The math on paid media has been getting worse for six years. Meta CPMs are up around thirty percent since 2023. Google CPCs are up twenty-five. Click quality is down. Attribution is increasingly broken. Most of the post-2020 paid-marketing playbook now produces lower ROAS at higher cost.
Meanwhile a quietly-built newsletter with four thousand of the right readers, reaching them weekly for two years, is now worth more than most six-figure ad campaigns. The math here moved in the opposite direction. The cost of producing the work fell. The cost of placing it stayed near zero. The half-life of a piece extends from days to years.
The category has a name. Earned attention. Most marketing teams under-invest in it because the unit economics are alien to anyone who came up through performance.
The unit economics nobody on the team has run
A useful number. The average B2B newsletter with engaged subscribers, properly mailed, drives a marginal pipeline lift in the range of one to three percent of subscribers per quarter. For a list of four thousand qualified readers in your category, that is forty to one hundred and twenty pipeline conversations a quarter, on a cost base of roughly the salary of one good writer.
The equivalent volume of pipeline through cold outbound, at typical 2026 conversion rates, costs four to six times as much per opportunity, before you count the brand damage from cold sequences sent to people who would rather not be cold-sequenced.
The catch is that the newsletter takes eighteen months to two years to compound to that number. The first six months produce almost nothing measurable. The next six produce something. Year two is where the asset starts paying. Most marketing teams in 2026 cannot tolerate eighteen-month investments, so they keep buying ads instead, and keep wondering why their pipeline costs more every quarter.
Three formats that work, one that mostly doesn't
- The category newsletter. Weekly cadence. One named writer. A specific take on the industry the audience is already in. Examples that worked in the last five years: Lenny Rachitsky for product, Stratechery for tech, Workweek's portfolio for everything else. The format is forgiving as long as the writer has a point of view and ships on time.
- The interview podcast with a specific audience. Not "general business". A specific cut of your category. Twenty minutes is enough. Editing matters more than equipment. The compounding effect comes from the guests you book, who become advocates, not from the listenership which is usually smaller than the team hopes.
- The single-track conference. Yearly or twice-yearly, two to three hundred people, no sponsor village, no panels. Most modern conferences are too big and too noisy to compound for anyone. A small, opinionated, well-curated event with a clear point of view creates a community that buys from each other and from you for a decade.
The format that mostly does not work is the corporate blog with five contributors and no consistent voice. Search killed it in 2023. Assistants will finish the job by 2027.
Why most companies still don't do it
Three reasons, in order of how often we hear them.
The patience problem. Earned attention compounds slowly. A quarterly review cycle is the worst possible cadence to evaluate it on. Most teams cancel the project at month nine, exactly when the line was about to start curving.
The author problem. A newsletter with no named author reads like a press release. A podcast hosted by a marketing manager who clearly does not know the audience sounds like a corporate event. Earned attention needs a real person with a real point of view at the front of it, and most companies do not have that person ready, or the person who is ready does not have the time, or the time is not budgeted, or the person has left for a competitor before the asset compounds.
The attribution problem. Earned attention sources pipeline that arrives months after the touch, through channels the dashboard cannot see. Marketing teams beholden to last-touch reporting find it impossible to justify the spend. The CFO sees a line that does not move and a line that does, and cuts the one that does not, exactly when it was about to.
How to start, cheaply
You do not need a media operation. You need one person with a point of view, twenty hours a fortnight, and a two-year horizon nobody can revoke at the quarterly review.
One option. Pick the smallest possible audience cut you can compound against — not "everyone in finance", but "CFOs at fifty-to-two-hundred-person UK fintechs who report to a non-finance CEO". Three hundred named accounts. Write something for them every two weeks that they would forward to one another. Build the list by hand. Do not buy ads to grow it. Do not gate the content. Year two will tell you whether the bet was right.
Earned attention is the only marketing line that gets cheaper per quarter, indefinitely, after the first year. Everything else costs more next year than it did this one.
The teams that get this right look identical to the teams that fail at it for the first year. The dashboards say almost nothing. The pressure to abandon the project peaks at month ten. The teams that survive the pressure are the ones whose CMO took a deliberate decision, in writing, at month zero, to not measure the asset on quarterly pipeline contribution.
The companies that ship the work get to a place, around year three, where they are no longer paying for distribution. Their distribution is paying them. Most of their competitors are still buying clicks at twice the price.