The Three Filter Test: Choosing Content Pillars That Earn Their Keep

Every content team has a graveyard. It is usually a spreadsheet tab or a forgotten Notion page, and it holds the pillars that seemed obvious in a kickoff meeting: thought leadership, industry trends, company culture, tips and tricks. Eighteen months later, those pillars have produced forty articles, three of which anyone reads, none of which sales has ever sent to a prospect. The pillars were never wrong exactly. They were just never tested.

The test this article proposes is deliberately unforgiving. A content pillar has to pass three filters, in order, and failing any one of them is disqualifying. The order matters because each filter is cheaper to check than the one after it, so weak candidates get eliminated before you spend real effort evaluating them.

Before running anything through the filters, it helps to be precise about the thing being filtered. A pillar is not a category label or a blog tag. It is a standing commitment: a territory you promise to keep covering, deepening, and interlinking for years. If that definition feels heavier than the one you have been using, this breakdown of what a content pillar is and everything it quietly stands for is worth reading first, because the test below only makes sense once you treat pillars as commitments rather than labels.

FILTER 01

Proof of Demand

Is there evidence, not intuition, that a defined audience keeps returning to this topic?

The first filter asks for receipts. Not "our buyers probably care about this," but observable, recurring demand you can point to. Demand shows up in four places: search behavior, community conversation, sales conversations, and support tickets. A topic that appears in at least two of those four is demonstrating pull. A topic that appears in none of them is a topic you find interesting, which is a different thing.

Search volume alone is the most common way teams fake this filter. Volume tells you a phrase gets typed; it does not tell you the people typing it resemble your audience. The corrective is to cross-check against where your audience actually spends attention. Public data helps here: the Pew Research Center's ongoing social media fact sheet is a useful reality check on which platforms different demographics actually use, which often contradicts where teams assume their audience lives.

PASS SIGNAL: THE QUESTION KEEPS ARRIVING WITHOUT BEING INVITED

FAIL SIGNAL: YOU HAVE TO EXPLAIN WHY THE TOPIC MATTERS

A practical threshold: if you cannot list five distinct questions your audience has asked about the topic in the last quarter, from real channels rather than keyword tools, the candidate stops here. Roughly half of most first-draft pillar lists die at this filter, and that is the filter doing its job.

FILTER 02

Right to Win

If you publish on this topic for two years, is there a reason your coverage becomes the reference, rather than one more voice?

Demand is necessary but nowhere near sufficient, because high-demand topics attract everyone. The second filter asks what you bring that the existing coverage does not. There are only a few honest answers: proprietary data, direct practitioner experience, access to a specific community, a genuinely contrarian and defensible position, or the resources to go meaningfully deeper than anyone else has bothered to.

"We can write it well" is not on that list. Writing quality is table stakes, and search engines have been explicit that the bar is experience and expertise a reader could not get elsewhere. Google's own guidance on creating helpful, people-first content frames the question almost exactly like this filter does: would someone leaving your page feel they learned enough from a source with real first-hand knowledge, or would they need to search again?

THE UNCOMFORTABLE VERSION OF THIS FILTER

For each candidate pillar, name the current best resource on the internet for that topic. If you cannot name one, you have not researched enough to claim the pillar. If you can name one, write a single sentence explaining what your coverage will have that it lacks. If that sentence is generic, the candidate fails.

PASS SIGNAL: YOU OWN AN INPUT COMPETITORS CANNOT COPY

FAIL SIGNAL: YOUR ADVANTAGE IS EFFORT, TONE, OR ENTHUSIASM

FILTER 03

Distance to Revenue

When a reader finishes a piece under this pillar, is your product or service a natural next thought, or a non sequitur?

The last filter is the one teams resist, because it sounds mercenary. It is not asking every article to sell. It is asking whether the pillar as a territory sits close enough to the problem you solve that the commercial connection never has to be forced. Distance to revenue is measured in the number of logical steps between the topic and your offer. One step is ideal. Two steps is workable. Three or more steps means every piece under the pillar will end with an awkward pivot, and readers feel awkward pivots instantly.

There is a reader-behavior reason to take this seriously, not just a commercial one. People do not read web content the way teams imagine; the Nielsen Norman Group's long-running research on how users read on the web shows readers scan, extract, and leave. A pillar that requires a paragraph of throat-clearing to connect back to your world loses the scanner long before the connection lands. Pillars that sit one step from revenue can make the connection in a sentence, which is all the attention you get.

PASS SIGNAL: SALES WOULD SEND THESE PIECES UNPROMPTED

FAIL SIGNAL: THE CTA ALWAYS FEELS BOLTED ON

The Three Filters, Side by Side

Each filter interrogates a different failure mode, uses different evidence, and costs a different amount to run. Seeing them in one view makes clear why the order is fixed: you check the cheap, external evidence first and save the expensive strategic judgment for candidates that deserve it.

COMPARISON A: WHAT EACH FILTER ACTUALLY CHECKS

Earning Its Keep vs Occupying Space

The phrase "earn their keep" implies a standard of behavior after the choice is made, not just at the moment of choosing. A pillar that passed all three filters behaves differently in the wild than one that was chosen by vibe. The contrast below is what you should expect to observe six to twelve months in, and it doubles as an audit rubric for pillars you already have.

COMPARISON B: A WORKING PILLAR VS A SQUATTING PILLAR

A Worked Shortlist

Abstract frameworks are easy to nod along to, so here is the test applied. Imagine a company selling payroll software to restaurants, running five candidate pillars through all three filters. Notice that a candidate can be strong on two filters and still fail, and that the verdict column allows a middle state: hold, meaning the topic is parked as a subtopic under a stronger pillar rather than promoted to a pillar itself.

COMPARISON C: FIVE CANDIDATES THROUGH THE FUNNEL

The end state is two pillars and one parked subtopic, from a list of five. That ratio feels alarming to teams accustomed to five or six pillars, but it reflects an honest budget: a small team can genuinely deepen two territories. It cannot deepen six, and pretending otherwise produces six shallow ones.

What to Do With the Failures

Cutting a candidate does not mean the topic is banned. It means the topic is demoted from commitment to opportunity. The tipping culture piece can still exist as a one-off if a news moment makes it timely. The trends topic can survive as an annual recap inside a stronger pillar. What changes is the promise: nobody is expected to feed those topics monthly, nobody reports on their performance as a strategic line, and nobody defends them in planning meetings out of sunk cost.

The "hold" verdict deserves its own habit. Held topics go on a calendar for re-testing in six months, because filters are snapshots. Demand grows, competitors abandon territories, and products move closer to topics that were once three steps away. A pillar list should be a living document with a review date, not a founding constitution.

THE TEST IN ONE SITTING

[ 01 ]  List every candidate pillar without judging. Ten to fifteen is normal for a first draft.

[ 02 ]  Run Filter 01 on all of them in a single session. Demand evidence is fast to check and kills quickly.

[ 03 ]  For survivors, name the current best resource per topic and write your one-sentence advantage. No sentence, no pillar.

[ 04 ]  Bring sales into the room for Filter 03 and count the steps to revenue out loud. Two is the ceiling.

[ 05 ]  Record the verdicts, the evidence, and a six-month review date. The document is the deliverable.

The Point of Being This Strict

The three filter test will leave you with fewer pillars than you wanted, and that is its gift. A pillar that survives demand, right to win, and revenue distance is a territory where every hour of writing compounds: readers keep arriving, your coverage keeps deepening past what competitors will match, and the path from reader to customer stays short enough to walk in a sentence. Everything else was never a pillar. It was a topic wearing a pillar's badge, quietly collecting a salary it never earned.