic transformation encompassing DevOps culture and development practices. Perfecting Sprint Planning won't do much if your test code coverage remains close to zero. Agile Manifesto states clearly “We are uncovering better ways of developing software by doing it and helping others do it.” It’s about developing software in a better way, not just putting processes around it. I actually agree with Capital One assessment: “The agile role in our tech organization was critical to our earlier transformation phases, but as our organization matured, the natural next step is to integrate agile delivery processes directly into our core engineering practices,” — Capital One. Yet, I still think they need some Agile Practitioners, operational people to help with the team collaboration and change management for the development good practices adoption. That’s not as easy as it seems. As a society of people working in technology we arrived at the plateau or some even the transcendence phase (Shu-Ha-Ri) of the Agile frameworks. It’s time to break them and come up with our own way. A way that can be based on the already crafted and successfully used agile practices. While Scrum as defined in the Scrum Guide might slowly disappear, the concept Sprints might not. The backbone of the framework provides teams with valuable feedback loops, a sense of predictability and also some stability. Looking ahead, the future of Agile lies in crafting a customizable fusion of practices from Scrum, Kanban, XP, DevOp and more. All while prioritizing collaboration, code quality, and—dare I say it—getting our hands dirty with technical aspects of software development.

On the contrary, practitioners of actual Scrum guide-based Scrum will grow, and the Agile mentality will prevail in the tech industry for a long time. Simply because it’s efficient and beneficial for practitioners, and profitable for stakeholders. What will indeed fade and slowly disappear is the blown-out-of-proportion Scrum industry, certificate and course market, and hordes of semi-charlatan ‘Agile coaches’, ‘Scrum masters’ and ‘Product owners’. Scrum will be practiced on a team level where it belongs, run by Scrum masters whom developers will elect themselves, or hired managers with hands-on experience of the framework; and product owners from the CEO’s office to ensure the pipeline. Without any fanfare and long, useless ‘Agile transformations’, but simply as a series of 5 events (which would take no more than 20 hours a month) and managing backlogs. For the purpose of doing twice the work in half the time, and not for the sake of great philosophical and existential goals. As simple as that

From Scrum Master to SAFe Advanced Practitioner: A Roadmap for Agile Growth Transitioning from a Scrum Master to an advanced role in the Scaled Agile Framework (SAFe) is a rewarding journey that demands a shift in mindset, leadership skills, and the ability to navigate complexities at scale. Here's a roadmap I’ve followed (and recommend!) to make this leap successfully: ?? 1. Strengthen Your Foundation Start by mastering the core responsibilities of a Scrum Master—facilitating high-performing teams, delivering value, and embodying servant leadership. Success at this stage lays the groundwork for scaling Agile. ?? Goal: Build a proven track record of delivering successful projects while fostering a culture of collaboration and continuous improvement. ?? 2. Learn the Fundamentals of SAFe Pursue the Leading SAFe certification to understand how Agile scales across the enterprise. Dive into concepts like Program Increment (PI) Planning and the Lean-Agile mindset. ?? Goal: Contribute to Agile Release Trains (ARTs) and demonstrate alignment between team outputs and business objectives. ?? 3. Expand Your Leadership & Coaching Skills Evolve your skills to coach not just teams but also stakeholders and leadership. Focus on driving cultural shifts, enabling collaboration, and fostering a Lean-Agile mindset across all levels of the organization. ?? Goal: Become a trusted advisor for teams and leaders navigating the complexities of Agile transformation. ?? 4. Pursue Advanced Certification The SAFe Advanced Scrum Master (SASM) certification takes your expertise to the next level. Learn to scale Agile practices, resolve cross-team challenges, and drive program-level success. ?? Goal: Bridge the gap between team-level execution and enterprise-level strategy. ?? 5. Master SAFe Program & Portfolio Management Collaborate with Release Train Engineers (RTEs), Product Managers, and Solution Architects. Use tools like Portfolio Kanban and SAFe metrics to ensure value-driven delivery. ?? Goal: Align team deliverables with organizational strategy while managing dependencies across Agile Release Trains. Challenges and Lessons Learned ?? Overcoming resistance to change. ?? Aligning team work with broader business objectives. ?? Gaining leadership buy-in for Agile transformations. ?? Pro Tip: Build strong relationships with stakeholders and use SAFe metrics to demonstrate tangible business value. The End Goal: Driving Enterprise Agility This journey positions you to lead enterprise-wide transformations, influence strategic decisions, and deliver measurable outcomes at scale. Whether you aim to become a Release Train Engineer (RTE) or a SAFe Program Consultant (SPC), this roadmap will guide you. ?? Question for You: Where are you in your Agile journey? Let’s discuss the challenges, wins, and lessons learned in scaling Agile. #AgileTransformation #SAFe #Leadership #CareerGrowth #LeanAgile #ScrumMaster

Great InfoQ writeup by Bryan Finster on the failure of Agile delivery frameworks at Walmart and the success of DevOps. https://lnkd.in/gi68Huwi I'll copy the top two bullets here but you should read it: "- We saw negative outcomes from agile scaling frameworks. Focusing on "Why can’t we deliver today’s work today?" forced us to find and fix the technical and process problems that prevented agility. "- We couldn’t deliver better by only changing how development was done. We had to restructure the organization and the application. I know many will claim Agile and DevOps are compatible, and of course they share a common history. But lately it's become clear that the formalized Agile delivery frameworks ("industrial Agile" or whatever you care to call it, and this includes the Scrum ecosystem as well as SAFe) are as misguided in their own way as ITIL was back in the day, and are encountering the same fate: advocates, trainers and coaches let go en masse, and shattered credibility among IT leaders. Many stories circulating, many pained "where did we go wrong" blogs and articles... What's very interesting to me in this Walmart writeup is that it's directly supportive of Marty Cagan's recent statements in Transformed: "...many of the best product teams in the world have mastered the ability to consistently deploy very small releases (referred to as continuous integration and continuous deployment, or CI/CD), yet they don't follow any formal Agile process or methods. Contrast this with organizations that invest so much time and money in adopting Agile coaches, rituals, roles, methods, and processes yet, at the end of that considerable expense and time, are still releasing quarterly and inflicting the associated pain on their customers." (p. 34) I have many thoughts on the implications, but given the limits of a LinkedIn blog will have to share those elsewhere. In the meantime, comments greatly appreciated.

Agile is dead, said someone in the Internet, and much debate ensued. Not dead, but can't deny the fact that countless organizations fail trying to implement it. It's one of the two things really. Agile without direction and the ageless classic: it was waterfall all this time. I even wrote my thesis about it, the way how Agile gets dumbed down to implementation phase of waterfall project. Just a way to manage task queue in a bit shorter cycles. Everything is top down, Scrum just models shorter painfully unchangeable cycles and purest of the pure, Kanban, really only seems to work when it's exercised in small box within larger, pre-planned framework around it. Doing Agile so that it actually works is an uphill battle because it's our brain chemistry that we're fighting against. Neurotypical brain is literally wired to think top-down, planning the big things before it can start doing details. That's waterfall thinking built into our brain. And what's Agile if not embracing bottom-up thinking, the very thing that's less known territory for neurotypical brain. Intellectually we get the idea. Promise of vastly improved efficiency intrigues us. But the disappointment comes when implementation fails. What to do? Give up and fall back to old pattern that doesn't really work but at least makes sense? Call neurodivergents for help? Okay, happy to be of assistance here but I actually believe in the power or realization. If you acknowledge those mechanisms that create the gravity pulling your mind towards top-down waterfall-like thinking, it opens the potential to consider other options. There's a big difference between "I think we need to do it like this" and "I feel we need to do this because my brain is wired to prefer this approach". Understanding it is power. That's what people do in therapy too. They identify issues so that they can then work on them. No, I'm not casually referring to top-down thinking as an issue that you need to heal from. I'm telling that understanding how your own mind works helps to remove obstacles and opens up potential to grow your own mental toolset. So to summarise: Agile, as general set of principles, leans more towards bottom-up thinking than traditional old school processes. It comes with a promise of better flexibility and efficiency, but the caveat is that model itself relies on thought patterns that neurotypical people are often not very comfortable with. Acknowledging that issue might very well be one of the keys that help implementing agile methodology successfully. Agile processes here are of course just an example. The core of this, let me be honest, written from bottom-up train of thought that I started in previous 2 posts, can be applied to wide array of planning and problem solving contexts. Just needs a bit of introspection and willingness to widen ones own horizons.

I posted a bit ago about the shortcomings of Agile/Scrum without providing many details. Upon reflection, I realize that many of my concerns are not about the methodology itself but what I'll call "Agile Anti-Patterns." I commonly see these pop up when teams try to adopt Agile/Scrum with an inexperienced Scrum master or Agile coach. For example, one of the first things I see new teams do is try to build a mapping table between story points and some fixed-time duration (e.g., one story point == 1 developer day, etc.) instead of using them as a relative measure of complexity. The first truth teams need to accept is that any time/duration-based estimate given with a one or two-sentence definition of done is, at best, a fantasy. At that level of detail, a person can only make a (very poor) assumption about what the product manager is asking for. Sure, even this will self-correct to provide a useful velocity over time. But how long do teams have to get here? The problem new teams are trying to solve by going this route is overcoming the fact that they have little to no history of completed stories to "point off of." A better approach would be to think of the scale to represent the universe of all the projects they have ever done - with 1 being the simplest, most straightforward thing they can think of and n (the top-end of their scale) being the most painfully large, complex, and involved project they have ever done. Ever. Then, look at the story they are "pointing" and ask, "Where in the universe of complexity does this lie, based on what I know?". Sure, the metrics won't map person-to-person, but the conversation the scrum master should lead will yield precious insights. As teams gain experience walking this path, the universe of comparative projects should shrink only to the work the team has accomplished together, and then, recently. This narrowing of focus and the history of points closed each iteration are the magic that makes story points worth the time invested in doing them in the first place. Teams need patience, both from stakeholders and the team itself, to get here for any sort of payoff. Some techniques I've used to help teams overcome this "new agile team" hurdle are things like "Tee-Shirt Sizing" to avoid straight-up numerical conversions. You must avoid mapping "Small = 1 day" at all costs; otherwise, you'll be right back where you started. The other mistake people make when adopting Agile is taking a team of engineers and thinking they can do it all themselves without buy-in or support from the surrounding business, specifically the product owners (PO). Having POs skipping stand-up, grooming, and refusing to write user stories with enough clarity and an articulate definition of done (or at all) is a recipe for disaster. Teams in this situation would be further ahead just giving up on Agile and returning to Waterfall methodologies (or something else, hint, hint). More to come soon. // John

Agile development has been the stated approach of every software product I have worked on for the past 15 years, but almost all have been shoehorned into some form of traditional waterfall management, creating issues. ?????? ???????? ???????? ????????????? Senior leadership and the investors are looking at the long term, and even the exit strategy, and need to see both a plan for progress and milestones to measure against. And because these roles are essentially financial, they are invariably tied to quarters and year end. In addition, commercial reality appliers immovable external deadlines to release and launch dates, whether this is trade shows, tender deadlines, Christmas breaks or specific industry annual cycles. This leads to requests from management to product and development for 6 and 12 month plans, which get horrendously mixed up with the Agile planning, and leads to a lot of (unnecessary) detailed planning work, distracting from the actual desired build. ???????? ?????? ???? ????????? First, it is the job of product management and the head of development to provide an isolation layer between these competing drivers, working with the developers on short term agile detail, and with senior management on the long term big picture plan – translating and coordinating as they go, but protecting each side from the other. A key component of this is the outline solution or product vision, a very high level version of the current view of the final end product. This may look like a series of very roughly sketched and named web or app pages to the developers, and more like a list of modules or top level feature capabilities to management – essentially it is an attempt at the chapter headings of the solution (user admin; customers; finance; integration API etc) so everyone can see what is the target. The development plan then probably breaks each of these areas down into something like ‘Basic’ (MVP),’ Phase 2’ and ‘Advanced’ features. Note that at this level, very little to no definition of what each of these actually contain is required. A vital use of this outline solution is to provide the management timeline or roadmap – which is about coming up with an ordering these chapters for development based on the commercial imperatives for the business (on the understanding this order can be changed at any time – Agile at management level!). It is usually of great help to align and present this timeline with commercial planning (ie quarters) and not as the timeboxes and increments used by developers Agile planning. As a final step, commitments to the board and investors - to allow them to check progress against plan - should ideally be of the form ‘deliver 2 Basic modules’ or ‘1 Basic and one Phase 2’, leaving the exact selection to be made, planned and managed in true Agile fashion as resource becomes available, but paired with the current roadmap. This provides the necessary reassurance and confidence that there is a plan, and it is working.

My first "agile team experiment" started in 2007 when I proposed "Scrum" to my partners in our quite-fast-growing startup (a software development company focused on web developmemt, content management, ecommerce). I still remember my white board drawings in my quite tiny "corner office" at Villa Media in Wuppertal: We had just won a bid for a "larger" ?? software project and needed a better organization for a full-time team of maybe 5 or 6 people. And we needed something that worked: I had already developed this inner resistance against "traditional project management" (PMI, PRINCE2, etc.) which was so focused on planning, predicting the future and a very static, nearly mechanical way of managing (that was quite some time before I discovered Complexity Management as a sciece, though-leaders like Dave, or the Cynefin Framework) - it just "felt" wrong in relation to the things we were trying to do. So, we went for this "new thing" called "agile" and became quite good at it. I introduced it at many companies I joined afterwards, and every team actively did what you are supposed to do in real agile: Take the principles, learn some practices, adapt it, make it your own, develop it further. Fastforward 18 years, and the discussion around the end of agile is already going for quite some years in the leading circles - and I totally understand why: The top large consulting companies are now doing "agile" (at least they pretend to) and selling "transformations" to their clients, pretending they never did something else - and now there is a multi-billion dollar industry around consulting, coaching, training, certification, and content. It is very debatable whether "agile" is still evolving (or even helpful) as most people & companies lost that way years ago, turning to bullshit-"frameworks" (like SAFe) and perverting the original ideas of agile into all forms and sizes. Me, personally, having been a big fan for many years, I can clearly see that "agile" is at the end of its lifecycle - and real Complexity Management, with lots of scientific underpinnings, is what people should be applying these days - something I drive my teams and my whole company towards for already 5 years now. But even for people who thought that there is hope, who are still klinging to "agile" and who believe that the "large-scale transformation" is still not finished in their own company, the following developent should be a sign: It's over, people - it's really time for something better, something real, something more effective now.

I was seeking ideas on how to effectively scale agile and was exploring creative solutions that might benefit my company. Familiar with "Nexus" (Scrum.org) and having read about Large Scale Scrum (LeSS), I was eager for more pragmatic examples and clear directions, that is pre-conceived solutions. Although I missed the meetup hosted by Stefan Wolpers where he introduced Alexey Krivitsky ?? and Roland Flemm, their course "Org Topologies? ??? ??" caught my attention. To my surprise, the course exceeded my expectations and challenged my preconceptions about scaling agile. Initially, I sought new directions for scaling (to critically assess them, of course). However, Alexey's and Roland's course thankfully reframed my questions, highlighting an unagile thinking approach I had been unconsciously taking. They shifted the focus from prescriptive frameworks to the underlying value and outcomes of scaling at all. They posed critical and fundamental questions: Why scaling? What is the current state of your teams and their composition in your organization? What behaviors are rewarded, and what is your ultimate scaling goal? What compositional or topological level do you aim to achieve to realize the best possible outcomes for your product? These questions are so intuitive, yet it is humbling to be reminded that the focus should be not on outputs but on real value and outcomes -(how easy is it to fall into the output trap...). I consumed the entire self-paced course in one sitting and have now revisited it a second time. Key learnings emphasized the importance of reducing dependencies, enhancing cross-functionality, fostering experimentation, and quickening feedback loops—shortening the "time to learn." The course advocated for creating teams responsible for delivering value from discovery to delivery, highlighting a shift from scaling up to scaling down to minimize overhead and maximize value creation, feedback; inspection and adaptation. This thought-provoking course also included over 90 minutes of case studies and interviews, providing concrete steps that could be tried. However, a summary of these interviews could benefit those who prefer written reinforcement (as a possible feedback for the creators). What you don't get here is a one-size-fits-all solution; rather, the course excels at offering valuable insights to start mapping your path and asking critical questions—a kind of continuous topology improvement. The recommended approach is empirical: experiment, observe, learn, and adapt with the guidance of org topologies. Lastly, integrating Michael Lloyd's "Dysfunction Mapping" with Org Topologies? ??? ?? could provide a comprehensive approach to identifying and addressing organizational compositional pain points, tracking solutions, and measuring their impacts. In short, I am not only grateful to the creators of Org Topologies but also excited to continue this learning journey, becoming more proficient as a professional "de-scaler".