The Production Operating System
Map and redesign the workflow from marketing request to production launch, identifying bottlenecks, unnecessary handoffs, waiting time and duplicated work, and establishing clear ownership at every stage. Define what information must exist before work can begin, create work-in-progress limits and prioritization rules, and build separate production lanes for urgent changes, standard pages, experiments and complex technical projects so that urgent work stops destroying the predictability of planned work. Create clear escalation paths for blockers and for requirements that change mid-build.
Requirements and Task Definition
Build standardized briefs for landing pages, checkouts, funnels, upsells and technical changes, and establish a Definition of Ready for every category of work so that incomplete or ambiguous requests never enter development. Ensure every requirement carries complete copy, design direction, references, responsive behaviour, tracking requirements and acceptance criteria, and translate marketing intent into specifications that developers and page builders can execute without repeated clarification. Create a controlled process for requirements that change after development has begun.
Audit the full production process for opportunities to use AI and automation, evaluate AI coding tools, AI page builders, no-code platforms, reusable components and template-based systems, and determine the right balance between developers, page builders, designers and AI-enabled operators. Build workflows where AI converts approved briefs, copy and designs into high-quality first drafts, install guardrails that allow non-developers to build safely without damaging performance, tracking or brand standards, and reduce repetitive development through reusable sections and approved design systems. Measure the actual impact of AI on cycle time, cost, quality and throughput rather than assuming it.
Dependencies and Capacity
Create visibility over the workload and capacity of development, design and other shared resources, identify dependencies before work enters production, and sequence work based on when copy, graphics, video, development and approvals will realistically be available. Prevent developers from being assigned work that cannot be completed because a dependency is missing, build a capacity planning system that lets marketing understand what can actually be delivered and when, and surface constraints early enough for the business to reprioritize or find an alternative. Reduce reliance on informal messages and on individual managers holding critical information in their heads.
Quality Built Into the Process
Move quality control upstream instead of relying on final-stage inspection. Introduce developer self-QC before any task is submitted, create standardized QC checklists by page and funnel type, define acceptance criteria before development begins, and establish an independent final QC process that does not depend entirely on one person. Implement automated testing wherever practical, including visual regression, broken links, responsive layouts, tracking, checkout functionality, browser compatibility and page performance. Categorize defects, track root causes, and ensure repeated defects change the underlying system, checklist, component or training rather than being corrected one at a time.
Team Structure and Capability
Lead the existing production team and raise its operating standard. Determine which work should be completed by developers and which should be completed by AI-enabled production specialists, and make sure capability is matched to the work. Coach the team on estimation, task clarification, self-QC and AI-assisted workflows, set clear performance expectations for speed, predictability and quality, and build enough documentation and shared visibility that the function does not depend on any single person holding critical context.
Build and maintain reporting on request-to-launch cycle time, active production time versus waiting time, throughput by work category, on-time delivery, clarification interactions after assignment, first-pass QC rate, average revision rounds, escaped defect rate, rework hours, blocker age, the share of work completed through reusable components or AI-assisted workflows, cost per page or funnel, and marketing launch delays caused by technology or creative dependencies. These metrics exist to improve the system, not to generate reporting.
How Success Will Be Measured
Your first thirty days are about establishing a baseline. We do not currently measure cycle time, waiting time, clarification loops, rework or escaped defects, so your first task is to map the current state from request to launch and put real numbers against it, then produce a prioritized roadmap for improving the system. We would rather agree ambitious targets against real data than hand you numbers invented before anyone measured anything.
By sixty days we expect a standardized intake and requirements process, clear Definitions of Ready and Done, service-level expectations by category of work, visibility over development and creative capacity, dependency planning across copy, design, video, development, tracking and approvals, piloted AI-assisted and template-driven workflows, and a live operating dashboard covering speed, quality and throughput.
By ninety days we expect the function to be trending strongly toward a materially shorter median request-to-launch cycle time, the large majority of work delivered by the committed date, the large majority of work passing final QC without another development round, a sharp reduction in issues identified after approval by the CEO or the media buyer, full visibility over active work, dependencies, blockers and capacity, standard page types produced without bespoke development, and higher throughput without a corresponding increase in headcount. Exact targets will be finalized against your baseline.