Diversification has two axes
Years ago, an agency went quiet. An agency whose medical textbook chapters had been a steady part of my calendar, and they stopped sending them. No new files, no updates , nothing. I knew something was wrong, and I decided it was temporary. Volume dips happen all the time, and publishers pause series and pick them up next quarter.
That agency was half my income.
I had three main clients at the time, and three felt like a lot. All three were translation agencies, and all three sent me the same deliverable: medical textbooks. When the largest went quiet, I had hoped that the other two would fill the gap, but that didn’t happen. Their volume rose and fell with the same market.
It took about nine months to replace the income. I went looking for clinical trial and publication work, and I went back to two deliverables I had offered before and let lapse. Linguistic validation came first, because the pitch was short: the work asks for exactly the clinical background I had, and CME medical writing followed. Neither depended on textbook publishing, and neither came through a translation agency.
Nine months, it took me nine months to diversify my income. The number is the least useful thing in this story.
What I had been counting was clients. Diversification has two axes, but I had been ignoring one of them.
The first axis is who pays you; the second is what they buy: two clients who buy the same deliverable from you are, for risk purposes, closer to one client than to two. Their budgets answer to the same market. When it moves, they move together.
Concentration on one axis is a strategy. Concentration on both is an accident, and the accident is common because translators and writers each guard the axis they were told to guard.
Translators guard the deliverable. The specialization is the asset, and protecting it from dilution is a reasonable instinct: a medical translator who also does marketing copy is harder to refer. But the same instinct keeps your attention on what you translate and away from who sends it. Two or three agencies feel like a client base, but an agency is a channel, though, and a channel can change vendors or move the work to post-editing at a lower rate while demand for your specialization stays exactly the same. The specialization is still there but the income is gone. That was my case.
Writers guard the client. A writer with five clients and one deliverable is the mirror of my three agencies and their textbooks. Five CME providers, all buying needs assessments. The client axis looks healthy, and no single name stands out. But the deliverable is a single point, and this one can move under the writer's feet. ACCME’s guidance on AI, out since December 2025, lists identifying learning gaps as an appropriate use, and taking the gap analysis in-house means that the provider can try to send the result out for validation at a lower rate. If five providers change their workflows in the same year, five clients cut the same line item, and five names on your client list send less work.
Here is what I currently do to try to protect myself from these shifts:
Take last quarter's invoices and sort them twice. First by client. My rule is that no client goes above 25% of a quarter's income. When one does, I intensify outreach, and the aim of that outreach is a new client, so that next quarter's list is longer.
Then by deliverable. If one deliverable carries most of the quarter, the response is an adjacent one you can invoice within the next quarter. For a writer, from needs assessments alone to the content and slides that follow them. For a translator, from translation alone to the linguistic validation or plain-language work on the same documents. Adjacent work is a relatively fast route because the same clients will buy it, and it covers you against a squeeze on one deliverable. It does not cover you against a squeeze on the market those clients share. For that you need a deliverable whose demand does not move with your main one, which is the longer project and the one I ended up doing. I had offered other services before, and reactivating them still took most of the nine months.
Whichever axis you have never prioritized is the one that deserves your attention now. It is the one that can destabilize you.
Photo by Lewis Fagg on Unsplash