Sustainable pace was never about pace
We were working on an organisational change programme at a global research and advisory firm, inside one of its largest business units. Nine teams did the product and technology work there, on a revenue base of hundreds of millions of dollars every year, and they had been told to double it. Along the way the programme needed evidence, so we produced an assessment: twenty-five dimensions of scored questions, and a qualitative pass that captured what the numbers couldn’t hold.
The analysis started with the map.
Start with the map
Correlation, not causation: every edge is a measured statistical association (p<0.05) from this engagement, not an asserted cause. Edges below |r|=0.35 are omitted for legibility; the rest are drawn thicker and more opaque the stronger they are. The map stays focused on Sustainable pace: hover any dimension to preview its own connections against it, and use the strength slider to keep only its strongest links.
We noticed a cluster of strongly connected dimensions with red concentrated inside it, and one prominent node sitting in the middle of that neighbourhood: sustainable pace, the dimension people wrote about more than any other, carrying fourteen strong connections of which all but one sit on the red side of neutral. In a connected system you don’t act where the pain is loudest; you act where the connections concentrate. That is a leverage point, and it is why this dimension, not the lowest-scoring one, became the centre of the analysis.
Look at what the neighbourhood is made of: product management roles, testable requirements, quality metrics, technical debt, automated testing, team-level planning: the mechanics of how work arrives and ships, almost all of it on the red side of neutral.
The rest of this page walks it. The question behind it, asked plainly:
How would you rate the team’s ability to maintain work-life balance while meeting project goals?
What the numbers said
Median 7 out of 10 came back from the scores. If you stopped here, and most assessments stop here, you’d file this under “fine” and move on. The distribution is wide, but the weight of it sits comfortably above the middle.
What the words said
The two dark blobs not lining up is the chart.
Written observations told a different story. Of twenty coded observations on this dimension, sixteen were negative and only two positive. On the same scale, the writing sits around 3 where the scores sit around 7.
“Teams are working day and nights, often on weekends when a major event or upgrade is planned.”
“They are always overburdened with something or the other… at the end it affects work-life balance and daily routine.”
“Some teams are at times required and expected to work very late, as late as 1 or 2 am their time during the busy season.”
That gap between what people score and what they say is the finding. People rate the system they’ve adapted to. They describe the system as it is.
The clearest tell is how often both halves live in a single sentence:
“Team members are very dedicated. However, everyone has a large workload and split focus in their positions.”
Dedication first, strain second. A culture this committed rounds its own pain upward, which is exactly why the scores stayed comfortable.
Where it lived
Nine teams, anonymised, best to worst. Thin rows render wider and flatter; less data looks uncertain, not falsely precise.
Pace wasn’t evenly distributed. Two teams sat comfortably at 8-9. One team sat at 3; its whole distribution shifted left, not just a few unhappy voices. The written comments clustered on exactly the teams the scores flagged, independently. When the numbers and the words point at the same team from two different directions, you stop debating the instrument and go look.
What pace turned out to be entangled with
Back to the map’s finding, now with the numbers on it. Here’s the part nobody in the room predicted. Sustainable pace didn’t correlate most strongly with workload, staffing, or delivery pressure. Its strongest statistical neighbours were:
- Role clarity in product management (r = .63), the strongest pull in the network
- Testable requirements (r = .57)
- Quality metrics (r = .56)
Every one of those is upstream of the evening. Unclear ownership means work bounces until someone absorbs it after hours. Untestable requirements mean rework arrives at the worst time. Weak quality signals mean regressions show up as emergencies. The teams weren’t burning out because there was too much work; they were burning out because the work was shapeless.
You can hear the entanglement in the observations themselves. The strain quotes rarely complain about volume; they describe shape:
“Product owners create monolithic requirement docs that need to be translated into individual stories.”
“Product ownership responsibilities seem to have spilled over into project management and lead developers.”
“The QA team frequently works extended hours to complete regression testing on time.”
That last one is coded under sustainable pace, but read it again: it is a testing sentence. The dimensions aren’t just correlated in the numbers; they are entangled in how people talk.
Unsustainable pace is rarely a workload problem. It’s a clarity problem wearing a workload costume.
What happened next
The scores on their own would have been comforting, and that is about all they would have been. The written observations confirmed what leadership already suspected, but the connections between dimensions turned suspicion into a target list, and that’s where the effort went: serious investment in CI/CD and automation, plus training aimed at the product and portfolio teams. One finding resolved itself in an unexpected way. Remote and onsite teams were showing different outcomes, and the difference traced back to proximity: relationships formed differently when people shared a room. Then COVID sent every team home at once, and that gap shook out on its own. The organisation didn’t stop watching once the training landed; it kept an eye on the teams and adjusted its focus as things moved. What happened to the numbers after that isn’t mine to report, that belongs to them. What is worth taking from this is the method itself. They acted on where the data connected, not on its averages.
A note on the data
A real engagement; an unnamed client, and it stays that way. Team names and identifying details are removed, and quotes are lightly edited for anonymity. This page states no numbers from memory: every chart and figure is built fresh out of the underlying assessment data, nearly 2,000 data points, qualitative and quantitative both.
Read Next
- Measuring Technical Debt in SAFe
Technical Debt is the strongest whole-estate correlate the container classifier has a headed page for (see the correlation network above).
- Stakeholder management: keeping Business Owners engaged productively
Stakeholder Management is the measured edge (r = .71 with Technical Debt) the pilot edge-insert targets next.