Suitability filters are models that offer simplified views of a decision-making process. They help people identify and discuss the variables at play and make a recommendation. The British statistician George Box said it best “All models are wrong, but some models are useful.”
A good model provides insights, usually a basic view of how something works or the interactions between variables. From this perspective, models are interesting and can speed up our comprehension of something. However, the greatest benefit of models is as a group alignment tool.
Much of today’s work is knowledge work. Subject Matter Experts (SMEs) contribute their expertise and collaborate to solve problems or build new products and services. This work is largely invisible and intangible. There are few physical or visible increments in the process. So, when someone describes a requirement, decision, or problem, there is a high chance that other people interpret that description differently.
The Telephone Game and Tree Swing cartoon are classic examples of how even well-meaning interpretations can be distorted, leading to bad outcomes. Left unchecked, our mental models of a product, project or process can diverge from our teammates. This is where visual models offer significant benefits.
Models visualize the previously invisible and, if dynamic, allow us to manipulate the intangible. In doing so, they bring SMEs together by pointing at and discussing the same variables, goals, and interactions we see.
Since knowledge work is now pervasive, tools that can get people on the same page are incredibly useful. This is why I created PMillustrated.com to explain project management to visual learners, and I am a fan of the PMI product Wicked Problem Solving (WPS). They illuminate the invisible and allow us to manipulate the intangible nature of knowledge work. Models unite stakeholders and help avoid divergence.
Project Plans, Release Roadmaps, and Kanban Boards are examples of familiar visual models designed to help get people on the same page about timelines, what will be released, and work in progress. They are all simplifications since listing every task or feature for non-trivial projects or products would not be practical – yet they are still valuable.
Project plans and release roadmaps are mainly static. While we can edit them to pull and push at the variables (scope, people, estimates), it is generally discouraged since we quickly lose the original view created by the project manager or product owner. Kanban boards are deliberately more interactive. Team members move their own tasks, select new items to work on, and we can see the impacts on queue length, WIP and maybe cycle time or throughput.
Models that invite more people to interact, capture more diverse insights, facilitate better conversations, and generate stronger consensus to the final configurations. So, ideally, good models are interactive, easy to reset and invite manipulation and conversation.
“Good models are interactive, easy to reset, and invite manipulation and conversation.”
I have long been a fan of approach suitability filters. Not because these are infallible or inherently superior to letting people discuss how to undertake something. Instead, because they make the invisible factors visible and get people talking about what common aspects are significant.
Without suitability filters, people can be stuck in old ways or recommend unsuitable approaches because “that’s just how we do things here”, “what our standards say”, “a cool new approach I just read about”, etc.
My first exposure to suitability filters was in 1994, working on the DSDM Project and Organizational Suitability Filters. I have written about the evolution of suitability filters several times on this blog. This comprehensive list post dates from 2007. More recently, I created the Agile Suitability Filter for the Agile Practice Guide by synthesizing elements from Boehm & Turner’s model, DSDM’s Organizational Suitability Filter and new elements.
The trouble with these models was that they were published as Standards, books, or static web pages. This meant people had to print and hand them out or implement them as spreadsheets to share for use and get feedback. These clunky implementations hampered the quick reset and shared manipulation we want from a good model. DA also has some good material around scaling factors, but they are currently static models with links to static pages of text. There is no interactivity or feedback gained from manipulating values.
Web-based tools are less likely to get lost or need preparation before use. Anyone with a web browser can access them, so they promote use by many stakeholders. Plus dynamic web controls also allow users to interact and manipulate the model parameters and see, then discuss, how the model responds.
1) The Beyond Agile Model
The Beyond Agile Model is a web-based tool initially sparsely populated with suggestions for approaches, training, webinars, etc. Dragging the Project Characteristics sliders expands or contracts the dotted-red line depicting the Recommends Lens. Choosing the “Simple”, “Medium”, or “Complex” options from the Choose Model drop-down adds more sliders.
My book “Beyond Agile: Achieving Success with Situational Knowledge and Skills” describes the model, which is fully reconfigurable based on a public Google Sheet spreadsheet as a starting point. You can copy the sheet, amend it to your organizational needs, save it privately, and point the tool to this new private version. Then continue adding or amending Recommendation Types (Approaches, Training, Artifacts, etc) Recommendations (Team Charter, Release Roadmaps, etc) and Project Characteristics you wish to model (such as Team Size, Criticality, etc.)
Organizations use it to describe their proprietary approaches or create hybridized implementations of DA and SAFe with other techniques. You can read more about the Beyond Agile Model here and here and here and here, or just get the book for all the info.
2) Agile Suitability Filter
This is a pilot implementation of the Agile Suitability Filter published in the Agile Practice Guide. Making the model available online makes it quicker to share, easier to reset and open for all to play with and discuss as a group.
3) Citizen Developer (CD) / Low-Code No-Code (LCNC) Suitability Filter
This is a new pilot filter based on the existing Citizen Developer suitability assessment 2x2 matrix and personal experience. Are the assessment parameters complete? - Probably not. Are they helpful in discussing with stakeholders about the risks and opportunities of using LCNC – Probably.
What factors do you assess when determining if a product or problem is a candidate for LCNC? I would love to hear from you and create alternative models for testing.
Likewise, what are your thoughts on dynamic models to help facilitate better conversations and decisions? What models do you use, and what would you like to see developed?