Acontent pillaris one of the few topics a brand reports on consistently, so that over time it becomes recognized for them. We use three. Together they are the identity layer: what people remember about you.
A content format is the repeatable way a pillar gets presented to an audience. Same topic, a fixed shape. The format is the testable unit, and the closest thing to it in media buying is an ad set.
Most brands pick pillars and stop there. That is why their content still does not teach them anything.
Pillars: the identity layer
Three is not an arbitrary number. It is few enough that each one repeats often enough to register, and wide enough that you are not saying the same thing forever. Pillars are deliberately slow: you hold them for a long time, because recognition is built by repetition and destroyed by rotation.
But the part almost every guide misses is that three pillars are not three topics you happen to like. Each pillar should do a different job.
That is the difference between a content plan and a strategy. If all three pillars do the same job, you are loud about one thing and silent about the rest of what a buyer needs to believe before they act.
Give every pillar a role
Here is our own, so you can see the shape rather than the theory. Milkman's three pillars, each with its role:
- Always-On Service Explained. How we build Always-On systems for brands: the workflows, the production pipelines, the mechanics. Role: clarity. It answers "what do you actually do."
- Cinematic Production for Always-On Brands. Premium campaign content built for long-term brand visibility and distribution. Role: proof of quality. It answers "can you actually make something good."
- Execution Proof. Results, and what a brand should expect from working with us. Role: trust. It answers "do you deliver, or do you just talk."
Read those three roles in order and you have a sequence: understand, believe the craft, believe the outcome. A buyer needs all three. Drop the first and nobody knows what you sell. Drop the second and you sound like a consultancy. Drop the third and you are a portfolio with no evidence.
So when you pick your three, do not ask "what should we post about." Ask what three things someone has to believe before they work with you, and make each pillar own one.
Formats: the testable layer
A pillar is too broad to measure. Inside one topic you can go twenty directions, in five tones, across four formats, and when one post outperforms another you have no idea which variable did it. That is why comparing posts inside a pillar teaches you almost nothing.
The format is where measurement becomes possible. A format fixes the shape: the premise, the structure of each episode, the title pattern, the hook. Same topic, same shape, repeated. Now the audience's reaction actually means something, because you have held everything else still.
Think of it as an ad set. You do not test a campaign objective, you test ad sets inside it. You keep one variable moving, you read the response, and you learn something you can act on. A content format works the same way: it is the unit you present to the world, the unit the audience reacts to, and therefore the unit you can compare.
Which leads to the rule that makes or breaks this: change one thing at a time. A format where the premise, the length and the hook all shift between episodes is not a format. It is a series of one-offs wearing a shared name, and it will never tell you anything.
Pillars make you recognizable, formats make you measurable
That is the whole model in one line, and it is why you need both layers.
Pillars give you clarity, and clarity is what earns recognition. Formats give you comparability, and comparability is what earns knowledge. A brand with pillars but no formats is consistent and blind. A brand with formats but no pillars is measurable and forgettable.
Formats are how you decide where the budget goes
This is the part that turns a creative exercise into a commercial one.
Once formats are the unit, you can do two things you could not do before. You can improve inside a format, because you can see which hook or opening earns the watch time. And you can compare formats against each other, because each one is a clean read of one way of presenting a topic.
Then the decision gets easy. The formats that earn attention are the ones that deserve media budget. You are not guessing which creative to scale, and you are not spreading spend evenly out of fairness. Organic finds which format works, and paid puts weight behind it. That is what "creative is the main lever" means in practice: the lever has to be a comparable unit before you can pull it.
Where this goes wrong
Three pillars that all do the same job. Usually all three are proof of quality, because that is the fun one to make. Nobody explains what the company does.
Formats that are really just topics. "Behind the scenes" is not a format until it has a premise, a structure and a rule. Without those, every episode is improvised and nothing is comparable.
Too many variables per format. The most common failure. If you cannot say what stayed the same between two episodes, you did not run a test.
Rotating pillars. Changing them resets the recognition you were building. Pillars are meant to feel repetitive to you long before they register with your audience.
Killing a format too early. One weak episode is not a signal. A format needs enough repetitions to be read, exactly like an ad set needs enough delivery.
How this fits the Always-On System
The two layers sit between strategy and production. Strategy decides what the brand stands for. Pillars turn that into three things you will be known for. Formats turn each pillar into repeatable, testable shapes. The engine produces the episodes, distribution puts them out, and measurement reads which formats earned attention.
Then the loop closes, and it closes on the format, not the post. What worked tells you which format to make more of and which to fund. That is how volume stops being noise and starts compounding into recognition, and recognition into trust.
We set pillars and formats with the brands we run Always-On for, then build the production rhythm on top of them. See how in our media production capability.
