A marketing campaign workflow that survives contact with real teams: intake, approval gates, baselines and feedback loops, with hot takes on what to stop doing.
Marketing lead sorting a campaign request
A marketing campaign workflow is the repeatable path a campaign takes from a request to a measured result: who asks, who accepts, who makes what, who approves, who launches, and who reads the numbers. Most teams have a version of it. Few have one that works, because they wrote a checklist and called it a system. Below: the opinions we would defend in a meeting, then a build you can copy.
Alexdly
alexdly.ai
Make Campaign Handoffs Clear
AlexDLY helps teams assign marketing tasks with clear outcomes, checklists, and deadlines while keeping progress and ownership visible.
Table of Contents
- Hot take 1: your workflow is a checklist, and that is the problem
- Hot take 2: the real campaign starts at intake
- Hot take 3: approvals should check the test, not just the typos
- Hot take 4: attribution is a debate you schedule in advance
- Hot take 5: brand and demand are one workflow, not two calendars
- The build: a marketing campaign workflow in seven steps
- Hot take 6: the problem is rarely the tool, it is the handoff
- Our perspective: execution is the missing layer
- Hot take 7: automate the follow-through, not the thinking
- Where AlexDLY fits
- A one-week rollout that does not need a reorg
- Common failure modes
- FAQ
- Sources
Hot take 1: your workflow is a checklist, and that is the problem
Search this topic and you will find a familiar shape: plan, create, review, launch, report. Tidy boxes, arrows pointing right. It looks like a process. It behaves like a to-do list.
The trouble is the arrows. Real campaigns loop. The UK Government Communication Service’s OASIS campaign planning guide names five stages: Objectives, Audience and insight, Strategy and idea, Implementation, and Scoring. It also calls for review and refresh throughout, using feedback to improve implementation while the campaign runs. That is a loop, not a conveyor belt.
Campaign stages linked by a feedback loop
If your workflow only has a retrospective at the end, you are doing archaeology, not management. Put a feedback gate after each phase. Ask one question at each: did what we just learned change what we do next?
Hot take 2: the real campaign starts at intake
People search for “project intake process” and “marketing request intake” because that is where the pain lives. A Slack message that says “can we get a webinar going?” is not a request. It is a mood.
A request earns a spot in the queue only when it carries:
- The outcome. One measurable result, not “awareness.”
- The audience and intended behavior. Who, and what you want them to do differently.
- The baseline. Today’s value for the metric you are trying to move.
- The deadline and the owner. One name, one date.
- The decision rule. What you will do if it works, and what you will do if it does not.
The GCS approach is blunt about audience: insight has to identify the barriers to the behavior you want, and that shapes the proposition, channels and journey. In workflow terms, that means creative tasks should not be accepted until the audience insight and intended behavior are on the request. Writers and designers are not the place to discover that nobody knows who the campaign is for.
The baseline is the item teams skip most. GCS guidance warns that a target with no no-intervention baseline cannot show what the campaign changed. Record the starting value before anyone gets assigned work. Otherwise a good quarter and a good campaign look identical.
A request form that earns its place
Keep it to a page. Required fields: campaign name, requester, business goal, primary metric and its current value, audience, intended behavior, channels, hard deadline, budget ceiling, and who signs off. Optional fields breed ignored fields. If the requester cannot fill a required one, that is the useful finding, and it cost you ten minutes instead of three weeks.
Hot take 3: approvals should check the test, not just the typos
Most approval queues exist to catch a wrong logo or a legal problem. Fine. But the expensive mistake is a campaign that launches fast and teaches you nothing.
The American Marketing Association’s experiment planning workbook makes the point that experiments only help when the test is planned clearly: a focused hypothesis, a measurable success metric and an agreed decision rule. Agree the rule before launch and you cut down the post-test arguments and hindsight bias that eat the week after a campaign ends.
So the approver’s checklist has two halves:
| Check | What the approver confirms |
|---|---|
| Creative | Copy, design, brand and legal compliance |
| Measurement | Hypothesis, one main variable where possible, success metric, decision rule |
| Baseline | Starting value recorded and dated |
| Attribution | Which model answers the question, and what action follows if credit shifts |
That last row deserves its own take.
Hot take 4: attribution is a debate you schedule in advance
Nobody has the true attribution number. Google’s attribution playbook starts from the sensible premise that modeling works best when you define your analysis questions and plan what you will learn. First-interaction, last-interaction and linear models answer different questions, so comparing them is more honest than picking a favorite.
The workflow fix is dull and effective: write down, before launch, what you will do if channel credit moves. If paid social looks stronger under one model and weaker under another, do you shift budget, hold, or run a test? Decide while nobody’s channel is on the line.
Hot take 5: brand and demand are one workflow, not two calendars
Many teams run a brand campaign calendar and a demand campaign calendar, with different owners and different dashboards. The AMA piece on brand and demand marketing quotes Rally’s Ashley LaPorte on connecting the two through shared objectives, integrated planning, test-and-learn work and unified measurement, instead of managing isolated campaigns.
In practice, one intake form, one objective hierarchy, one scorecard. If sales needs leads and marketing needs reach, the shared objective is the revenue outcome both are accountable for. Sales and marketing also need a shared definition of where a lead sits; a short piece on CRM lifecycle stages argues for starting with four or five, which is a good limit for the handoff between campaign and pipeline.
The build: a marketing campaign workflow in seven steps
Here is the version we would put in front of a team on Monday. It follows the OASIS logic but is written as tasks with owners.
Step 1: Intake and triage
One queue. Every request uses the form above. A single person, often a campaign lead or marketing ops, triages weekly: accept, send back, or park. Rejection with a reason is a feature. It teaches requesters what a good request looks like.
Step 2: Objectives and baseline
Turn the request into one measurable outcome. Record the baseline and the date you measured it. If the outcome is a behavior, say which behavior.
Step 3: Audience and insight
This is the stage most likely to be skipped under deadline pressure. Make it a gate: no creative tasks open until the audience, the barrier and the intended behavior are written down. Signals about what your audience is already doing help here, and a guide on matching intent signals to sales moves is a reasonable primer if you do B2B outreach alongside campaigns.
Step 4: Strategy and idea
One proposition, a channel plan, and the experiment design: hypothesis, main variable, success metric, decision rule. Approver signs off on the plan, not only the creative that follows.
Step 5: Production
Now assign the work. Each task needs an owner, a deliverable described so someone can check it, a checklist, and a deadline. “Write the landing page” is a wish. “Landing page copy, 400 words, one CTA, includes the three objections from the sales call notes, due Thursday noon” is a task.
Step 6: Approval and launch
Two approvals at most. Creative and compliance first, measurement second. Every approval should leave a visible record: who approved, when, and against what criteria. Then post to every channel on the schedule, not from five logins.
Step 7: Scoring and refresh
Compare results to the baseline and apply the decision rule you wrote in step 4. OASIS scoring looks at outputs, outtakes and outcomes, which is a good discipline: posts published and emails sent are outputs, and they are not results. Then feed the learning back into intake so the next request starts smarter.
Hot take 6: the problem is rarely the tool, it is the handoff
Teams buy a new project tool every two years and still miss launch dates. The cause is almost never missing features. It is the handoff: a task that was assigned in a meeting, clarified in a thread, edited in a doc, and approved in a DM. Nobody can say who owns it or what “done” meant.
A workable workflow makes four things visible at any moment:
- Who owns each task right now.
- What outcome counts as accepted.
- What is waiting on an approval, and for how long.
- What slipped, and why.
If your system cannot answer those four in under a minute, you have a communication habit, not a workflow. Sales follows the same logic. Pieces on which automations to start with in sales workflows land on a similar point: pick one or two processes, make them reliable, then expand.
Our perspective: execution is the missing layer
We build AlexDLY around one belief: most teams do not lack tools, they lack a clear line between “someone should do this” and “this is done and approved.”
That is why the product is organized around delegation that preserves ownership. You assign a task with an outcome, a checklist, acceptance tests and a deadline, and you keep a live view of progress and who holds each piece. Content goes through an approval queue instead of ad hoc DMs. Sales calls can be reviewed with coaching feedback so a manager is not listening to every recording. And AlexDLY sits above the work OS or CRM you already run, so you manage execution without ripping out the tools your team knows.
Our hot take on the whole topic: a campaign workflow is an accountability design problem first and a software problem second. If you can describe who owns what, what accepted looks like and what happens when a deadline slips, almost any setup can run it. If you cannot, no tool will rescue you. We would rather you fix the design and then pick the layer that enforces it.
- Thomas
Hot take 7: automate the follow-through, not the thinking
The tempting move is to hand an AI model the whole campaign. The better move is to give it the routine, checkable work and keep humans on the decisions.
Good candidates for automation:
- Routing a new request to the right department with the missing fields flagged.
- Drafting campaign concepts and channel variants for a human to approve.
- Posting approved content to several social pages at once on a schedule.
- Chasing overdue tasks and surfacing blocked approvals.
- Summarizing sales calls into coaching points a manager can act on.
Bad candidates: choosing the objective, deciding what the audience believes, and setting the decision rule. Those are the parts the OASIS framework puts first, and for good reason. Automation speeds up whatever you feed it, including a vague brief.
Where AlexDLY fits
If your campaigns stall between the idea and the approved, published result, that is the gap AlexDLY, the AI execution layer, is built for. You delegate tasks with outcomes, checklists and deadlines, route work from ChatGPT or Claude, keep content in an approval queue, and post to your social pages from one agent. Want to see what lost hours are costing you first? Try the ROI calculator. Prefer a wider look? Start at alexdly.ai.
A one-week rollout that does not need a reorg
You do not need a transformation program. Try this:
- Monday: Write the one-page request form. Cut every field you cannot defend.
- Tuesday: Name the triage owner and the two approvers. Publish the rule that nothing enters production without a request.
- Wednesday: Pull your last three campaigns and record, for each, the baseline you wish you had captured. This shows the cost of skipping it.
- Thursday: Write the approver checklist with the creative and measurement halves.
- Friday: Run one live request through the full flow and note every place someone asked “who has this?” Those questions are your next fixes.
Expect the first run to be slower. The point is that the second one is not.
Common failure modes
Too many approvers. Each extra signature adds a day and removes accountability. Two is plenty.
Intake as a suggestion box. If anything can enter the queue, nothing has priority. Say no in writing.
Metrics chosen after the fact. If you pick the winning metric once the results are in, you did not run a test, you told a story.
Reporting outputs as outcomes. Impressions and posts shipped are activity. Compare to the baseline.
No owner for the loop. Someone has to carry what you learned back to intake. Name them.
Sources
- A guide to campaign planning (OASIS framework), UK Government Communication Service
- Marketing Experiment Planning Workbook, American Marketing Association
- Attribution Playbook, Google
- Brand and Demand Marketing: Marketing’s Powerful New Duo, American Marketing Association
Alexdly
Discuss Your Campaign Workflow
Reach out to discuss how AlexDLY supports task delegation, progress tracking, and coordinated marketing work across your team.
AAlexdly
Email AlexDLYtom@alexdly.ai
+1 317 213 2436
Recommended
- The Complete Guide to Agentic Sales Workflows for Solo Founders
- 5 Sales Methodologies That Actually Move Revenue (with Free Training)
- How to Hand Customer Follow-Up to AI Agents Without Losing Trust
- AEO vs GEO vs SEO for AI Search
What are the 4 stages of a marketing campaign?
There is no single standard, so check which model your team uses. A common simple version is plan, create, launch and measure. The five-stage OASIS model from the UK Government Communication Service adds audience insight and treats review as ongoing, which is closer to how campaigns behave in practice.
What are the 5 steps of a workflow?
For a campaign, a practical five are request, plan, produce, approve and launch, and measure. The key is that each step has an owner and a clear definition of done. Add a feedback point after each so learning does not wait for the end.
What should a project intake process include?
At minimum: the outcome, the audience and intended behavior, the baseline metric, a deadline, one owner and an approver. Add a decision rule if the work is an experiment. Requests missing required fields should go back to the requester, not into production.
How is a campaign workflow different from marketing automation?
Marketing automation sends messages triggered by customer behavior, such as nurture emails. A campaign workflow governs the human and tooling steps that create and launch campaigns: requests, tasks, approvals and review. You usually need both, and they solve different problems.
See what an execution layer does with your week
AlexDLY runs on the ChatGPT and Claude keys you already pay for, adds your live business context, and brings you finished work to approve.
Build my free War Room
