The Five Stages of the Velocity Curve

The cap is ownership. Without an owner there is no vision, without a vision there is no roadmap, and without a roadmap every piece of work is a favor or an emergency. Effort at this stage produces output but not capability, because each delivery leaves the system exactly as it found it.

At Stalled, every dimension reads as absence. Digital is treated as a cost or a purely technical task, delivery bottlenecks on a person or a committee, work moves sequentially through tickets with frequent rework, content has no reuse or ownership, decisions rest on opinion because nothing is measured, and the platform is aging and carrying heavy debt. The gate reads the same way: no governance, and whatever AI use exists is unauthorized.

The cap is mandate and capacity together. The champion can improve what they touch, but the vision is unowned above them, so process survives only as long as their attention does. Capacity is thin enough that anything new displaces the work of keeping the current experience running, which means improvement happens in the gaps, and the gaps close whenever the business gets busy.

At Starting, every dimension reads as emerging but fragile. A champion exists but the vision is unowned, a few specialists carry the load while the agency does the core, process is forming but handoffs dominate, content standards are partial and metadata inconsistent, analytics are present but underused, and the platform is functional but debt-laden with fragile integrations. The gate shows basic compliance, curiosity about AI without a policy, and risk discussed informally when it is discussed at all.

The cap is the operating model itself. The practices that make Steady reliable are the same practices that hold it in place: marketing and IT work in series because the handoff is controlled, content is templated because templates are predictable, and the roadmap is release-driven because releases are what the machine was built to produce. Every part of the system is optimized to run, and no part of it is designed to change how the system works. Reliability also removes the internal alarm. Nothing is visibly broken, so nothing justifies the disruption of rebuilding the way work moves, and the organization stays exactly where it is.

Steady has a signature: every dimension reads as competent but static. An owner exists but the roadmap is release-driven, roles are defined but capacity is capped at running the current experience, the cadence is repeatable but marketing and IT work in series, content is structured and documented but built to be maintained rather than recombined, reporting is reliable but reactive with little testing, and the platform is stable and well run but template-oriented. The gate is the tell: platform governance exists, AI-safe use cases have been defined and reviewed, and none of them have been deployed. An organization that recognizes itself in four or five of these rows at once is almost certainly at Steady, because the signature is the pattern, not any single row.

The cap is that the gains are still sponsored rather than self-sustaining. The sponsor's decision rights, the change program, and the governance cadence are what hold the new operating model in place, and if that sponsorship moves on before the practices become the team's own, the system can slide back toward series work and release thinking. Progress to Compounding requires the team, not the sponsor, to own the vision and revise it.

At Accelerating, every dimension reads as invested and governed. The sponsor holds decision rights with an outcome-tied roadmap, the cross-functional team has slack to experiment, work runs in parallel with templates and automation under managed change, content follows a reusable model across channels with a governed editorial workflow, experimentation and personalization are routine, and the platform is modern, composable, and integrated with debt retired. The gate covers testing, and risk-managed AI augmentation is deployed.

Compounding sustains itself through reuse and renewal together. Reusable components and evergreen content mean every delivery adds to a library that makes the next delivery cheaper, which is the mechanism behind the curve's steepening. Renewal keeps the library from ossifying: the team sets and revises its own vision, continuously upskills, and swaps platform components as needs change rather than rebuilding around them. Governance holds the pace, because leadership at this stage governs risk rather than approving work, and adoption of new capability, including agentic AI, moves at the speed the audit cadence can absorb.

At Compounding, every dimension reads as self-renewing. The team sets and revises the vision while leadership governs risk, the team is self-sufficient and AI-fluent, delivery is continuous with cycle time compressing, structured content is evergreen with the supply chain optimized, measurement is continuous and predictive with personalization at scale, and the platform is flexible, swapped and extended as needed. The gate shows mature AI governance, deliberate agentic adoption, and an audit cadence.