Nobody abandons a content strategy in a meeting. They abandon it at 4pm on a Tuesday, staring at an empty slot for Wednesday, with a trending audio open in one tab and a competitor post open in the other.
That moment is the whole game. Everything upstream of it, the positioning deck, the audience research, the tone of voice document, exists to make that decision easier. Most of it does not, because none of it tells you what to say no to. A brand pillar tells you who you are. A content pillar tells you what you are not allowed to post.
The reason so many teams end up with a feed that looks like a shared drive is not laziness. It is that the empty slot has to be filled and no rule exists to fill it. So the slot gets filled by proximity: whatever is nearest to hand, loudest in the group chat, or easiest to shoot that afternoon. Over a quarter, proximity produces a body of work with no discernible position.
Pillars fix this by moving the decision earlier and making it smaller. You are no longer choosing from everything. You are choosing from four things. The cost of that is real and worth stating plainly: you will reject ideas that would have performed. The return is that what remains accumulates into something recognisable.
What actually counts as a pillar
Before choosing, it helps to know which object you are choosing. The phrase covers three separate things: a recurring theme your brand commits to, a long-form hub page at the centre of an SEO cluster, and a large production asset you cut into many derivatives. They are not competing definitions and they are not the same job. Our complete guide to content pillars disambiguates all three and shows how they stack. This article stays inside the first one and goes deeper on the part that guide could only summarise: the selection decision itself.
A theme pillar, defined tightly, is a subject your brand commits to returning to, repeatedly, for at least two quarters, such that a stranger scrolling twelve of your posts could name it without being told.
Three things are load bearing in that sentence. Repeatedly, because a subject visited once a month is a topic, not a pillar. Two quarters, because association forms through duration and nothing you rotate seasonally will build it. A stranger could name it, because if the pillar is only legible from inside your organisation, it is an internal filing label wearing a strategy costume.
The test that separates the two. Take your last thirty posts and sort them into your proposed pillars. If every post fits somewhere, your pillars are categories, not constraints. Categories are designed to accommodate what you already made. A constraint is designed to have refused something. If nothing in the last thirty days would have been blocked, the pillars are not doing any work.
Why the answer is three to five, and how volume decides which
Three to five appears in almost every practitioner source, usually without a reason attached. The reason is arithmetic, and it is worth doing rather than accepting, because it also tells you whether your number should be three or five.
Association is built by repetition inside a bounded space. Both human memory and recommendation systems need repeated, consistent exposure before they will confidently categorise you. The practical threshold that most teams find workable is roughly four pieces per pillar per month. Below two, the pillar is invisible. Between two and four, it is present but weak. At four or more, it starts to compound.
That single number converts your posting volume into a pillar count. Do not choose the count you like. Choose the count your capacity can actually feed.
| Posts per week | Monthly volume | At 3 pillars | At 5 pillars | Verdict |
|---|---|---|---|---|
| 3 | 12 | 4 each | 2.4 each | Run three. Five would leave every pillar below the threshold and none of them would register. |
| 5 | 20 | 6.7 each | 4 each | Three or four. Five is technically viable but leaves no slack for anything unplanned. |
| 7 | 28 | 9.3 each | 5.6 each | Four is the sweet spot. Enough range to avoid monotony, enough density to build recognition. |
| 12 | 48 | 16 each | 9.6 each | Five. At this volume three pillars will exhaust their supply and start repeating angles within weeks. |
| 2 | 8 | 2.7 each | 1.6 each | Run two. The standard advice does not apply at this volume. Capacity, not strategy, is your binding constraint. |
Read the bottom row carefully, because it contradicts the rule and it is the row most solo operators and small teams are actually in. If you publish twice a week, three pillars means each one surfaces less than once a week. Two pillars executed with real density will build a position faster than three executed thinly. The number three is a floor for teams with volume, not a law for everyone.
The failure at the other end is more common and more damaging. Six pillars becomes eight becomes an unspoken position of covering everything. Adding a pillar without removing one does not add range. It taxes every pillar you already had.
Five gates a candidate has to pass
Guidance on choosing pillars usually stops at picking subjects your audience cares about, which is true and useless. Here is a harder filter. A candidate has to clear all five gates. Four out of five is a fail, because the gate it misses tells you exactly how it will collapse.
| Gate | The question | Pass bar | How it fails in practice |
|---|---|---|---|
| Recurrence | Does this subject generate new material every single week, indefinitely? | You can list 20 distinct angles right now without straining. | Runs dry by week six. The team starts rewording earlier posts and the pillar quietly dies of repetition. |
| Supply | Who inside the business actually produces this, and do they have the hours? | A named person, with recurring time already blocked. | Depends on one busy founder or a subject expert with no diary space. Publishes twice, then stops. |
| Distinctness | Could a reasonable person confuse this with another pillar on the list? | You can name three subjects that belong here and nowhere else. | Two pillars start absorbing each other's posts. Within a quarter you effectively have one. |
| Proximity | Trace the path from someone enjoying this post to someone buying. How many steps? | Two steps or fewer, and you can state them out loud. | Builds an audience with no route to revenue. Engagement rises, pipeline does not, budget gets cut. |
| Endurance | Will this still be worth posting about in eighteen months? | Tied to a durable problem rather than a moment. | Anchored to a trend or a launch. Becomes dead weight the moment the moment passes. |
Two gates deserve extra attention because they are the ones teams talk themselves past.
Supply is the gate most often faked. A pillar without a named owner and blocked time is a wish. The most common version is a behind-the-scenes or founder POV pillar that depends entirely on a founder who has no intention of filming anything. It looks excellent on the strategy slide and produces four posts a year.
Proximity is the gate creators fail. A pillar can be genuinely popular and still be a hobby with analytics attached. If the path from post to purchase requires three improbable leaps, the pillar is buying you reach you cannot convert, and it will be the first thing questioned when someone reviews the budget.
Extract pillars, do not invent them
Pillars invented in a workshop are guesses dressed as decisions. Pillars extracted from evidence you already hold are considerably more likely to survive. The session below takes about ninety minutes and works with a group of two to five people.
Move one: harvest the evidence
Before anyone proposes a theme, put four sets of real material on the table. Your last ninety days of posts with their engagement figures. The last fifty questions sales or support actually received, in the words the customer used. Your top twenty landing pages by organic entry. And a scan of what three competitors publish weekly, noted as subjects rather than opinions about quality.
This step exists to stop the session becoming a preference contest. Anyone can advocate for a theme. Fewer people will advocate for a theme when the evidence in front of them shows nobody has asked about it in three months.
Move two: cluster by question, not by topic
Group the harvested material by the question the audience was asking, not by the subject matter. This distinction does most of the work. A topic cluster gives you groups like pricing, integrations and onboarding, which mirror your product structure. A question cluster gives you groups like how do I know this will work for a team my size, which mirrors how buying actually feels.
Product-shaped pillars are the single most common reason a pillar set fails to connect. They are derived from the org chart, so each department gets representation and the audience gets a map of your internal politics.
Move three: name the clusters and write the boundary
Every cluster that survives gets three lines written down and kept somewhere visible.
- The name. Two to four words. If it needs a sentence, the scope is wrong.
- In scope. Five subjects that unarguably belong.
- Out of scope. Three adjacent subjects that do not belong, even though someone will argue they do. This line prevents drift better than anything else you can write.
The out-of-scope line is the part teams skip and the part that pays. Pillars widen by accident, one reasonable exception at a time, until an email marketing pillar has quietly absorbed SMS, push and in-app messaging and is no longer a constraint on anything.
Move four: cut to the capacity number
You will finish move three with six or seven clusters. Take the arithmetic from the section above, find your capacity number, and cut to it in the room. Do not defer this. A pillar set that leaves the session with an unresolved sixth candidate will have seven by month two.
Cutting is easier with an explicit rule. Rank the survivors by the two gates that predict outcomes best, Supply and Proximity, and cut from the bottom. A topic that loses the cut is not discarded. It becomes a recurring sub-angle inside the pillar nearest to it, which usually strengthens both.
Treat the first four weeks as a pilot. Publish two or three pieces under each candidate before you commit the set for a quarter. Pillars chosen in a session are informed guesses. Pillars confirmed by four weeks of your own performance data are decisions. If a candidate underperforms across every format you tried, that is a supply or distinctness problem showing up early, which is the cheapest time to find it.
Names that hold a line
A pillar name is a working instrument, not a slogan. It gets read at 4pm on a Tuesday by someone deciding whether an idea is allowed. Three rules make names hold up under that pressure.
Name the subject, not the format. Carousels is not a pillar, it is a shape. If you name formats, you will produce carousels about nothing in particular.
Avoid the abstract nouns. Education, inspiration, community and value are the four most popular pillar names and the four least useful, because every post can be argued into all of them. A name that cannot reject anything is not a constraint.
Prefer the specific over the impressive. Restaurant labour costs is a better pillar name than operational excellence. The first has an obvious boundary and an obvious next post. The second has neither, and it will drift within a month.
Pillars are not equal, and pretending they are will cost you
Choosing the set is the visible half of the work. Assigning weight is the half that decides how the feed actually reads. Four pillars at 25 percent each is the default nobody chose, and it produces a feed with no centre of gravity.
A more useful shape gives each pillar a role. One anchor carries the largest share and is what you become known for. Two supports provide range and answer different stages of the buying question. One spike takes the smallest share, carries the highest variance, and is where you take risks.
| Role | Example pillar | Share | What this role is for |
|---|---|---|---|
| Anchor | Workforce scheduling | 40% | The association you are buying. Highest volume, most consistent format, the subject you want named back to you unprompted. |
| Support | Customer results | 25% | Answers the question the anchor cannot: does this work for someone like me. Carries most of the conversion weight. |
| Support | Compliance changes | 20% | Reaches a segment the anchor misses and gives the calendar a reason to be timely without chasing unrelated trends. |
| Spike | Shift-work culture | 15% | Reach and personality. Highest variance in both directions, which is precisely why it gets the smallest allocation. |
The example mix above is drawn for a B2B software brand publishing seven times a week. The roles transfer across business models even when the subjects do not.
On the famous ratios. You will meet the 80/20 rule, the 5:3:2 rule and the 50/30/20 split. None of them traces back to a published study with a stated method. They are practitioner heuristics that hardened into rules through repetition, and they are a reasonable default when you have no data of your own. The moment you have four weeks of your own numbers, that data outranks every ratio in circulation.
The layer between the pillar and the post
A pillar tells you what a post is about. It does not tell you what the post is. The gap between those two is where most calendars stall, because the person filling the slot still has to invent a shape.
Close it by fixing a small set of repeatable formats per pillar. Three or four each is enough. The point is not creative range, it is that the person planning Wednesday is choosing from a list of twelve known combinations rather than from an infinite field.
| Pillar | Short video | Static or carousel | Long-form social |
|---|---|---|---|
| Anchor | Sixty second teardown of one common mistake. | Framework broken into five numbered panels. | The full argument with the reasoning shown. |
| Results | Customer says the outcome in their own words. | Before and after with the actual figures. | The full story including what nearly went wrong. |
| Timely | What changed, in plain language, under a minute. | A dated summary card people save and return to. | Second-order effects nobody else is writing about. |
| Spike | An unpolished moment from the working week. | A recognisable in-joke from inside the industry. | A short, honest position on something contested. |
One warning about platforms. Pillars should stay constant across channels while formats change. The same anchor pillar becomes a sixty second video on one platform and an argued post on another. Teams who let the platform choose the pillar end up with a different brand on every channel, which reads to anyone following you in two places as inconsistency rather than range.
What a constrained month looks like
Twenty publishing days, four pillars, weighted 40 / 25 / 20 / 15. Every slot was decided before the month began. Nothing here required a decision on the day.
| Week | Monday | Tuesday | Wednesday | Thursday | Friday |
|---|---|---|---|---|---|
| Week 1 | Anchor: teardown video | Results: customer clip | Anchor: framework carousel | Timely: rule change explainer | Spike: week note |
| Week 2 | Anchor: common mistake video | Results: numbers carousel | Timely: saved summary card | Anchor: long-form post | Results: full story post |
| Week 3 | Anchor: teardown video | Timely: second-order effects post | Anchor: framework carousel | Results: customer clip | Spike: industry in-joke |
| Week 4 | Anchor: long-form post | Results: before and after | Timely: rule change explainer | Anchor: reader question | Spike: short position post |
Three things are worth reading off this grid.
The anchor never goes more than two days without appearing. That spacing is deliberate. Density on one subject is what produces recognition, and recognition is the only thing that makes the other three pillars land.
No two consecutive days share a pillar. Weighting decides how often a pillar appears. Sequencing decides whether the feed feels repetitive. They are separate problems and the second one is solved by spacing, not by adding pillars.
The timely pillar carries the flexibility. When something genuinely newsworthy happens, it takes a timely slot. It does not get bolted on as an extra post, which is how a twenty slot month becomes a twenty six slot month that nobody planned and nobody has capacity for.
Keep a log of what you rejected
This is the practice almost nobody runs and the one that tells you most. Keep a shared note. Every time an idea gets blocked because it does not belong to a pillar, write it down with one line on why.
The log does three jobs at once, and none of them are achievable from your published output alone.
- It proves the constraint is live. An empty log after a month means your pillars are categories, not constraints. You are approving everything and sorting it afterwards.
- It reveals the missing pillar. When eleven rejected ideas cluster around the same subject, that is not eleven bad ideas. That is a pillar you failed to identify, arriving through the back door.
- It settles arguments. When someone asks in a review why the brand never posts about a given subject, the log answers with a date and a reason instead of a shrug.
Review the log at the end of each quarter alongside your performance data. It is the only place your strategy's opportunity cost is written down, and it is the input that makes the quarterly re-cut an evidence-based exercise rather than an opinion contest.
Six ways a pillar set fails
Ordered by how often they occur. Each has a symptom you can observe and a check you can run this week.
| Failure | What you observe | The check |
|---|---|---|
| Pillar bloat | The set has grown to six or seven and nothing feels like it is compounding. | Divide monthly output by pillar count. If any pillar receives fewer than two pieces a month, you have more pillars than capacity. |
| Silent overlap | Two pillars keep producing posts that could sit under either one. | Take last month's posts and have two people sort them independently. Disagreement above roughly one in five means the boundary is not real. |
| The absent owner | One pillar is always the one that slips when the week gets busy. | Name the owner and find the blocked time in their calendar. If either is missing, that is the cause, and no amount of planning fixes it. |
| Scope drift | The pillar covers noticeably more ground than it did two quarters ago. | Reread the out-of-scope line. If posts have crossed it, either rename the pillar honestly or split it into two and cut something else. |
| Themes with no home | Consistent social presence, no organic search footprint, all reach rented from platforms. | For each pillar, name the page on your own site it maps to. If you cannot name one, you are building an audience you do not own. |
| Weighting inversion | The promotional or product pillar has crept up to a third of output. | Count actual published pieces per pillar against the planned mix. Product pillars expand quietly because they always have an internal advocate. |
The quarterly re-cut
Every guide covers how to create pillars. Very few cover the other half of the lifecycle, which is where a mature programme spends most of its attention. Ninety minutes a quarter, four questions per pillar.
Did it earn its share?
Compare each pillar's share of output against its share of outcomes. A pillar consuming 30 percent of production for 8 percent of results is a candidate for reduction. Be careful with the anchor, though. It often shows weaker direct numbers precisely because its job is to make the other pillars credible, and cutting it can flatten everything else a quarter later.
Has it drifted?
Reread the in-scope and out-of-scope lines written at the start. Drift is not a crisis, but an undeclared drift means the constraint stopped operating some weeks ago and nobody noticed.
Should two become one?
Merge when two pillars keep producing interchangeable posts, or when neither generates enough volume to sustain itself. Merging is usually an upgrade rather than a retreat: two half-built themes become one complete one, and one complete theme is a materially stronger signal than two partial ones.
Should one be retired?
Retire when audience demand has structurally declined, when the business has moved away from the subject, or when you have concluded you cannot win it. Retirement is not deletion. Fold the best-performing angles into the neighbouring pillar and keep whatever still earns attention.
The rule that keeps a pillar set healthy. Add a pillar only by replacing one. This forces the question content teams avoid most reliably: what are we giving up in order to do this. Teams that add without subtracting arrive at nine pillars over two years and cannot work out why nothing compounds any more.
The verdict
Find your row. Each one names the single move that matters more than everything else in this article.
| If this is you | Verdict | The one move that matters |
|---|---|---|
| You publish twice a week or less | Run two pillars, not three. | Density beats range below twelve posts a month. Add the third pillar when volume genuinely supports it, not when the strategy document looks thin. |
| You publish daily or close to it | Run four, weighted. | Give one anchor around 40 percent of output and never let it go more than two publishing days without appearing. Recognition comes from that density. |
| Your set has grown to six or more | Cut before you plan anything. | Merge the two pillars that overlap most, then adopt the rule that a pillar can only be added by replacing one. Nothing compounds until you do. |
| Every idea you have fits a pillar | Your pillars are categories. | Rewrite the out-of-scope line for each pillar until it excludes something you would otherwise have posted, then start the rejection log to prove it is working. |
| One pillar keeps slipping every busy week | You have a supply problem, not a strategy problem. | Give it a named owner with recurring time already blocked, or fold it into a neighbouring pillar. No plan survives an absent producer. |
| You are starting from zero | Extract, pilot, then commit. | Run the ninety minute extraction session, publish two or three pieces per candidate for four weeks, keep what the data confirms, and hold the set for a full quarter. |
The short version. Let your posting volume decide the count. Make every pillar pass all five gates. Weight the set around one anchor instead of splitting it evenly. Fix the formats so nobody invents a shape at 4pm. Keep the log of what you turned down, and add a pillar only by replacing one. The calendar stops being a blank page and starts being a rule, which is the entire point.
