Agile teams love a clean forecast. A velocity chart, a burndown line, a confident date on a roadmap slide. But ask any delivery lead why a release slipped and the answer rarely involves coding speed. It’s almost always something else: an approval stuck in someone’s inbox, a compliance check that took longer than expected, a verification step nobody accounted for.
Forecasting models built purely on story points or historical velocity assume a frictionless system. That assumption is the problem. Work doesn’t move in a straight line through most organisations—it moves through gates, and those gates are where predictability quietly falls apart.
Why Forecasts Fail At Handoff Points
The handoff between “done” and “actually deployed” is where most forecasting models lose accuracy. A team can estimate coding effort reasonably well, but estimating how long a change sits waiting for sign-off is a different skill entirely—one most sprint planning sessions never attempt.
This isn’t a minor statistical quirk. A three-point story might finish in four days, while a one-point story sits blocked for three weeks because of a dependency or an approval queue. Size was never the variable that mattered—friction was.
Identity And Verification Friction Slows Delivery Silently
Verification steps are a particularly underrated source of delay because they often happen outside the team’s direct control. Security reviews, identity checks, compliance sign-offs—these processes exist for good reasons, but they rarely appear as a line item in sprint planning. Teams track story completion; they don’t track queue time in someone else’s approval workflow.
Industries built around high-stakes verification have had to get this right at scale. Payment processors handle identity checks in seconds without disrupting checkout flow. Banking onboarding platforms run document verification in the background while users complete other steps. Digital identity providers route approvals through automated pipelines that flag exceptions without pausing the entire journey. Casino platforms have pushed this further — the GamblingInsider review of no kyc casino sites shows how verification flows have matured into structured, predictable processes, with clearly mapped user journeys and minimal friction at each checkpoint.
Low-Friction Systems Outside Agile Offer Lessons
Software delivery has its own version of this challenge, and the data is sobering. Even with AI accelerating how quickly code gets written, enterprise delivery research shows median lead time to production still sitting at 30 to 45 days, with some business-critical features taking well over 200 days once manual approvals, brittle test suites, and governance processes are factored in. Coding was never the bottleneck. The gates around it were.
That distinction matters for how teams build forecasts. If the model only accounts for effort and ignores the queues, sign-offs, and checks surrounding that effort, the forecast is measuring the wrong thing entirely.
Building Friction-Aware Forecasts For Scrum Teams
The fix isn’t more precise story pointing. It’s shifting toward flow-based forecasting that treats throughput—including blocked time—as the primary signal. Teams need only ten to twenty historical data points to run probabilistic simulations that reflect how work actually moves, gates included.
Organisations that have adopted this structured approach to phase transitions and approvals show dramatically different outcomes. Software delivery timeline data indicates that while typical teams see delivery variance of 50 to 100 percent against committed dates, organisations with mature handling of approvals and phase gates keep that variance under 20 percent. The difference isn’t talent or tooling—it’s whether friction was built into the forecast from the start.
Scrum Masters and delivery leads who want more reliable timelines should start by mapping every approval, check, and verification step their work passes through, then measure how long each one actually takes. That data, uncomfortable as it might be, is far more predictive than another sprint of velocity tracking.
