Wishpond

Marketing automation platforms are priced per contact or per seat, but the work that determines whether an agency makes money is the number of live journeys, integrations, approval gates, and exception paths it has to maintain. Agencies that quote on seat count rather than journey count absorb every added branch as unpaid labor, and the gap widens as agentic workflows push more decisions into automated paths. The fix is a documented operating baseline at handoff: named journeys, named owners, and a maintenance schedule the retainer actually funds.

Journey Debt Ratio measures how many workflows are running in a client account against how many are documented, owned, and reviewed. The category's unit of work is the workflow, not the send, so every undocumented nurture sequence, behavioral trigger, or sales handoff is a liability that surfaces when the original builder leaves. Agencies feel this first: a retainer priced on three journeys quietly carries eleven, and the gap is invisible until a trigger misfires on client-facing email. The failure mode is well evidenced. Pact0 ran 13 blind agent sessions against its own onboarding flow and only 3 of 11 agents completed the full loop, roughly 27%, which is what happens when multi-step logic runs without tracing or review checkpoints. Audit each account by counting live journeys, then count the ones with a named owner and a written trigger map. Anything above a 2:1 ratio means the retainer is under-scoped or the delivery team is absorbing unpaid maintenance.

Autonomy tiering assigns every client-facing workflow a level from 1 (drafts only, human sends) to 4 (acts unsupervised on CRM and messaging data), then sets review checkpoints and monitoring to match. The tier, not the tool, determines liability. Agencies that tier explicitly can sell higher-autonomy builds at higher retainers because the risk is priced, documented, and bounded; agencies that skip tiering discover the ceiling during an incident. The evidence for caution is concrete: Pact0 ran 13 blind agent sessions against its own onboarding flow with a 30-turn limit and only 3 of 11 agents completed the full loop, roughly 27%. A WhatsApp broadcast flow built on AiSensy or a nurture sequence in ActiveCampaign may sit comfortably at tier 2, while an agent that writes lead scores into a client CRM belongs at tier 3 with sampled human review. Tier first, then choose the platform.

Handoff Fidelity Score measures how much of a marketing automation build survives the moment an agency stops touching it. The framework scores four dimensions on a 1 to 5 scale: documented journey logic, CRM field integrity, exception handling coverage, and a measurable operating baseline the client can read without the agency. A workflow that scores 4 or higher on all four can be maintained by a client-side ops hire; anything below 3 on two or more dimensions is a support liability disguised as recurring revenue. The category description is explicit that a complete handoff and measurable operating baseline matter more than claims of indefinite recurring revenue, and this score operationalizes that test. Consider a nurture build in ActiveCampaign or Brevo: if the client cannot explain why a lead exited a sequence, the handoff failed regardless of how many journeys were shipped. Agencies that score before invoicing the final milestone avoid the 90-day rework cycle that erodes margin on retainer renewals.

Price the engagement on the count of journeys, integrations, approvals, and maintenance obligations, and put that count in the contract before any build starts.

Cap the number of live workflows per client at what your current delivery team can monitor, test, and document weekly, and only add a new channel when an existing journey is retired or fully automated.