The workflow in one picture
Seen from above, almost every campaign workflow has the same shape, whatever the channel and whatever tool holds the work. A request comes in and is triaged. Accepted requests get a brief, which is agreed at a kick-off. Build, creative production and trafficking run, often in parallel. The finished work passes one or more review and approval gates. A launch check confirms that what is about to go live matches what was approved. After launch, the campaign is handed to whoever will monitor and optimise it.
| Stage | Entry condition | Exit condition | Typical owner |
|---|---|---|---|
| Intake | A request using the intake form | Accepted, declined or parked, with a reason | Operations or the channel lead |
| Brief and kick-off | Accepted request | Brief signed by the requester and the campaign owner | Campaign owner |
| Build, creative, trafficking | Signed brief | All assets and set-ups complete against the brief | Specialists in each channel |
| Review and approval | Complete build | Approved by every required reviewer, or returned with reasons | Reviewers named in the approval matrix |
| Launch check | Approved work | Check passed and recorded | Someone other than the builder |
| Handover | Live campaign | Monitoring owner confirmed, measures and review dates set | Campaign owner |
Two features of this picture matter more than the number of stages. First, every stage has an entry condition, so work cannot drift forward half-ready. Second, every gate has a named owner and a defined output, so "waiting for sign-off" always means waiting for a specific person to make a specific decision. Where either is missing, work stalls without anyone noticing until the launch date arrives. To see how far your own campaigns depart from this picture, mapping a marketing process traces a few recent items through the route as they actually ran, waits and loops back included. When campaigns keep missing their dates, the diagnosis for campaigns that keep launching late reads that route stage by stage to find where each late launch waited.
Intake and triage
Intake is the stage that does most to protect everything after it, because each piece of missing information at intake becomes a question asked later at greater cost. A good intake routeThe single route through which every request reaches the marketing team, usually a short form, so work is triaged and checked before it starts. has one entry point, usually a form in the work management system, and a short list of required fields. The test of a field is whether its absence has delayed a campaign before; if not, leave it out, because long forms push requesters back to email.
| Intake field | Why it is required | What happens without it |
|---|---|---|
| Objective and audience | Decides whether the request belongs in marketing's plan at all | Work starts on requests nobody would prioritise |
| Deadline and the reason for it | Separates fixed dates from preferences | Every request is urgent |
| Channels requested | Tells triage which specialists are needed | Capacity is discovered late |
| Budget source | Confirms the spend is available | Build completes and launch waits for money |
| Offer or claim, if any | Flags work that needs legal or compliance review | Review happens after build, and build is redone |
| Proposed campaign name | Starts the naming convention at the source | Names are invented at launch and differ by platform |
| Success measure | Defines what reporting will look for | The campaign launches without a target |
Triage is the decision on each request: accept, decline, park until a later window, or route to the fast lane for low-risk repeatable work. Which work qualifies for the fast laneA written, shorter approval route for low-risk, repeatable or urgent work, such as replies or posts from approved templates; anything risky leaves the lane., and what sends an item back to the full route, is written down before triage relies on it. Triage needs a person with the authority to say no, and a regular slot rather than a continuous trickle; a short triage meeting a couple of times a week is enough for most teams. Deciding which accepted requests come first is a planning decision about capacity rather than an operations one, and the guide to prioritising work and capacity covers it. What operations owns is that every request passes through triage and that the outcome is recorded with a reason the requester can see. Questions for the analytics team follow the same pattern with a different intake, one that asks for the decision behind the question, as running an analytics request process describes.
Brief and kick-off
The brief turns an accepted request into a description of the work that everyone can build against. Its structure belongs elsewhere: the campaign brief guide explains what a strong brief contains, and the campaign brief template gives a ready format. The workflow question is narrower: when is a brief finished, and who says so?
A brief is finished when the requester and the campaign owner have both signed it, and signing means agreeing that the objective, audience, message, offer, channels, budget and measure are right. The measure should connect to a KPIA metric tied directly to an objective. the business already reports, not one invented for the campaign. Changes after signing go back through the same signature, which sounds rigid but is the single most effective way to stop creative being rebuilt. Most reworkThe share of items sent back from a review or launch gate to an earlier stage; each return carries a reason, so the pattern points to the step that failed. at review traces back to a brief that was never agreed, or was quietly changed by one party after the fact. Edits to work that has already passed review follow the same logic, and the guide to handling changes to marketing work after approval sets out which review each kind of change reopens and how small fixes are confirmed.
The kick-off is a short meeting, or an asynchronous thread for small campaigns, where the people who will build the work read the signed brief together and raise questions. It is also where the plan for the build stage is agreed: which assets are needed, who produces each, which ones depend on others, and which reviews will be needed so that reviewers can hold time. Reviewers who learn about a campaign at the review stage will always be slower than reviewers who knew it was coming.
Build, creative and trafficking
Build is where most of the effort goes and where the workflow should intervene least. The specialists know their channels; the workflow's job is to make sure they start from a signed brief, work from current templates and approved assets, and record progress in the work system so that the calendar stays true.
Three things deserve explicit handling. Dependencies come first: copy usually blocks design, design blocks resizing, and every asset blocks trafficking. Recording them in the work system makes the critical path visible and stops the last week being spent waiting. Procedures come second: recurring build tasks such as setting up a paid social campaign, building an email or publishing a landing page should each have a written procedure, which the SOP tutorial shows how to write. Tracking comes third: every link needs its UTMURL tags identifying traffic source. parameters built from the naming convention at this stage, not added on launch day, because a parameter added in a hurry is the most common source of unattributable traffic.
Trafficking, placing finished creative into ad servers and platforms with the right dates, targeting and tracking, is a build task with its own discipline. For programmatic and ad-server work, the trafficking workflow guide covers the steps and checks in detail; the operations point is that trafficking is a stage with an entry condition (approved creative and a placement plan) and an exit condition (set-up complete and ready for the launch check), not an afterthought.
Review and approval gates
Review gates exist to catch errors and risks before they reach the public. They become a problem when there are too many, when reviewers are chosen by seniority rather than by the risk the work carries, or when every reviewer sees every version. The fix is to decide reviews by risk: a matrix that says which kinds of work need which reviewers, maintained by operations and applied the same way every time. The approvals and compliance guide builds that matrix; the workflow only needs to know who is on it for each campaign, which the brief records.
Within the workflow, three rules keep reviews fast. Reviewers see the signed brief alongside the work, so they review against intent rather than taste. Feedback is consolidated: one round of comments from all reviewers in parallel, with one person resolving conflicts, beats sequential rounds where each reviewer reopens the last. And every return carries a reason chosen from a short list, such as off-brief, inaccurate claim, brand standard, legal or technical, so that rework can be analysed. A gate that returns work without reasons teaches the team nothing. Social accounts publish much that never enters this workflow, such as replies and reactive posts, and the social media approval workflow tutorial sets the separate, faster route those need.
Service levels help, provided they are agreed with reviewers rather than imposed. A reviewer who knows the expected turnaround and sees a queue of incoming work can plan; one who receives an urgent request on the afternoon before launch cannot. The tutorial on setting service levels for marketing requests agrees them type by type from the team's own records, with the requester's side written in. When one campaign runs in several countries, the tutorial on rolling out a campaign across markets adds a reviewer for each market's language, commercial message and claims, with deadlines in each market's clock.
Launch checks and post-launch handover
The launch check confirms that what is about to go live is what was approved: the right creative, the right dates, the right audience and budget, working links and working tracking. It is done by someone other than the builder, because builders check what they meant to do rather than what they did. The published campaign launch QA checklist is a ready starting point. The result of the check is recorded against the campaign in the work system, including who ran it, so that a later problem can be traced to a gap in the checklist rather than argued about.
Launch is not the end of the workflow. The handover after launch names the person who will monitor the campaign, the measures they will watch, the dates of the first and subsequent reviews, and what they are allowed to change without going back through approval. Early signals such as CTRShare of impressions that result in a click. and delivery pacing tell the monitor whether the campaign is behaving as expected in its first days; the agreed success measure tells them whether it worked. A campaign without a named monitor tends to run unattended until someone notices the budget has gone. When the campaign ends, the campaign post-mortem template records which stage of this workflow contributed to any shortfall and the changes agreed for next time. If an error reaches a live ad or page after launch, the live ad and page error response SOP contains and corrects it, using the launch check's record to know what was approved. When problems on live campaigns are found days after they began, the diagnosis of live campaign problems found too late reads how each one was found, traces why nobody saw it and routes each pattern to the page that owns its cause.
A software company redraws its workflow
Fenwright, a B2B software company with a marketing team of fourteen, mapped its last ten campaigns from request to live. Median cycle time was 26 working days. Six campaigns had been rebuilt after review, and in five the reason was an offer or a message that changed after build started. Requests arrived through four routes. The team introduced a single intake form with seven required fields, a rule that build could not start without a brief signed by both requester and owner, and parallel review with a two-day turnaround agreed with the reviewers. A fast lane covered social posts and nurture emails from approved templates. Over the next two quarters, rebuilds after review fell to one in ten campaigns and median cycle time to 17 working days.
Common mistakes
- Designing the workflow around the tool. The stages should come from the decisions the team needs to make; the work management system is configured to match, not the other way round.
- Intake forms that ask for everything. Each required field should prevent a known failure. Long forms drive requesters back to email and the workflow loses its front door.
- Starting build before the brief is signed. It feels faster and is the main source of rework at review.
- Sequential reviews with no consolidation. Each reviewer reopens the previous round. Review in parallel and give one person the job of resolving conflicting feedback.
- The builder running the launch check. A second person catches what the builder cannot see.
- No fast lane. Forcing every small post through the full route teaches people to bypass the workflow altogether.
- Ending the workflow at launch. Without a handover, nobody owns the live campaign, and problems are found at month end rather than on day two.
- Never changing the workflow. Review cycle time and rework by stage regularly, and remove gates that never return work as readily as you add new ones. The process hub explains how to measure both.
