Sustainable Pace in Scrum, SAFe and Kanban

“Agile processes promote sustainable development. The sponsors, developers, and users should be able to maintain a constant pace indefinitely.” — The Agile Manifesto Principle

“programmers or software developers should not work more than 40 hour weeks, and if there is overtime one week, that the next week should not include more overtime.” — Extreme Programming

An unsustainable pace is unhealthy. It contributes to burnout, quality issues, and unpredictable results.

If you are an agile leader — do you know whether your teams are currently operating at a sustainable pace? Do you care? Would you rather not know because you’re afraid of the answer?

Beyond having a general idea of the sustainability of pace, perhaps it would help to have a concrete KPI related to it?

If we use the language of OKRs, maybe something along these lines:

Objective Achieve a healthy sustainable pace KR1 — People working reasonable hours AMB (as measured by) hours per week KR2 — People happy about their pace AMB continuous survey KR3 — Plans don’t assume unsustainable pace AMB ability to achieve forecasts without resorting to unsustainable pace measures

The last KR relates to the planning/forecasting approaches we use in agile. Agile approaches leverage empirical planning approaches. They look at the past (Yesterday’s weather) to try and forecast the future. Whether it is PI Planning, Sprint/Iteration planning. Whether by looking at Velocity, Throughput, or Cycle/Flow Times — most approaches tend to ignore how these past results were achieved when using them to predict future capacity.

For example, if our velocity in previous Sprints was 15–20 points, we will probably take on about 15–20 points. But what if these 15–20 points required a herculean effort that wasn’t sustainable?

Similarly — if we just concluded a PI in which the ART achieved a Program Predictability Score of 85% we will be tempted to assume we have a pretty serviceable approach to planning/forecasting. But what if this required killing it through nights and weekends and skipping any sort of Innovation in the IP iteration? Where does this come into the calculation?

If our cycle/flow times are 7–10 days 85% of the time we will be tempted to set that as an SLE (Service Level Expectation) to ourselves. But does that make sense if this was achieved while working 60 hour weeks?

What I’m advising teams/organizations I’m working with is to consider the “sustainability factor” when considering any past results for the purpose of forecasting the future and adjusting accordingly. This isn’t trivial. It requires making sustainable pace an explicit “citizen” of the measurement dashboard and conversation.

We have learned that speed and quality are related. We now understand that inappropriate speed might be at the expense of quality, so we look at a balanced scorecard of speed and quality. Moving forward, we need to add pace sustainability to this scorecard and to the conversation around how much work does it make sense to forecast.

A metric I’ve been toying with is Weighted Predictability/Sustainability:

As you can see here, achieving reasonable predictability scores is weighted down by the unsustainable pace required to achieve them. so a score of 80% is actually weighted down to 53%. This 53% is what should be used for future planning purposes. For example in SAFe I&A and PIP.

First, we need to come up with a good name for this metric / KPI. Flow Sustainability? Pace Sustainability? Work Sustainability? Burnout Risk?
 Are you explicitly measuring anything related to Sustainable Pace? Do you have goals around it? Do you take it into consideration when planning? Please share in the comments!

Improving flow means fewer bottlenecks, faster feedback, and less wasted capacity. Explore flow advisory or let’s talk.