Using Velocity As An Estimation Tool Rather Than A Commitment
Why Velocity Gets Abused as a Commitment
There are cultural pressures in many organizations to meet deadlines and estimates at all costs. Managers often treat velocity – the amount of work a team completes in a sprint – as capacity rather than as a trend. This leads to velocity being used as a hard commitment for planning and staffing projects, rather than as a flexible estimation tool.
Cultural Pressure to Meet Deadlines and Estimates
Project managers are frequently measured based on their ability to deliver according to the initial committed plan. There are often organizational incentives and cultures that reward meeting estimates and punish missing deadlines, regardless of external circumstances. This pressure leads managers to treat velocity as a commitment in order to appear able to meet forecasts.
Velocity Treated as Capacity Instead of Trend
Because velocity represents the work a team has historically completed, it is often misused as a representation of the team’s capacity for future work. However, capacity suggests a consistent maximum output, while velocity is meant to provide insight into trends and averages to inform probabilistic forecasting. Treating velocity prescriptively as capacity ignores variability and leads to unrealistic plans.
Problems With Velocity as a Commitment
Velocity Varies and Is Unpredictable
Numerous internal and external factors lead to velocity varying significantly across sprints. Bugs, technical debt, team changes, employee leave, training needs, and task uncertainty all affect how much work a team can handle sustainably. Using yesterday’s weather as an exact predictor of tomorrow is poor meteorology, just as last sprint’s velocity makes for unreliable forecasting.
Causes Death Marches and Technical Debt When Targets Missed
When velocity is codified as a commitment, there is no room for flexibility in plans when reality does not meet predictions. To meet fixed deadlines based on inflated velocities, teams end up in continuous crunch mode, leading to burnout, turnover, and lower quality via technical debt accumulation. By the end, the initially committed timeline and estimates are usually woefully incorrect.
Recommending Velocity as a Planning Tool Instead
Guide for Forecasting Based on Past Throughput
While velocity should not be used as a hard target, it retains value for planning when treated as a guide. By looking at past throughput trends, probabilities can be assigned to ranges of work likely to be completed in a future sprint. Velocity anchors this forecasting while accounting for variability based on historical data.
Account for Variability Via Ranges and Monte Carlo Simulation
By bucketing previous velocities into ranges and frequencies instead of averages, variability can be captured and propagated into probabilistic forecasting models. Using Monte Carlo simulation, many simulations of a project’s timeline can identify the likelihood of ranges for completion given the input velocity ranges.
Update Regularly Based on New Data
No plan survives contact with reality completely intact. By monitoring actual output across sprints and incorporating new data into velocity ranges, forecasts will remain grounded in team evidence rather than fixed abstractions. Continually updating keeps plans just ahead of reality.
Setting Expectations Around Velocity
Velocity Communicates Team Throughput, Not Guarantee
It is vital that stakeholders understand velocity represents the work a team can handle, not a promise for delivery. Much like card dealing rates communicate capability rather than certainty, velocity is a guide not a guarantee. Setting and adhering to these expectations reduces chances of misuse.
Plans Should Have Slack for Discovery and Unknowns
Project plans must include slack time for exploration (spiking solutions), unplanned work (bugs), and unknown activities that emerge in complex projects. Mandating that teams maintain 100% utilization of their velocity allowance leads to unsustainable race conditions and unanticipated delays when surprises arise.
Commit to Sustainable Pace Aligned with Business Needs
Agree with stakeholders that team throughput will align with long-term business objectives rather than short-term cost efficiencies. Mandating excessive velocities may produce interim results but cause costly attrition, turnover, and technical debt in the project later. Commit to pace that is assertively sustainable.
Sustainable Planning With Evidence-Based Forecasting
Leverage Velocity Ranges for Scenarios
Maintain target, conservative, and aggressive velocity ranges backed by historical data to build project plans. Model plans using these ranges to provide scenarios tied to evidence, facilitating stakeholder decision-making given the uncertainties involved.
Simulate Schedules Statistically
By performing Monte Carlo simulations using the various velocity ranges and their probabilities, reasonable project timelines can be defined statistically. Mathematical simulation captures the variability more responsibly than estimating unknowns.
Involve Team and Stakeholders in Planning Process
Collaborative planning sessions across team members and business stakeholders foster shared grounding and consensus on the uncertainties and risks inherent in a project. Rather than enforcing velocity as a commitment, driving participation enables trust and support when deviations occur.
Conclusions
Velocity Flawed When Used as Commitment
Basing project targets and promises solely on velocity ignores the variability teams necessarily encounter in practice. As a commitment, velocity misrepresents true capability and ensures unrealistic plans.
Excel as Input for Forecasting Instead
However, velocity applied as a distribution for estimating team throughput probabilistically provides value. Velocity informs data-driven forecasting without hard mandates that ignore reality’s inherent variation.
Anchor Plans to Evidence and Update with Data
Intertwining velocity ranges, simulations, and stakeholder participation produces plans that accurately reflect uncertainties, facilitate buy-in, and adapt responsibly to ongoing inspection and adaptation. Plans stay anchored in evidence and responsive to new data.