Content strategy that compounds: how we plan a year of publishing

Ask a mid-size company how their content strategy is going and you'll usually hear a number. "We're publishing four posts a week." "We did two hundred articles last year." Volume is the one metric everyone can quote, which is suspicious, because volume is the one metric that says almost nothing about whether content is working.

Here's the number we ask for instead: of everything published in the last twelve months, what share of organic sessions now comes from pieces published before those twelve months began? On a healthy programme the answer climbs every quarter — it's routinely 60–80% on the programmes we run as part of a growth retainer. That's compounding: old work appreciating while new work ships. On an unhealthy programme the answer is near zero, which means every article has the shelf life of a sandwich and the whole team is a hamster wheel with a CMS.

This is how we plan a year of publishing so the second outcome doesn't happen. It's the system behind programmes like the journal that grew Meridian Climate's organic pipeline, and it borrows shamelessly from portfolio management, because that's what a content library is.

Start with the maths, not the topics

Every content library we've ever audited follows the same rough distribution: about 10% of URLs drive 80% of organic traffic, and the long tail is either appreciating slowly or quietly rotting. Planning a compounding year means planning for three asset classes at once:

Compounders. Pieces targeting durable, intent-stable queries — the questions your buyers will still be typing in five years. They often start slow (the query has real competition) and appreciate for years. These are the pension fund of the library: unglamorous, reliable, the source of eventual wealth.
Spikes. Timely pieces — a platform deprecation, a regulation change, a new framework release. They earn links and attention in a burst, then decay to near zero. Spikes are worth doing, sparingly, because they buy the authority and links that help compounders rank. They are marketing for the pension fund, not the fund itself.
Decaying assets. Existing pieces losing rankings as the query's intent shifts and competitors refresh. Ignoring these is how libraries rot. A mature programme spends a fixed share of its capacity — we hold 25–30% — on refreshing, consolidating and occasionally deleting.

The single most important strategic decision is the ratio. For most B2B and service businesses we plan roughly 60% compounders, 15% spikes, 25% maintenance. A startup chasing awareness in a category it invented might invert the first two. What never works is 100% new URLs forever — that's the hamster wheel.

Topic selection: intent first, volume second, ego last

The topic-selection meeting in an unhealthy programme sounds like this: "What have competitors written about? What's the CEO been thinking about? What's trending?" All three are recipes for mediocrity — you'll write a worse version of an existing article, a piece nobody searches for, or a spike with no follow-through.

We select topics by walking the buyer's actual questions, tiered by intent:

Problem-aware queries. Someone has the pain but not a name for the solution. "Why does our checkout conversion drop on mobile." These build the compounding core of most programmes, because the underlying pains change slowly.
Solution-comparing queries. "Headless vs monolith commerce." High commercial value, fiercer competition, and your piece must be honestly comparative or it deserves to lose. Our piece on headless commerce trade-offs is unambiguous about when headless is theatre — that honesty is precisely why it converts.
Brand and navigational queries. Small volume, near-certain conversion, often neglected because nobody wants to "do SEO" for their own name. Own them properly anyway.

Search volume is a tiebreaker among intents, not a target. A 200-searches-a-month query asked by exactly your buyer, at exactly the moment they feel the pain, beats a 5,000-volume vanity query you rank 11th for. And ego topics — the CEO's hobby horse — get one honest test: would a stranger with the problem bookmark this? If the answer requires the phrase "well, it's really more about our vision," it goes in the newsletter, not the journal.

This is where content clusters earn their keep: a cluster of twelve interlinked pieces around one problem space lets a site with modest authority outrank giants on specific intents, because relevance compounds within the cluster even before links arrive.

The refresh engine: where compounding actually happens

New articles get all the applause; refreshes do all the quiet work. Our refresh cycle runs like this.

Quarterly triage. Pull the last four quarters of Search Console data per URL. Sort every piece into four buckets: growing (leave alone — do not touch things that are compounding, redesign urges be damned), flat (candidates for consolidation or rewriting), decaying (refresh now), and dead (no impressions for six months — refresh ruthlessly or 301 it into a stronger sibling).

The decaying bucket gets the full treatment. Not a new date stamp — a genuine revision: re-check the query's current intent (it will have drifted), rewrite sections that no longer answer it, update every example and screenshot, strengthen internal links from pieces published since it launched, and tighten the title against what the current page-one results promise. A serious refresh takes a third to a half the effort of a new piece and routinely returns multiples of what the same effort spent on a new URL would yield, because the URL already has links, history and impressions. You're compounding interest on an asset you already own. Skipping refreshes to fund new content is refusing free money to buy lottery tickets.

Consolidation is the unglamorous winner. Most decaying libraries have five thin articles all nibbling at one query. Merging them into one definitive piece — with proper redirects, per the migration rules — consistently outperforms five rewrites. People resist deleting their old work. The redirect preserves the equity; the ego cost is the only real loss.

Distribution: plan the second audience before writing for the first

SEO distribution has a half-life of months; newsletter and social distribution have half-lives of hours. Compounding programmes treat them as one system: a piece is written for search intent, then re-cut for owned channels — never merely linked. The newsletter version is an argument, not a summary; the LinkedIn version is one idea with the jargon scraped off. We've written about running this as an editorial operation that survives contact with reality; the short version is that distribution is scheduled before the article is written, not improvised after it's published.

One underrated mechanic: route every new piece's internal links deliberately. Each new article should link forward to the compounders it supports, and the refresh queue should link back to it. Internal linking is the only distribution channel you wholly own, and most companies treat it as an afterthought their CMS autocompletes.

Attribution: honest maths or no maths

Content's dirtiest secret is attribution. A compounder piece read in January is invisible in a March demo request that closes in June. The programme that only counts last-touch conversions will defund its best assets within two quarters — we've watched it happen, twice, at companies that then came to us to rebuild what they'd killed.

Our position, elaborated in the attribution essay: use directional, modelled attribution and say so in the report. Track first-touch source on self-identified leads, ran through a simple model that credits organic content for assisted pipeline, and report a range rather than a false-precision number. Pair it with the one honest experiment available: when you refresh or prune a section, pre-register what you expect to move, then measure it. The discipline of a well-run CRO programme applies to content too — the learnings ledger matters more than any single number.

A content programme you can't defend with honest maths is a content programme one budget cycle from death. Build the defence while you build the library.

The year, actually planned

Our annual plan fits on two pages. Page one: the portfolio allocation (60/15/25 or whatever the business warrants), the clusters we're building and the intents they serve, the refresh calendar with named owners. Page two: the quarterly distribution plan and the measurement model, written down before the year starts, so December's report is an answer to a question asked in January, not a story invented in December.

Then we publish on a schedule a small team can keep forever — for most clients, two to four pieces a month plus maintenance — because the strategy that compounds is the strategy that survives contact with a hiring freeze, a rebrand and a product pivot. Consistency is the whole trick. The hamster wheel feels fast for a year and goes nowhere. The portfolio feels slow for two quarters and then it starts paying you back.

Key takeaways

Judge a programme by the share of traffic from pieces over a year old — healthy libraries hit 60–80% and climb.
Allocate capacity deliberately: roughly 60% compounding pieces, 15% timely spikes, 25% refresh and consolidation of the existing library.
Select topics by intent tier, use search volume only as a tiebreaker, and subject ego topics to the bookmark test.
Refreshes typically return multiples of new-content effort on the same hours — a decaying URL with links and history is appreciating capital.
Model attribution directionally, say so in the report, and pre-register expected effects of refreshes and pruning.
Plan two pages, not forty: portfolio allocation and clusters; distribution and measurement. Then keep the schedule boringly consistent.

FAQ

How long before a compounding strategy shows results?

Honest answer: two quarters of feeling slow. Compounders on a fresh domain typically take four to nine months to reach their plateau. That's why the portfolio mixes in spikes and why refresh work starts paying first — your existing URLs respond within weeks. Anyone promising meaningful organic pipeline in 30 days is selling you the spike and hiding the crash.

We're a small team. Is 60/15/25 realistic?

The ratio matters more than the volume. A solo marketer publishing one piece a month should spend one month in four on refreshes. Small teams actually compound faster proportionally, because a neglected library of fifty pieces has more consolidation upside than a neglected library of five hundred.

Should we delete underperforming content?

Dead content (zero impressions for six months, no links, no links out to it internally) costs you crawl attention and topical clarity. But delete via 301 into the nearest strong sibling, not 404 — you keep the equity and tidy the library at once. Deleting into the void is throwing away the only thing the piece ever earned.

How do newsletters fit a compounding model?

As the distribution engine for the pension fund: every compounding piece gets an owned-channel cut; every newsletter issue eventually becomes or feeds a searchable piece. The newsletter spikes attention; the library banks it. Programmes that only email are renting their audience's memory. Programmes that only publish for search are banking attention nobody yet gives them.