Business Growth

Content Compounding Economics: Why Owned Media Appreciates While Paid Depreciates

Every dollar spent on ads buys an impression that disappears the moment it is served. Every strong article, benchmark, or framework you publish can keep producing attention for years. That asymmetry, paid depreciates, owned can appreciate, is the economic core of the owned-audience strategy, and it is supported by two very different bodies of evidence: vendor research showing a minority of content assets compound into the majority of returns, and peer-reviewed experiments showing paid advertising returns are routinely overstated by attribution. This article assembles that evidence honestly, vendor flags included, examines how AI search changes the compounding curve, and gives 5-7 figure service firms a two-ledger model for funding content like capital and paid like rent.

Joshua Agonya Pi'Rwot

By Joshua Agonya Pi'Rwot

Founder, Business Growth Accelerator

Executive summary

Paid traffic stops the day spend stops; evergreen content keeps producing. We examine the evidence, compounding post data, paid search incrementality experiments, and CAC math, behind treating content as a balance-sheet asset.

Section 1

The five challenges at a glance

The economic argument for owned content rests on a simple asymmetry: a paid impression is consumed the moment it is served, while a published asset can keep generating attention, links, and AI-citation visibility for years at zero marginal cost. The evidence base, however, requires honest handling. The strongest pro-content data comes from vendors who sell content tooling, HubSpot's compounding research and ProfitWell's CAC comparisons should both carry a vendor flag, while the strongest anti-paid data comes from peer-reviewed experiments like eBay's paid search study (Blake, Nosko & Tadelis, 2015). Read together, they triangulate the same conclusion from opposite directions: paid's measured returns are systematically overstated, and content's long-tail returns are systematically invisible to short-window analytics. The five challenges below explain why most service firms still get the allocation wrong, and the later sections build the corrected economic model. Two further dynamics sharpen the stakes for 2026. Auction prices on the major ad platforms keep rising as more competitors bid for the same in-market minority, so the depreciating asset is also getting more expensive per unit. And AI answer engines are redrawing how organic attention flows, punishing commodity content while rewarding citable original work. Both trends push in the same direction: firms that own appreciating content assets gain a widening cost advantage over firms that rent attention.

Section 2

Challenge one: paid media's depreciation curve and overstated returns

The canonical evidence on paid depreciation is the eBay field experiment. Blake, Nosko, and Tadelis, then at eBay, ran large-scale tests that paused paid search in selected US markets while continuing it elsewhere, publishing the results in Econometrica (Blake, Nosko & Tadelis, 2015). The findings were stark: brand keyword ads showed no measurable short-term benefit, because users simply clicked the organic listing instead, and the overall return on ad spend in the geo experiment was approximately negative 63%. Returns on non-brand keywords were positive only for new and infrequent users, while frequent users, who would have purchased anyway, accounted for most of the spend. The mechanism generalizes: click-based attribution credits ads with conversions that organic demand would have produced regardless, so platform-reported ROAS systematically overstates true incrementality. Important caveats: eBay had exceptional brand awareness and dominant organic rankings, so smaller advertisers typically see better incrementality, and subsequent literature finds paid search can work for unknown brands. But the structural lesson holds for service firms: every paid dollar buys a depreciating impression at an auction price set by competitors, and the dashboard measuring it is biased upward. When spend pauses, the asset value remaining is zero. That is rent, not investment, defensible as a harvest mechanism, indefensible as the entire growth model.

Section 3

Challenge two: the compounding evidence and its honest limits

The pro-content evidence is real but must be vendor-flagged. HubSpot Research, analyzing its own blog and a sample of business blogs, found that one in ten posts were compounding, meaning organic traffic grew over time rather than decaying, and that these compounding posts represented about 10% of posts while generating roughly 38% of total traffic (HubSpot Research, 2016, vendor study). In HubSpot's own data, 75% of blog views and about 90% of blog leads came from older posts rather than the current month's output, and a single compounding post generated as much traffic over its life as six decaying posts (HubSpot, 2015, vendor). HubSpot sells content marketing software, so the incentive to flatter content is obvious; the pattern, however, is mechanically plausible, evergreen search demand plus accumulating links produce growth curves paid media cannot replicate. ProfitWell's analysis, another vendor source, reported content-sourced customers carrying meaningfully lower acquisition costs than paid-sourced ones (ProfitWell, vendor). The honest limits: most posts decay, compounding status concentrates in evergreen problem-led topics, and these studies predate AI search. The economic conclusion survives the caveats, content portfolios behave like venture portfolios, where a minority of assets produce the majority of returns, and the firm that never builds the portfolio never holds the winners.

Section 4

Challenge three: AI search changes the compounding curve, not the logic

The 2025-2026 search environment forces an update to compounding math. Semrush's study of AI Overviews across millions of keywords found AI Overviews appeared on 6.49% of tracked keywords in January 2025 and roughly 15-25% through the year, increasingly on commercial queries, with independent analyses reporting substantially reduced organic click-through where Overviews appear (Semrush, 2025, vendor; ppc.land, 2025). For content built purely to win informational clicks, this is genuine depreciation risk: the asset still ranks, but the click never arrives. Three factors keep the compounding logic intact. First, AI answer engines cite sources, and well-structured, evidence-dense pages earn citations that function as the new impression, brand exposure inside the answer itself. Second, the assets least affected are those answer engines cannot replace: original research, strong opinion, first-hand experience, tools, and brand-named frameworks. Third, content compounds across more surfaces than search, email lists, communities, podcast back-catalogs, and social archives are unaffected by SERP redesign. The portfolio rule changes from 'rank for queries' to 'be the source machines and humans both cite.' Service firms publishing commodity explainer content will watch its value erode; firms publishing citable proof, benchmarks, client data patterns, named methodologies, will find AI search amplifies their compounding, because answer engines need exactly what thin content farms cannot produce.

Section 5

Innovative solutions

Sophisticated operators are restructuring content economics around five moves. First, capex accounting for content: treat each substantial asset as a capital project with an expected payback window of 6-18 months, tracked in a simple asset register, publish date, cumulative sessions, conversions, citations, so compounding becomes visible instead of buried in monthly dashboards. Second, the portfolio bet structure: since roughly one in ten posts compounds (HubSpot Research, 2016, vendor), plan in batches of ten-plus assets per theme and reallocate refresh effort to proven compounders; HubSpot's own historical-optimization work showed updating old winners outperformed publishing new posts for incremental traffic (HubSpot, 2015, vendor). Third, incrementality audits of paid: replicate eBay's logic at small scale by pausing brand-keyword spend for two to four weeks and watching whether total conversions actually fall (Blake, Nosko & Tadelis, 2015). Fourth, multi-surface syndication: every flagship asset becomes an email series, short-form clips, and a podcast talking track, so one production cost yields returns across surfaces with independent decay rates. Fifth, citation engineering for AI search: original numbers, named frameworks, clear claims with sources, and machine-readable structure increase the odds answer engines quote you, converting the zero-click threat into a zero-cost branding channel. Together these moves convert content from a marketing expense line into a managed asset portfolio.

Section 6

Solution framework

The framework is a two-ledger model. Ledger one, operating spend (paid, depreciating): every paid channel is evaluated on true incremental contribution, not platform-reported ROAS, with the eBay finding as the standing prior that dashboards overstate returns (Blake, Nosko & Tadelis, 2015). Paid's legitimate role is harvesting existing demand and accelerating tests; its budget should never exceed what survives an incrementality discount. Ledger two, capital investment (owned, appreciating): content, email list, community, and research assets are funded like capex with multi-quarter payback expectations. Allocation discipline follows the effectiveness literature's direction, a meaningful majority of growth investment into demand-creating assets once basic capture is in place (Binet & Field / IPA, 2013). Within ledger two, apply the 10x portfolio rule: commission assets in thematic batches, expect roughly one in ten to compound (HubSpot Research, 2016, vendor), and concentrate refresh investment on demonstrated winners rather than spreading effort evenly. Each asset must pass the citation test before production: does it contain something, a number, a framework, a documented result, that a human or an answer engine would have to credit us for? Governance: review the asset register quarterly, write down assets that have decayed, and capitalize the lesson. The output is a firm whose marketing balance sheet grows every quarter, instead of a P&L that resets to zero every month.

Section 7

Evidence-based action plan

Days 1-15: build the asset register. List every substantive content asset with publish date, trailing-12-month sessions, conversions influenced, and external citations. Identify your existing compounders, most firms find a handful of old assets quietly producing the majority of organic value, mirroring HubSpot's 10%/38% pattern (HubSpot Research, 2016, vendor). Days 16-30: run the paid incrementality check. Pause brand-keyword spend for a defined window with a clean before/after comparison; if total conversions hold, you have located pure waste, exactly as the eBay experiments did at scale (Blake, Nosko & Tadelis, 2015). Reallocate verified waste into the content capex budget. Days 31-60: commission one batch of ten assets against a single theme your buyers research before they are in-market, each engineered for citability, original data, a named framework, sourced claims. Refresh your top three proven compounders simultaneously; updated winners typically return faster than new bets (HubSpot, 2015, vendor). Days 61-90: install the two-ledger review. Monthly: paid channels versus incrementality-adjusted contribution. Quarterly: asset register growth, cumulative organic sessions, email subscribers per asset, AI-answer citations, and the share of inbound leads referencing specific content. Success at day 90 is not a traffic spike; it is a register showing appreciating assets, a paid budget pruned of non-incremental spend, and a publishing system that produces the next batch without founder heroics. For adjacent evidence in this pillar, see [The Trust Transfer: Borrowing Audiences Through Guesting and Partnerships](/blog/growth-trust-transfer-borrowed-audiences) and [First-Party Data After Cookies: The Factual State of Tracking in 2026](/blog/growth-first-party-data-after-cookies).

FAQ

Direct answers for operators.

What is a compounding content asset?

A compounding asset is content whose traffic and influence grow over time instead of decaying after launch. HubSpot's research, a vendor study, found such posts were about 10% of business-blog articles yet drove roughly 38% of traffic, with monthly visits multiplying for nearly two years post-publication. They are usually evergreen, problem-led pieces aligned with persistent search and buyer questions rather than news or campaigns.

Is paid advertising actually a bad investment for service firms?

Not inherently, it is a mismeasured one. The eBay experiments published in Econometrica found brand-keyword ads produced no measurable benefit and overall paid search returns were far below attribution claims, because ads were credited for demand that already existed. Paid works best for harvesting demand and for brands without organic visibility. The error is funding paid on dashboard ROAS and treating it as the growth engine.

How should a small firm budget between content and paid?

Run two ledgers. Fund paid as operating expense capped at its incrementality-verified contribution, test by pausing brand keywords and watching total conversions. Fund content as capital expenditure in batches of roughly ten assets, expecting a minority to compound and carrying a 6-18 month payback horizon. Most growth-stage service firms land near a 50:50 to 60:40 split favoring owned-asset creation once capture basics work.

Joshua Agonya Pi'Rwot

Written by

Joshua Agonya Pi'Rwot

Founder, Business Growth Accelerator · Country Director, AVODA Group Uganda · EMBA

Joshua helps service-business operators turn scattered marketing into a clear path from first attention to booked call. He is Founder of Business Growth Accelerator and Country Director of AVODA Group Uganda.