Content Pillars: Pick 4, and Know When One Has Died

Four pillars, four different jobs, four different signals that one has stopped working.

Most pillar sets fail the same way. Someone runs a workshop, sticky notes get sorted into six columns named "Educate", "Inspire" and "Engage", and four months later nobody can say which column works. Those were never pillars. They were adjectives.

This is the pillar chapter of our social media marketing guide. Cadence and slots belong to our content calendar guide; platform and objective decisions to our social media strategy guide.

Content pillars are not pillar pages (and the confusion costs you)

A pillar page is an SEO structure. HubSpot, which popularised the term, defines it as "the basis on which a topic cluster is built", one page that "covers all aspects of the topic" with cluster posts linking back (HubSpot, updated 2 May 2025). One URL, on your site, built for search.

A content pillar in social media is a theme you post within, repeatedly, across many separate posts on someone else's platform. No URL, no cluster, no internal links.

Blurring them costs you twice. Teams inherit the SEO habit of covering every aspect of a topic and build pillars broad enough to absorb any post ever made, which makes them useless for deciding what not to post. Keyword volume is also the wrong input: a fat-volume theme is often something people read and never send on.

What a pillar actually is, and the test for a good one

A content pillar is a recurring theme defined by the audience problem it answers, not by its format or its mood. "Reels" is a format. "Inspire" is a mood. "How to tell whether a ₹900 dinner plate beats a ₹400 one" is a pillar. Run every candidate through three questions and keep only what passes all three.

1. Name the audience problem in one sentence, without naming your product. If the sentence collapses without the product, you have a promotion schedule.
2. List twenty specific post ideas inside it, today, in fifteen minutes. Under twenty and it is a series: run it, let it end, do not build a quarter on it.
3. Check whether a competitor could post it unchanged. If any brand in your category could publish it with their logo on it, it is filler and will perform like filler.

Question three gets skipped most, and it separates a pillar from a template. "5 signs of dehydrated skin" is what eleven other skincare brands posted that month.

How many content pillars? Three or four, and here is the arithmetic

Run three or four, five if one is seasonal. Past five you lose the ability to read any of them, for reasons of arithmetic rather than taste.

Our own calculation, so you can redo it with your numbers. A two-person team ships twelve posts a month. Four pillars means three posts each per month, so a quarter gives nine posts inside every pillar. Nine is enough to take a median seriously.

Seven pillars at the same volume is roughly 1.7 posts each per month, so nine posts in any one takes five to six months. By then the format has changed and the team has turned over: you are comparing January to June and calling it a trend. With seven ideas, merge to four and park the rest.

Choose pillars from evidence, in one working session

Bad pillars come from brainstorms, because your own team produces themes your team finds interesting. Use what the audience already did. Ninety minutes, one document, five inputs.

1. Sort your last 30 comments and DMs into questions. Not sentiment, questions. A repeated question is a pillar candidate with proven demand. Replies and routing sit in our community management guide.
2. Ask whoever answers the phone for their five commonest objections. Objections make the strongest pillars: the reason people did not buy, said out loud.
3. Sort your last 60 posts by saves and by shares, separately. Saves show what people kept; shares show what they were willing to be seen sending. Different pillars win on each.
4. Read three-star reviews across your category, not just yours. Three stars is where people explain what they expected and did not get: a content brief in disguise.
5. Write each candidate as a problem sentence and run the three-question test. Keep the four that survive; park the losers with the date, so the same idea does not return in November as a fresh insight.

One caveat: your current followers describe the audience you have, not the one you want. Chasing a different buyer? Weight the objections and category reviews above your own post history.

Four pillar types that work for Indian SMB and D2C brands

These four keep earning their slot across the categories we work in. The examples use Nilaya Ceramics, a hypothetical Pune D2C tableware brand (invented for this article, not a client) selling handmade dinner sets at a price people struggle to justify.

• Proof in use. The product doing its job in a recognisably Indian setting: a dinner set surviving a fourteen-person Sunday lunch, a stacking test in a Mumbai kitchen with one shelf.
• Buying education. How to judge quality, including when the cheaper option is fine. Why a thin glaze chips at the rim, what "microwave safe" actually claims, which pieces you will never use.
• The operator's view. A category argument someone could disagree with. "We stopped making mugs under 300ml, and why" beats "we love our craft".
• Making, and the people making it. The workshop, the kiln, the rejects, the seconds sale, the reason an order is late.

Notice what is missing: no "engagement", "trending audio" or "motivation" pillar. A trend is a tool you use inside a pillar, since no audio answers an audience problem. If the four feel thin, the gap is upstream, in brand strategy.

One row per pillar. Nobody fills in the last column, which is why dead pillars keep getting posted.

If a pillar has no metric it should obviously win on, it is not distinct from the one beside it. Merge them.

Mapping pillars to formats and platforms

The pillar is the theme; the format is how it survives a feed. Check the metric names before building a report on them, because they move.

Instagram. Meta introduced views on 21 January 2025 and deprecated impressions and plays across all API versions by 21 April 2025; reels_skip_rate and reposts arrived on 3 December 2025 (Meta, Instagram Platform changelog, checked 24 September 2026). Skip rate is the percentage of views where someone skipped within three seconds (IG Media insights reference, same date).

LinkedIn. Engagement is a defined ratio: organic clicks, likes, comments and shares over impressions, organic only, on a rolling twelve-month window (LinkedIn share statistics docs, last updated 30 April 2026). Want a year-on-year pillar read? Export before the data rolls off.

YouTube Shorts. The Content tab reports "Shown in feed" and "How many chose to view", the percentage of times viewers watched your Short versus swiped away (YouTube Help, checked 24 September 2026). The cleanest hook test any platform hands you, and the right way to judge a making-and-people pillar.

Pillars skew by platform: the operator's view carries a B2B LinkedIn page and underperforms on Instagram, where proof in use does the work. Keep the same pillar names everywhere and let the mix change, because four pillars on Instagram and a different four on LinkedIn is eight pillars in disguise. Someone has to own that mapping, a job scoped in our social media manager guide and work we take on in social media marketing.

The pillar every Indian brand adds that nobody engages with (our view)

Opinion, not data. Almost every Indian brand we see adds a festivals-and-national-days pillar. Diwali, Holi, Independence Day, Teachers' Day, Engineers' Day, and by March somebody is designing a World Water Day creative at 11pm.

Our view: it is not a pillar. It is a calendar obligation dressed as a theme, it answers no audience problem, and the output is interchangeable with every other brand posting the same greeting the same morning. It fails the competitor test outright.

What we would do instead: keep the two or three occasions where your category genuinely does something (a sweets brand at Diwali, a gifting brand at Rakhi) and run those as campaigns inside an existing pillar. Delete the rest. If leadership insists, pull last year's festival numbers first.

How pillars feed a calendar without becoming a straitjacket

Pillars decide what a slot is about; the calendar decides when it runs. Keeping those jobs apart stops a pillar system becoming a filing cabinet nobody opens.

• Assign most slots, not all. Leave two or three a month open for whatever happens: a complaint worth answering in public, a category story, a product problem.
• Let pillars borrow slots. Strong idea in one pillar, nothing in another this week? Swap. A month running 5/4/2/1 is fine. A quarter running 30/2/1/0 means you have one pillar and three hobbies.
• Tag every post with its pillar when you schedule it. A column in the sheet, a tag in the tool, anything retrievable. Without it you cannot measure per pillar later, and retrofitting 200 posts is a job nobody does.

On ratios: the 80/20 split and the rule of thirds get repeated everywhere, and we could not find a primary study or a named originator for either this week. They are conventions somebody wrote down once, not findings, and your own pillar numbers over a quarter beat both. Cadence and approval flow: our content calendar guide.

When to retire a pillar, and how to test the replacement

Retire a pillar when its rolling median on its own metric has fallen across two consecutive quarters and you have no new angle left. Two quarters, not one bad post, and the median rather than the average so one outlier cannot keep it alive.

Three other triggers: the pillar repeats itself, the audience problem got solved (a delivery-complaints pillar is pointless once delivery is fixed), or the business moved and the pillar now speaks to the wrong customer.

Test the replacement first. Instagram's trial reels exist for this: the reel goes to non-followers, your followers do not see it in Feed or the Reels tab, you get views, likes, comments and shares after 24 hours, and it can auto-share to everyone on the first 72 hours of views (Instagram for Creators, 10 December 2024). Six trial reels answer the question faster than a committed quarter.

Measure per pillar, not per post

Post-level reporting is where pillar systems die: one post's number is mostly noise, the day, the audio, the thumbnail. Do this monthly instead, on eight or more posts per pillar.

• Use the median, not the average. One post that did 40x hides four that did nothing.
• Judge each pillar on its own metric from the table. Comparing a buying-education pillar's saves against a proof pillar's shares is not one race, it is two.
• Watch direction, and log what changed. Six months of a pillar's median is a signal; one month is a mood. A format switch, a new editor or a platform change like the views migration above all move numbers for reasons unrelated to the pillar.

A pillar winning on its metric while selling nothing is not automatically a kill. Proof and making pillars move saves and shares; the operator's view does the convincing quietly. Judge the set on business outcomes, each pillar on its own metric.