top of page
Search
  • Writer's pictureJohn Clapham

Flow Motifs - What football passing patterns can teach tech teams

Updated: 4 days ago


Getting past an impediment. Is forward the only option?

Photo by Laura Rincón


I like to keep things simple, but not over simplified. Look up ‘flow’ in regard to agile teams and you’ll find an awful lot of large stylised arrows pointing left to right. It’s seductive and easy to believe that work and value flow through teams this way. If we optimise for this it is bound to happen, right?


Football teams have the same desire to continually move forward, although you seldom encounter presentations with big arrows pointing towards the goal. Like organisations they have something else impeding flow, the other team is far more visible, and perhaps more predictable, than organisational impediments.


This is where flow motifs come in, these describe passing (flow) between players, and crucially recognise flow is backwards, forwards, almost any potentially productive direction. It acknowledges the value of multiple side passes that lead to opportunities for forward progress. Recognition of these common flow motifs enables analysis of team playing styles and tendencies when faced with different circumstances and objectives. (Håland, 2020)


Above: Football passing motifs identified by Peña & Navarro. Rather than being between players, these could readily represent flow of work between teams. The A,B,C,D motif is what we often try to force, neglecting other useful flows.


I think we can utilise this concept, also described as network motifs, during team and organisation design, to further improvement and the general business of getting things done.

The first step is to acknowledge that, no matter how much we want it, it’s a fallacy to assume all work will only move forward and we should design solely for that circumstance. (I’m sure this one is already recognised, right?)

Chasing this nirvana is likely to be as successful as a naive football team which refuses to pass or move in any direction other than forward. I've seen it lead to accumulating tech debt as work passed back is de-prioritised or ignored in favour of pushing forward, and idealised visions of how we 'should be' working.


The second step is to observe and recognise flow motifs in the organisation, this means common interaction patterns between teams, both formally recognised and amongst influential groups. From this we have a visualisation for discussion and to determine if these patterns are useful for our goals and how we might adjust and improve.


This might be extended by a kind of gap analysis - between motifs we habitually see, and how we’d like to work. This enables a more intentional, lower risk change in operating model compared to the more common ‘throw everything good away with the bad and try to mimic what someone else is doing’ approach.


The wall of confusion, not so much 'flowing' as 'lobbing'. Source: http://dev2ops.org/2010/02/what-is-devops/

A good example of this is the original motivation for DevOps. Companies worked with the assumption that work moves in one direction, across the infamous wall of confusion, which caused headaches (and a lot more) for Ops when they received inoperable components. In the language of Flow Motifs, DevOps recognises the common pass forward pass back behaviour between Dev and Ops, and suggests this is desirable to the extent that we should improve collaboration (back and forth flow) and even smush the teams together.


Of course not all motifs are as obvious or ubiquitous, the greater the number of teams involved, the more useful a motif becomes. Consider five teams iterating on aspects of a product, you’d get a circle shaped motif, with crossing spokes, and multiple cycles. This would demonstrate a healthy situation where each team uses it’s specialist skills to improve the product each time around. Not all relationships are cyclic, Dev, Test, Sec, Ops interactions are more complex, like players on the field the idea is to be ready and predict when and where the pass (work) will arrive, so it can be handled quickly and effectively.


The interesting part would be the discovery that the work frequently rests longer at one of the teams, or that there is hidden interaction with a sixth team. Frequency of iteration with other teams is also significant. These relationships, good or bad, don't necessarily surface in single team metrics, especially where work is represented as a new item when it enters a new team or crosses a boundary.


This is particularly important during change and transformation when teams declare themselves to be a particular archetype and defend against specific work types or interactions. The tendency is to define what they do without considering all dependencies, and therefore neglect relationships which are useful and currently necessary.

This is similar to a goal keeper, mid-field player and a striker loudly stating their purpose but not figuring out how to work together or respond to their current situation. There is a reason players play and practise together, and there’s a reason teams should too.


Closing thoughts

It's important to design systems with the intent of continuous forward flow, but obsessive focus on forward motion can be to the detriment of team's capabilities. This is not unlike a puppy putting all its energy (and increasing frustration) into pushing through a fence while neglecting to look sideways for a gap.


Another point we can take from football is that while specific flow motifs may be likely or desirable, they are not mandatory, and will be more useful in some situations than others. We still need room for improvisation and novel approaches. It's a notation to encourage reflection and improvement, rather than a set of rules. What we can do is talk about likely interactions. Team A will frequently pass work back and forth with team B and require sporadic input from team C, so let's acknowledge that and reduce friction. If a new interaction pattern (motif) is desirable then let's work towards in manageable, deliberate steps using a dialogic approach.


There are plenty of similar concepts in the community, but I'm fond of the term flow motif, it intuitively speaks to patterns of interaction that include more than just neighbouring teams. It highlights that flow is not always left to right towards the user and that passing work back for iteration is a healthy part of product development. They provide a little more fidelity than over-lapping circles and allow us to add time and level of service attributes.


Flow motifs provide useful sketches for orientation and new starts, helping to articulate make sense of the team's context. They also invite a nuanced conversation, because they follow work, rather than team structure, making them amenable to different collaboration styles and team setups.



References and reading

Håland, Else Marie, Wiig, Astrid Salte, Hvattum, Lars Magnus and Stålhane, Magnus. "Evaluating the effectiveness of different network flow motifs in association football" Journal of Quantitative Analysis in Sports, vol. 16, no. 4, 2020, pp. 311-323. https://doi.org/10.1515/jqas-2019-0097


Peña, Javier & Navarro, Raúl. (2015). Who can replace Xavi? A passing motif analysis of football players. https://www.researchgate.net/publication/279310008_Who_can_replace_Xavi_A_passing_motif_analysis_of_football_players



Comments


bottom of page