Substack Notes and Boost buying for paid newsletters

Substack sells no advertising against Notes. There is no auction, no impression inventory and no targeting layer. The paid routes into that feed are three: the Boost network, where you pay another publisher per subscriber delivered; direct restack and cross-post agreements negotiated one at a time; and traffic bought outside Substack that lands on a Note rather than a signup page. We price all three against one figure, cost per subscriber still opening at day 90, and on the family and local-services lists we have worked that has landed between four and eleven euros. A quote priced per follower is pricing the wrong object. The first decision is whether to spend at all: We take the publication as it is and decide, inside the first two weeks, whether paid acquisition is the right instrument. That rests on your own retention curve, not on ours. We pull six months of subscriber and open data from the Substack export, cohort it by acquisition source, and look at where free-to-paid conversion actually happens. A publication converting below roughly two percent of its free list does not have an acquisition problem, and we say so before invoicing anything past the audit fee. Where the money goes: Boost spend is negotiated publisher by publisher, never set as a blended bid. We read each candidate's Notes activity and recommendation graph before committing, because a Boost from a publication whose own list came from a giveaway delivers subscribers who behave like giveaway subscribers. Restack agreements carry a defined post count and a defined window, and we hold the brief for the restacking author instead of letting them improvise a caption. Outside traffic goes to a Note, then to the about page, then to subscribe. Three steps, because the one-step version buys addresses rather than readers. What we do not do: We do not buy followers, restacks or Notes engagement from any source that cannot be named in your report. We do not run lead magnets, PDF gates or referral-prize mechanics; they move the signup number and destroy the day-90 number, and we would rather not have that argument in month three. We do not write your Notes. We do not touch your editorial calendar, and we will not tell you to post more often and call it a strategy. Reporting: Every fortnight, one sheet: spend by source, subscribers delivered, day-30 and day-90 open rate by source cohort, free-to-paid conversion by source cohort. Platform-reported figures and our own reconciliation sit in adjacent columns, and where they disagree the gap is printed rather than averaged away. On escrow-measured placements the delivered subscriber count is what settles the invoice. Who this is not for: Publications under roughly two thousand free subscribers, where the arithmetic on Boost pricing never closes. Publications whose paid tier launched in the last eight weeks and has no retention history to read. Anyone whose objective is a headline subscriber count for a fundraise or a sale; we have declined that twice and would again, because the channels that produce it are the ones that make the list worthless afterwards. Languages: Work runs in English and Spanish. Spanish-language publications are read and briefed by the strategist who runs the buy, not handed to a translator afterwards.