Pay
New entrants compress the bottom and leave the top alone
Crowding does not lower every price. It lowers the price of whatever is easiest to supply.
Guides on Pay: Pricing, from first principles to the annual review, The ceiling is hours, and it arrives sooner than people expect, What rating work actually pays
When a category fills up with new earners, the average price falls and the spread widens at the same time. Those two things sound contradictory and they are not: the bottom of the band drops hard, the top barely moves, and the average follows the bottom because that is where the new supply lands.
Which end you are standing on decides whether crowding is a problem or a non-event for you.
Why new supply lands at the bottom
Someone arriving with no record has one lever and it is price. They cannot compete on reviews, they have no repeat buyers, and their turnaround claim is unverified. So they enter at or below the cheapest visible tier, because that is the only position where being unknown is not disqualifying.
This is not a coordinated race to the bottom. It is a hundred separate people each making the same individually sensible decision, and the aggregate effect is a thick new floor of cheap, undifferentiated, fast-turnaround offers.
Nothing in that process touches the earner three tiers up who does long-form written work with a two-day turnaround and forty specific reviews. The new arrivals cannot supply that, so they cannot price against it.
The general shape economists describe
Price dispersion under competition is a well-studied result: in markets where buyers cannot verify quality before purchase, adding sellers widens the gap between the cheapest and the dearest rather than pulling everyone toward one number. Stiglitz's 1989 survey work on imperfect information markets makes the mechanism plain enough - when search is costly and quality is hidden, price stops being a single clearing number and becomes a signal that different buyers read differently. A recent measurement from freelance work has the same shape: Demirci, Hannane and Zhu's 2024 CESifo working paper found that after ChatGPT, job posts for automation-prone writing and coding fell 21% within eight months, competition among freelancers rose, and the jobs that remained were more complex and better paid.
The practical version for you is short. Buyers who were already choosing on price get more options and pay less. Buyers who were choosing on evidence get more noise to filter and pay about the same, sometimes more, because filtering has become harder and a clear signal is worth more than it was.
What holds up and what does not
The test is whether a new entrant can supply your thing in their first week.
| Offer | Suppliable on day one | What crowding does to it |
|---|---|---|
| Short generic score, fast | Yes | Price falls, often below the flat fee floor |
| Short written note | Yes | Price falls, volume rises |
| Structured written assessment to a published rubric | Not credibly | Roughly flat |
| Recorded review with a stated length | Partly | Falls, then recovers as the nervous ones drop out |
| Comparison across several submissions | No | Flat or rises, because supply is thin |
| Named specialism with a public example | No | Rises relative to the band |
The pattern is that anything a beginner can do on their first afternoon becomes cheap, and anything requiring a record, a rubric or nerve does not. The piece on what buyers pay more for across written and scored work has the underlying price differences that this table is only redistributing.
If your segment is the one being compressed
The instinct is to match the new floor. That is the one move guaranteed to make next month worse, because you will be competing on the single dimension where a person with nothing to lose beats a person with a record to protect.
Three responses actually work.
Move the offer, not the number. Add the thing the new supply cannot do - a rubric, a length guarantee, a second pass, a format nobody else lists. The reasoning behind holding your rate card while the market moves under it applies directly here.
Take the volume while it is there, if you are still early. Crowding raises total buyer traffic to a category as well as total supply, and if your problem is a thin record rather than a thin margin, more jobs at the current price is still the right trade.
Let the cheap tier go. If your entry offer is now underwater against flat processing charges, delete it rather than defend it - the arithmetic of a floor price against flat fees is unforgiving and does not improve with volume.
What crowding actually signals
A category filling up is evidence that buyers are arriving, not just sellers. People do not flood into a market with no demand in it; they flood into one where somebody visibly got paid.
So the honest read on a crowded segment is mixed. The bottom of your band is gone and is not coming back, the top is intact, and the middle is where the fight is. Buyers, meanwhile, have got better at filtering, and the buyer-side account of how someone actually shortlists a reviewer is worth reading precisely because it describes the filter you now have to pass.
Some of the compression is not competition at all. Generic instant assessment stopped being scarce when it became automated, and the technical account of what an automated score already does marks the line below which no human price is defensible, along with the tools that put that scoring in front of buyers before they ever consider paying a person.
If you want to see where a live category currently sits rather than reason about it in the abstract, the visible spread across the judges listed on Rate Cock is the cheapest read on a band available, and it updates faster than any post can.