Why process beats heroics
Every team has a few people who can rescue a launch at the last minute. They know which platform setting breaks the tracking, which stakeholder must see the copy before it goes out and where the current logo files are. Teams lean on them because it works, until one of them is on holiday, leaves, or is simply doing three rescues at once. Heroics scale with headcount at best; usually they scale worse, because the rescuers become the bottleneck for every launch.
A designed process moves that knowledge out of heads and into the route itself. The stage that asks for a campaign name in the agreed format at intake removes the evening spent renaming ads before launch. The written procedure for setting up a paid campaign means a new starter gets the tracking parameters right without asking. The gate that requires a signed brief before build stops creative being made against an idea nobody approved. None of this is bureaucracy for its own sake: each element exists because a specific, recurring failure costs more than the element does. If a stage or a check cannot be traced to such a failure, it is a candidate for removal, and good process design removes as readily as it adds.
Workflow, procedures, handovers, improvement
Process design has four parts. They are easy to blur together, and most process problems come from fixing the wrong one.
| Part | What it decides | What it looks like when it works | Where it is covered |
|---|---|---|---|
| Workflow | The stages work passes through and the gates between them | Every item has a visible stage; nothing skips a gate | Campaign workflow guide |
| Procedures | How a recurring task inside a stage is done | A new starter completes the task from the document alone | SOP tutorial |
| Handovers | What passes between people or teams, and in what state | The receiver never has to chase missing information | The core guides, at each stage boundary; Handing over work when someone leaves |
| Improvement | How the process is reviewed and changed | Measures are reviewed on a cadence and stages are retired as well as added | This page, below; Prioritising operations work, Mapping a marketing process |
The workflow is the skeleton and the procedures are the muscle. A workflow without procedures tells people where work goes but not how to do it, so quality varies by person. Procedures without a workflow are a library nobody opens at the right moment. Handovers are where the two meet, and they deserve explicit design: a handover is complete when the receiving person has everything needed to start, not when the sender has pressed send. Improvement is the part teams forget, and the reason documented processes drift away from practice within a few months.
The core guides
The core guides are written to be read in order.
- Campaign workflow from intake to launch draws the end-to-end route: how requests are taken in and triaged, how the brief and kick-off work, how build, creative and trafficking run in parallel, which review and approval gates sit before launch, and what the handover after launch must contain. It is the map on which every other operations fix is drawn.
- Writing a marketing SOP step by step is a tutorial that produces one standard operating procedure for one recurring task, from choosing the task to rolling the procedure out and scheduling its review. It includes a template you can copy.
The published campaign launch QA checklist is the companion to both core guides. The workflow decides where the launch check sits; the SOP tutorial shows how to write a procedure that ends in a checklist of that kind. Checks on tracking parameters, including every UTMURL tags identifying traffic source. on every link, belong in that checklist rather than in someone's memory. For the wider context of process alongside systems and governance, start from the marketing operations hub.
Further guides
Further guides cover the work around the route: seeing it as it really runs, learning from errors, agreeing what requesters can expect, deciding which fixes come first, controlling how much is open at once, and keeping work going when people leave or are away:
- Reviewing a marketing error after the fix — a tutorial for a timeline, the checks that missed it, causes traced past the person, owned actions and a shared error log.
- Mapping a marketing process before changing it — a tutorial for tracing real items, drawing steps and handovers by role, marking waits and rework, then choosing what to change.
- Setting service levels for marketing requests — a tutorial for a catalogue of request types, what each needs, turnaround from your own records and both sides' duties.
- Prioritising marketing operations work — one backlog fed by rework, errors, audits and requests, sorted by effect, risk and effort, with protected time.
- Limiting work in progress in a marketing team — why a team with too much work in flight finishes less, and how to limit it: seeing what is open, limits by stage, blocked work and pulling new work in.
- Handing over marketing work when someone leaves — a tutorial for recurring tasks, work in flight, accounts and admin rights, files, suppliers and unwritten know-how.
- Reducing key-person risk in a marketing team — finding work only one person can do and reducing the risk: a coverage matrix, admin rights in more than one name, procedures and tested cover.
Measuring a process: cycle time and rework
A process that is not measured cannot be improved except by opinion, and opinions about process are usually shaped by the last bad week. Two measures do most of the work, and both can be recorded in any work management system that tracks stage changes.
Cycle time is the elapsed time from an accepted request to live work. Break it down by stage and the picture becomes useful: most of the time a campaign spends in a workflow is waiting rather than working, and the stage where it waits longest is usually a review queue or an asset request. Shortening the work itself rarely moves the total; removing the wait does. The theory of constraints and its focusing steps turn that observation into a method: find the stage that limits throughput, get the most from it, pace the rest of the workflow to it, then add capacity there. Waits also grow when too much is open at once, and the guide to limiting work in progress sets limits by stage or by person from the team's own records and pulls new work in only when something finishes.
Rework is the share of items sent back from a gate to an earlier stage. A high rework rate at the approval gate usually means the brief was weak, not that the reviewer is difficult. Rework at the launch check usually means the build procedure is missing a step. Each returned item should carry a reason, so the pattern can be read.
Treat both as operational measures, not as a KPIA metric tied directly to an objective. for the people in the process. If cycle time becomes a target for individuals, gates get skipped to hit it. Changes to a process can be tested much as campaigns are: run the new stage on one stream of work, keep the old route on another, and compare over a reasonable period. It will not be a clean A/B TestControlled comparison of two variants., but it beats changing everything at once and guessing which change helped.
A furniture retailer reads its cycle time
Oakhurst Living, an online furniture retailer, logged stage changes for every campaign for one quarter. The median cycle time was 19 working days, of which 8 were spent in "awaiting brand review" and 4 in "awaiting assets". Rework at the brand review gate was 41% of items, and the most common reason given was "offer wording not agreed". The team added an offer field to the intake form that required sign-off from the trading manager before a brief could be written. In the following quarter, rework at brand review fell to 17% and median cycle time to 13 working days, with no change to the reviewer or the build team.
Common process problems
Most process complaints are one of a few symptoms. The table routes each to the guide that fixes it.
| Symptom | Usual cause | Guide |
|---|---|---|
| Requests arrive by email, chat and corridor | No single intake route with required fields | Campaign workflow |
| Creative is rebuilt after review | Brief not signed before build; reviewers see work too late | Campaign workflow |
| The same launch mistake keeps recurring | The task has no written procedure, or the procedure lacks a check | Writing a marketing SOP, Reviewing a marketing error |
| New starters ask the same questions | Procedures exist only in people's heads | Writing a marketing SOP, Building an onboarding pack |
| Tracking is wrong on live campaigns | Launch check missing or skipped | Campaign launch QA checklist |
| Documents describe a process nobody follows | No owner and no review date on procedures | Writing a marketing SOP |
| Nobody knows what happens after launch | No post-launch handover defined | Campaign workflow |
| A problem on a live campaign ran for days before anyone saw it | No named monitor, unclear limits on changes, no cover, no normal range, silent tracking or no route for reports | Live campaign problems found too late |
| Much is started and little is finished | Too much work open at once; new work pushed in before current work finishes | Limiting work in progress |
| Requesters cannot tell when work will be done | No agreed turnaround by request type; the requester's side never written down | Setting service levels for requests |
| Work stops when one person is away or leaves | Recurring tasks and admin rights held by one person; cover never tested | Reducing key-person risk, Handing over work when someone leaves |
When a process symptom has already reached the reports, such as campaigns that cannot be grouped or a monthly pack that takes days, the marketing operations troubleshooting section diagnoses it cause by cause.
