On this page
The donor who still doesn't know what you do
A long-time donor, after five years of steady giving and several cordial visits, sends a short note to the executive director. The note is kind. It also contains a sentence that, if the leader is honest, lands harder than anything a critic has said this year: I still don't really know what you do.
The organization has an archive of roughly four hundred pieces of work. Articles, blog posts, white papers, talks, conference decks, podcast episodes, a small shelf of published books, three years of campaigns, a handful of courses, a regularly published newsletter, a set of case studies, a flagship teaching or program series, a research report, and two dozen board-facing memos that were repurposed as public-facing commentary. The archive is not thin. The staff worked hard on all of it. Each individual piece, read on its own, is competent and well-intended.
And yet, after five years, the donor still does not know what the organization does. That sentence is not a failure of the donor's attention. The donor has read. The donor has given. The donor has come to events. The donor has nodded in the right places during site visits. The donor's confusion is, precisely, what a fragmented archive produces. It is the predictable result of an archive in which each piece is an island — unlinked to the next piece, built on no shared foundation, written without reference to the work that came before, and filed without any mechanism for being found again by anyone who might build on it.
The organization has been paying a tax on all four hundred pieces, quarter after quarter, without ever writing the tax down as a line item. That is the subject of this piece.
What the tax actually is
Every piece of work an organization produces is either connected to the rest of its work or it is not. Connected work compounds. Unconnected work pays interest.
The interest shows up in several places, each of which is small enough to dismiss on its own, and all of which are crushing in aggregate.
The first kind of interest is discoverability interest. An unlinked piece is findable only by someone who already knows it exists. Nothing downstream points at it. Nothing upstream references it. Search engines rank it as a one-off. Internal readers cannot surface it when they need it. External readers cannot move from it to the next relevant piece. The work gets produced, published, and buried on the same day. The cost of burying it is carried forward every quarter in the form of work the organization has to redo.
The second kind of interest is staff attention interest. Every time a staff member cannot find the prior piece on a subject, they rebuild it. Every new hire spends months discovering, piece by piece, the positions the organization has already taken, because no one has ever said those positions in one place. Every internal framework gets reinvented in parallel by three teams who do not know the other teams have already done the work. The tax on staff attention is enormous, and it is almost never visible on the scorecard because no line item exists for work we did twice because we forgot we had done it once.
The third kind of interest is donor mental model interest. Donors hold a rough mental picture of what the organization is and does. That picture is built slowly out of every touchpoint the donor has. If each touchpoint is a different-shaped island — a long-form talk here, a white paper there, a campaign in between, a podcast appearance off to one side — the donor's mental picture cannot assemble. The donor will describe the organization differently to two of their friends, and neither description will quite match what the organization says about itself. At scale, across hundreds of donors, this produces an organization whose external reputation is quieter, fuzzier, and less specific than the sum of its actual output would support.
The fourth kind of interest is organizational coherence interest. The organization that cannot describe itself in one paragraph, and cannot point at the twenty pieces that paragraph rests on, is not more complex than its peers. It is more fragmented than its peers, and the fragmentation is paid in every strategic conversation that has to begin by re-establishing what the organization is. Leadership time, board time, partner time, hiring time — all of it carries a surcharge because the foundation is not laid down.
Four kinds of interest, compounding quietly. Most leaders are paying all four. Very few have written the line item.
How the fragmentation happens in practice
The mechanics are almost always the same, regardless of sector. A recurring teaching series goes out. It is good. Nobody turns any of it into an article, because the article would require a different mode of work and there is no one whose job description includes the translation. Three months later, a staff member writes an article on a related theme, unaware that the teaching series exists, and the article restates things the teaching series had already said better. A year later, someone proposes a course. The course curriculum is built from scratch, because no one remembers either the teaching series or the article, and the person building the course is new. A book comes out of the course two years later. The book is published, sells modestly, and is not referenced in any subsequent writing the organization does, because the book lives in a different operational workflow than the blog, which lives in a different workflow than the newsletter.
Each handoff is where the connection should have happened. Each handoff is where the connection did not happen. The organization produced five pieces of work on substantially overlapping material, and a reader who found one of them has no way to get to the other four. The reader encounters fragmentation. The organization experiences production. Both are correct about their own experience, and the gap between them is the tax.
Multiply this pattern across a decade and you get the four-hundred-piece archive in which the donor, reasonably, cannot recognize the organization.
Why AI raises the rate
The dynamic above has existed for as long as organizations have produced content. The reason this piece lands here, at this moment in the book, is that AI is about to raise the tax rate sharply, in a way most organizations have not yet registered.
AI lowers the cost of production. It does not lower the cost of coherence. That is the sentence the rest of this piece hangs on, and the entire argument follows from it.
An organization that was producing four hundred pieces in ten years will, with AI-assisted tooling, easily produce four hundred pieces in two. The unit cost of a draft is now small enough that most of the internal friction that used to prevent production will disappear. Staff will ship more. Teams will launch more. Campaigns will cycle faster. The archive will grow at a rate that was physically impossible before.
The cost of coherence, meanwhile, has not moved. Coherence requires a human to read across the body of work, hold the organization's position in their head, and make the connection decisions that AI does not make on your behalf. A single staff member can do this for a four-hundred-piece archive only with effort. No one can do it for a four-thousand-piece archive. The cost of that human-held coherence is roughly linear in the size of the archive, and AI has produced a step-change in archive size with no corresponding step-change in coherence capacity.
The result is that organizations about to enter their AI-assisted production era will produce more work per quarter than they have ever produced, and less of it will be connected to the rest of their body of work than at any point in their history. The tax, already invisible, will double. The donor who did not know what the organization did after five years will, in five more years, know even less — not because the organization has been idle, but because it has produced a volume of work that has outrun its own capacity to make sense of itself.
Any organization whose response to AI is primarily a production response is, without knowing it, volunteering to pay this higher tax rate for the rest of the decade.
What connection actually requires
The remedy is not "more content strategy." That phrase is a symptom of exactly the fragmentation it claims to address. Connection requires something that most organizations do not have and most peer organizations will not develop: a core library.
A core library, in the sense this piece uses the phrase, is not a reading list. It is the small body of work — somewhere between twenty and fifty pieces — that the organization stands on. The arguments, frameworks, stories, and positions that define what this organization, specifically, is for. Everything else the organization produces is downstream of that library. An article riffs on an anchor argument. A course is built on two or three anchor pieces. A campaign extends an anchor story into a new surface. A book consolidates ten anchor pieces into a form that can sit on a shelf.
Without that spine, nothing can connect. Each piece has no anchor piece upstream to point back to, and no framework to point down from. Connection, at that point, is literally impossible — not hard, impossible — because there is nothing to connect to.
With the library in place, connection becomes cheap. Every new piece can reference anchor arguments by name. Every staff member can be onboarded onto the spine rather than onto the full archive. Every campaign can draw on anchor stories rather than inventing fresh narrative from scratch. Every partner, donor, and reader can encounter the library as the backbone of the organization's public position, and every subsequent encounter — an article, a talk, a podcast appearance — will read as part of the spine rather than as a free-floating piece of content.
This is why the path through the AI era cannot start with production. Starting with production, at the new cost curve, is volunteering for the higher tax rate. The only move that works is to build the core library first, and then to let the AI-era production serve that spine rather than dilute it.
Who is paying, and what they are paying in
Staff pay in morale. A year of reinventing frameworks that already exist somewhere in the archive, a year of producing pieces that do not compound, a year of watching good work get buried on the day it ships — this is how communications teams, research teams, and teaching teams burn out. It is not hours. It is the interior experience of producing work that does not seem to matter past the week it goes out, in an organization that cannot tell its own staff what the work is building toward.
Donors pay in incoherence. The small, slow, mostly unspoken drift of a donor from a clear mental model to a fuzzy one is not recoverable with a better annual report. It is a predictable consequence of fragmentation. By the time a long-time donor sends the sentence I still don't really know what you do, many other donors are thinking it and not writing it.
The people the organization exists to serve pay in a fuzzy understanding of what the organization is actually offering. In mission-driven work, that fuzziness translates into a lower rate of the thing the mission was founded to produce — people taking the specific action the organization's body of work was meant to make possible. The tax, ultimately, is paid by the mission.
And the mission is paid in velocity. An organization whose work compounds moves faster every year. An organization whose work fragments moves slower every year, regardless of effort. The interior experience is the same either way. The trajectory is not.
The reframe this piece is asking for
Stop measuring output. Start measuring connection. The first metric is cheap, flattering, and misleading in the numbers. The second metric is harder to produce, unflattering at first, and the only one that tracks the thing that actually matters for the organization's trajectory.
A smaller body of connected work outperforms a larger body of fragmented work on every metric that matters — clarity to donors, legibility to partners, recall among staff, durability against turnover, findability in a search-saturated environment, and compound growth of mission impact over a decade. This is not an aesthetic claim. It is about how the work is wired together, not how it looks. Connection compounds. Fragmentation compounds. Both do so silently. One serves the mission. The other taxes it.
The next chapter goes one layer deeper, into the individual unit of the archive. Because the tax is not only paid in the connections between pieces. It is also paid in what happens — or fails to happen — inside a single piece of work once it is published. Most organizational content, even when it is produced with care, simply does not move. Understanding why that is true is prior to fixing it.
Read next: Content That Doesn't Move — why most organizational output sits still, and what the alternative actually requires of a single piece of work.