The annual plan that was already fiction by March
You spent three weeks on your annual plan in November. By March, half of it no longer held. A channel did something unexpected, a competitor changed course, your agency proposed a different approach, and by then your team had picked up six new tools.
The reflex is understandable: plan faster. Quarterly plans instead of annual ones. More stand-ups. A more agile process.
None of that fixes it. The problem is not the frequency of your planning. It is what your organisation is built around.
Your org chart describes a landscape that no longer exists
Open the org chart of an average marketing team and you are reading an archaeological layer. Digital manager. Content manager. Social media manager. E-commerce lead. Brand manager. Each of those titles is named after a place where work lands, not after a decision someone has to take.
That was defensible as long as channels moved slowly. A channel was a craft back then. It took years to learn, and anyone who had mastered it was scarce.
Today the channel is the fastest-moving part of your entire system. Search has changed shape in eighteen months. Social platforms are filling with synthetic filler and are changing character simply because of it. New interfaces are arriving that none of your current job titles covers.
And you have bolted your organisation to exactly that part.
The consequence is not that you are slow. The consequence is that you are structurally fragile. Every time a channel shifts, you have to reorganise. Your org chart is not a map of your work, it is a list of things that will shortly stop being true.
There is a second consequence, and it is the more expensive one. A channel organisation puts every decision on the wrong clock. Your brand promise gets renegotiated in a weekly content meeting, because that is where the person who has to write the copy happens to sit. And approval on an adaptation waits three weeks for a steering group, because that is what the procedure says. Both are wrong. And both stay invisible as long as you look at job titles instead of decisions.
So the question is not how to re-slice your team. The question is: which decisions does this team actually take, and who takes them?
Five decisions, five clocks
Every marketing team, from three people to three hundred, repeatedly takes five kinds of decision. They differ in tempo, in reversibility, and in what counts as evidence. That is precisely why they do not belong in the same meeting.
| Domain | The question | Clock | Reversible |
|---|---|---|---|
| Promise | What we promise, to whom, and what we explicitly do not | yearly | barely |
| Allocation | Where money and attention go, and what we stop | quarterly | largely |
| Expression | What goes out with our name on it | continuous | entirely |
| Truth | What counts as evidence that something works | twice a year | yes, but rarely done |
| Capacity | Who or what makes this: in-house, agency, freelance, model | continuous | yes |
Lay those five beside a real org chart and two things stand out.
Truth and Capacity have an owner in almost no organisation. Truth is scattered across marketing, finance and the media agency, and there is no mechanism that decides which number wins when two dashboards contradict each other. So the winning number is whoever's voice is loudest, or whichever number carries the better news. Capacity is buried in procurement and in agency contracts, and it is shifting faster than any governance can follow.
And stopping belongs under Allocation, but is assigned to nobody. Every organisation has an owner for starting and none for stopping. It is the cheapest structural fault to repair and the most profitable. You do not have to hire anyone for it.
Four ways it goes wrong
The moment you look decision by decision instead of function by function, four failure patterns become visible. All of them are recognisable around the table.
Orphan decision. Nobody owns it. Tell-tale sign: the topic returns to every agenda and nothing ever changes. The classics are the measurement definition, the stop decision, and lately the question of what a model may publish in your name.
Collision. Three owners, so the decision is permanently reopened. Tell-tale sign: the phrase "didn't we settle this last quarter?" Usually caused by channel owners who each hold a piece of the same decision, each with a budget line underneath it.
Clock error. A slow decision on a fast clock, or the reverse. The brand promise reopened at every campaign. The adaptation waiting three weeks for a signature. This is the most expensive failure mode and the least visible, because nobody ever names it as a single problem.
No evidence. Decided on opinion where data exists, or decided on data that does not measure what the decision asks. Especially acute in Expression, where volume is exploding and approval capacity is not scaling with it.
Three interventions
None of this has to become a new org chart. That is also the best argument for starting: the first intervention costs you nothing but a conversation.
1. Name one owner per decision domain
A name, not a committee. That person owns three things: the rhythm, the evidence and the settled answer.
These are not new roles and not new hires. They are five mandates you lay on top of existing seniority. In a team of five, two people will each carry two mandates. In a team of forty they are five different people, and there the sharpness of the boundary is exactly the point.
The only real work is the political nerve to assign Truth and Capacity explicitly. Those are the two nobody wants today, because both mean you will have to contradict your colleagues.
2. Re-slice execution from channel to craft
Makers organise around a craft: writing, design, data, production, distribution. Not around a platform.
Channel knowledge stays crucial. It simply becomes a competence instead of an identity with its own budget line. Concretely: your social media manager does not become a different person. Their mandate shifts from "I own this channel and its budget" to "I am the distribution specialist on the team". What they lose is a budget line. What they gain is not being reorganised every eighteen months when the platform changes shape.
Expect resistance. Taking away a budget line feels like a demotion, even when it is not. Have that conversation explicitly, not in a footnote of a reorganisation memo.
3. Give every decision its own clock, and forbid the debate outside it
This is the intervention with the fastest effect and the lowest cost.
- Promise opens once a year and stays shut the rest of the year.
- Allocation opens every quarter, including the question of what stops.
- Expression runs continuously, inside a mandate set in advance. Write into that mandate what a model may draft, what it may publish, and what requires a human signature. Without that agreement your approval capacity will not scale with your production capacity, and in most teams that is already the real bottleneck.
- Truth twice a year.
- Capacity continuously, but with an explicit decision rule rather than case by case.
"Forbid" sounds harsh, and it is meant to. If the brand promise comes up in a weekly meeting, the right answer is not a debate but a redirection: this opens in September, note your argument. Without that mechanism you pay every week again for a decision you have already taken.
What you need to make it work
Two things. Neither requires a purchase.
A decision register. One table: what was decided, by whom, on what evidence, when we reopen it, and what would change our mind. That last column is the most valuable and it is always skipped. It is the difference between a decision and an opinion. A spreadsheet is perfectly fine. The bottleneck is never the tool, it is the discipline of keeping it current.
A context layer a machine can read. Your brand guidelines, your audience definitions, your tone of voice and your decision history currently live in a beautifully typeset PDF on a shared drive. That is dead material. Put them in plain text, with a clear structure, in one findable place, under version control. An ugly text file a model can read is worth more today than a gorgeous brand book nobody opens. Every colleague who reaches for a model tomorrow will do so with or without your context. You choose which of the two.
What happens next
If you do this, the question you put to the outside world changes.
Today you ask agencies for services: campaign, content, media, activation. That is a channel question, and it earns you a channel answer.
Once you know which five decisions you own, the question changes. It becomes: which capacity belongs inside, which belongs outside, and who holds the context that is needed for it. Agency choice then follows from organisational design, and not the other way round. That is a noticeably calmer conversation, on both sides of the table.
You cannot predict the future. That is not the assignment either. The assignment is to build an organisation in which it is clear who decides what, when, and on what evidence. Everything you do not see coming after that will at least be somebody's decision.
Check for yourself how your organisation decides
The Decision Card takes seven minutes. You work through twelve concrete decisions, tailored to the size of your team, and mark for each one who takes it, on what rhythm and on what evidence. They are not twelve at random: they are the twelve that most often have no owner, set beside a handful that usually run well. You get one finding at the top with the action that belongs to it, and below that the full card: your orphans, your collisions, your clock errors and your evidence-free decisions. Your card is stored, so you can set it beside a second run in six months.
The result is not advice. It is a mirror you can put on the table on Monday.