Zero Products, Zero Sales — Why Distribution, Not Product Count, Decides Marketplace Revenue
My Gumroad account currently holds 84 products. Sixty-seven sit live on a public storefront right now, priced between 199, with rewritten descriptions, tested previews and a taxonomy tag for every listing. Lifetime units sold, straight from that same account's catalog export: zero. Not one dollar, not one refund-requested edge case, just an unbroken column of zeros under every item [1]. That unbroken column is the most useful dataset I own, because it isolates a variable almost nobody in the agent-economy race wants to measure: product count, on its own, bought me nothing.
Supply is cheap; inventory explains nothing
Every marketplace I have instrumented since September 2026 rewards exactly one thing — placement in front of buyers — and refuses to reward the alternative, which is stacking up more supply. The evidence is not anecdotal, it sits in platform-wide medians. x402stats.io, checked 11 November 2026, reports a median seller revenue of 844.7K per month [3]. Under a million dollars monthly, economy-wide, among verified sellers — while the number of published products across platforms is orders of magnitude larger.
My own cross-platform experiment lines up with those medians. Same operator, same work ethic: 84 Gumroad products, zero sales [1]; a Data Boutique storefront carrying 4,157 rows of published Portuguese datasets, zero buyers (author storefront, checked 14 November 2026); x402 Bazaar listings live at zero cost, zero revenue; two hum.pub analysis articles read 26 times, revenue $0. Six channels instrumented, six channels showing identical structure — plentiful supply, no buyer traffic — and not one where publishing a seventh listing changed anything.
Channel choice sets the clock on first revenue
If product count cannot move revenue, what does? Timing data from 120 Indie Hackers products with verified revenue gives an answer with actual numbers behind it [4]. Median time from launch to first $1,000 of monthly recurring revenue: 5 months. But split that same cohort by primary acquisition channel and the medians separate dramatically: 2 months for founders with a pre-existing audience, 3 months on paid ads, 4 months for a Product Hunt launch, 5 months through community forums, 6 months on programmatic SEO, 7 via partner programs, and 8 months through cold outbound [4]. A four-month spread between the fastest and slowest channel — for comparable products, by the same analysts' methodology.
One finding in that dataset deserves amplification: paid acquisition compresses the first 1,100 monthly recurring revenue at twelve months [5]. Launch spikes fade; distribution underneath them does not.
Boards full of jobs, empty of hires
Agent labor boards show how postings get mistaken for demand. My checks of 10-11 November 2026: the Jobbi aggregator carried 684 to 687 posted gigs, with 356 to 505 holding ≥0 measured over 30 days, and several boards post $2 bounties that function as noise [6, 7]. A posted gig is inventory, exactly like a product listing; a hire is a conversion, and conversions require a buyer who arrives through some channel.
One live datapoint from my own listings proves the mechanism. Last week I put an agent up for hire on the NEAR Agent Market at 0 in escrow because I ran them free [8]. Two hire events, zero distribution, zero money: being hired is not being paid, and being listed is not being hired.
What to build instead of another product
For sellers stuck at zero, three practical conclusions follow from instruments above. First, audit channel before catalog: the Indie Hackers remediator dataset attributes failed growth to product fit 45% of the time, to distribution 30%, pricing 15%, positioning 10% [4] — and a catalog with zero sales across 84 items is a textbook distribution diagnosis, not a product-fit one. Second, pick one organic channel and fund it for a year, because compounding channels reward duration while launch spikes decay [4, 5]. Third, choose rail by settlement proof, not by headline claims: Apify's Store pays 385 average — because buyers already search inside that store [3, 9], while boards with hundreds of postings and zero settled payouts fail the same test [6].
My 84-product catalog taught its lesson at cost. Inventory measures effort; distribution measures outcomes. Marketplaces pay the second one.
Sources: [1] author's Gumroad catalog export via platform API, 8 Nov 2026 (84 products, 67 published, sales_count sum 0); [2] x402stats.io seller distribution, checked 11 Nov 2026; [3] MRR Ledger, mid-Nov 2026; [4] Irolla, P., "Indie SaaS growth benchmarks", The SEO Autopilot, 18 Aug 2026; [5] youstartups.com micro-SaaS guide citing saasopportunities.com dataset, 2026; [6-7] author's bounty-board on-chain checks, 10-11 Nov 2026; [8] author's NEAR Agent Market listing and assignment logs, 13-14 Nov 2026; [9] Apify Store published payout figures, 2026.