Agile Portfolio Management:
There is a new way of doing things in delivering a complex product portfolio. It focuses on delivering value both incrementally and iteratively. It utilizes empirical process control and hypothesis-driven planning. It utilizes test-driven development in both convergent and emergent delivery, even when budget and scope are fixed. It utilizes a Lean kaizen approach to maximize velocity.
This philosophy is by nature, object-oriented and modular. No one framework is right for every product, so it is highly customizable. It may sound new to you, but it has been around for quite awhile. But wait – I’m not talking about Agile, Scrum, or Lean software principles – I’m talking about Westside Barbell’s approach to powerlifting.
Waterfall Weightlifting:
Powerlifting is a sport in which the lifter competes for the highest single-repetition maximum in the Squat, Deadlift, and Bench Press for their weight class. The traditional approach to training powerlifters relied on linear periodization – a method still very valuable for beginning athletes because each phase builds on the last while progressing toward competition-specific strength.
At a basic level, here is a 12-week competition plan:
3 Week Hypertrophy Phase (muscle size, stamina): Sets of 12 to 15
3 Week Strength Phase (movement form, ability to move weight): Sets of 5 to 7
3 Week Power Phase (Explosive speed, maximum weight at progressively higher volume): Sets of 1 to 3
3 Week Peak & Rest (Highest weight, lowest volume): Sets of 1 to 3, tapering off to a few rest days
Competition: Three chances to get three lifts correct, competing against others who are doing the same
As agilists, this correlates perfectly with the “waterfall” approach we try to leave behind:
Hypertrophy phase: Business planning, creative design, and thorough documentation
Strength phase: Database layer, middle-tier
Power phase: Client-side logic, front end development
Peaking phase: Testing, beta release, focus group and stakeholder reviews
Rest days: Code freeze and marketing
Competition: Release to the market, in which you may not recover from failure
Then the lifter starts over. If there was a big loss (e.g. an injury) pre-competition, the weight lifter might not compete at all – just like software project that gets cancelled after key engineers leave or technical debt gets too high to meet the release date. More problematically, if there is a big loss or injury at the competition, the lifter may never compete again- just like the software team with a botched release that gets “reassigned” or laid off.
Repeating the Cycle:
The weightlifter who perseveres, win or lose, still has big “waterfall” problems. The lifter rests a little and repeats the linear progression cycle, an exercise in bodily context-switching. When the next hypertrophy phase starts post competition, most of what was developed in the previous cycle is gone! The same is true of each phase. When the lifter resumes focus on 3-rep max, some hypertrophy and stamina is lost. As the lifter peaks for competition, the 1-rep max may increase but the 5-7 rep range decreases. Studies show that after a few weeks in the subsequent hypertrophy phase, up to 15% of single-repetition strength is lost. The disconnect between foundational planning (by increasing stamina and size) sacrifices a considerable amount of value captured (ability to perform the same single-rep max).
What does this specificity-switching cost the lifter? As a beginner, not very much – any work will improve size, conditioning, and maximal strength, and fantastic progress can occur. The discipline of repeating the movement pattern likewise increases maximal strength even with little planning. However, once the lifter goes from a beginning athlete – a time when nearly anything will improve the lifts – to an intermediate athlete – subsequent peaking phases will see little or no increase.
The process requires disruption if total stagnation is to be avoided.
If this sounds like delivering software in waterfall, it is! As you read this quote from a strength coach describing the “waterfall” lifting approach, think about the Waterfall PMO:
Having now gotten away from this type of training and looking back as an outsider, I can see where the program is lacking and why I had so many problems. I used to feel it was the only way to train (mostly because it was all I ever knew). It was also the only type of program for which I could find a lot of research. Some of the limitations to this linear style of periodization include:
- It’s a percentage-based program
- It starts with a high volume
- It only has one peak
- Your abilities aren’t maintained
- The program has no direction to the future
– Dave Tate via T-Nation.com
Here are the parallel problems we see with waterfall:
- “It’s a percentage-based program” – accounting-based statistical process controls are applied to an emergent system
- “It starts with a high volume” – a significant portion of the budget is spent planning, designing, and fighting about features that no user wants (and if the project is cancelled, 100% of this sunk cost never drives user- or owner- value capture)
- “It only has one peak” – A major release attempts to market itself to all segments simultaneously and a flop may kill the product line completely
- “Your abilities aren’t maintained” – once the waterfall project plan is set in motion, market evaluation, user feedback, and stakeholder review is non-existent
- “The program has no direction to the future” – a waterfall project plan is delivered based on the knowledge available at the beginning of the project when the least is known and has no intrinsic method of looking to the future relationship between the user market that might exist and the software that could be produced.
Westside Barbell’s “Conjugate Method”
The Conjugate Method attempts to balance all phases across preparation for competition. At the “enterprise level” three movement patterns are continuously tested as the measure of the process. At the “business level” a new variation of a similar movement may become the focus for 3 to 5 weeks (e.g. training rack pulls instead of full deadlifts when “lock out”, the upper portion of the movement, is the weak link). At the “team level” (the lifter + coach), the two-week sprint has a consistent set of ceremonies and artifacts (workout plan, workout log, the workout, etc).
Here is an example:
Week 1
Monday – Max effort lower body day (squat + low back + hamstrings), focus on strength and power
Wednesday – Max effort upper body (bench press), focuses on strength and power
Friday – Dynamic effort lower body (squat, deadlift), focuses on speed and hypertrophy
Sunday – Dynamic effort upper body (bench press), focuses on speed and hypertrophy
Week 2
Monday – Max effort lower body day (deadlift + low back + hamstrings), focus on strength and power
Wednesday – Max effort upper body (bench press), focuses on strength and power
Friday – Dynamic effort lower body (squat, deadlift), focuses on speed and hypertrophy
Sunday – Dynamic effort upper body (bench press), focuses on speed and hypertrophy
This correlates nicely with “core” Scrum concepts:
- Maximal strength is tested every week – working software every sprint
- The metric (1-rep max / story points delivered), is improved (strength / velocity over time), through hypothesis and experiments (empirical process control)
- The entire body is trained for size, stamina, strength, and power per every week – vertical slicing and user stories
- The lifter gets to experiment with new exercises without fear of wrecking a 15-week cycle – sprint retrospective, sprint planning
- The coach focuses exercise planning on addressing weak points – a ScrumMaster, removing impediments
- The Power Lifting competition is not a unique event with a long lead time – working software every sprint, TDD, XP, continuous integration and release
Now the lifter, like our Scrum team, gets to plan, experiment, and deliver often. The overall roadmap (Lean + Scrum) might have a basic end-game or vision (increasing 1-rep competition max performed on 3 lifts the same day is equivalent to convergent product delivery), but planning only looks forward up to 5 weeks, commitment at 1 to 2 weeks. Likewise, the lifter and coach is always looking at the most recent data, the newest lessons learned, and quickly reacts to whether a behavior, practice, or process should be continued or not – just like the Product Owner, ScrumMaster, and Team are always planning and executing based on the most recent market and team data.
Applications to the SDLC:
Now we can extend the metaphor and draw conclusions. The powerlifter’s body equates to a complex large-scale digital portfolio. The lifter needs to increase value three programs that focus on convergent product delivery while also developing several programs that utilize emergent product delivery. In waterfall these two program methods are separated by functional division and project lifecycle, in conjugate (Scrum) these two are handled in tandem.
For the powerlifter, the three convergent products are squat, deadlift, and bench press. Quality must stay constant or the increase in value does not qualify. The same is true in software products – adding a high-value feature while allowing a 50% increase in crash on launch is absolutely unacceptable. Your users will disqualify you! Whether your have a three-application enterprise CRM program or a three-iOS app consumer program (see LinkedIn or Facebook as examples), adding an exciting feature to an app that causes mass user drop out is a risk no business can tolerate in today’s market. The competition is too fierce, barrier to entry too low; someone will blow you away.
At the same time, the powerlifter needs to maintain several emergent delivery programs, some for function (increasing grip strength), some for fun (increasing bicep size). Ongoing workout plans, building size, stamina, and maintaining joint health, addressing weak points by focusing on a new accessory exercise for 5 weeks – all of these priorities must be balanced and evolved. Keeping a workout log is the only way to be sure that exercise volume, intensity, and density are increasing. The relationship between the convergent product value and the emergent product investment is the only metric rationally applicable. The same is true in software delivery. Emergent-delivery programs like R&D, marketing, UX, product planning are all critical to the health and success of the portfolio as a whole – but the end goal must be clear.
- Over-planning and under-delivering is not acceptable.
- Over-researching and under-user-pleasing is not acceptable.
- Over-designing and under-testing is not acceptable.
- Over-marketing and and under-releasing is not acceptable.
Conclusion:
The Conjugate Method as an analogy for Agile, Scrum, XP, and Lean at scale works for me because I love lifting. I realize it may not be right for you, especially if neither agile or weightlifting are familiar territory. So, like everything, find how this applies to your life so that you can find inspiration in ordinary – then start a conversation about it. I’m happy to discuss anytime: 224.223.5248
41.878114
-87.629798