By late September the question is no longer whether you need help. It is which returns you hand over, and that question gets answered badly because piles are counted rather than weighed.
Forty returns is not forty units of work. It is a small number of genuinely hard files sitting inside a larger number of ordinary ones, and the difference between them is bigger than most firms carry in their heads.
Two scores, not one
Our platform grades every finished return on two separate one-to-five scales, captured by the preparer when the case goes to review.
The first is complexity — how hard the return was, as an objective matter. One is trivial, five is hairy. The second is client effort: how much the client cost in coordination and chasing, which has nothing to do with how difficult the return itself was. There is a free-text note alongside them for what actually made it hard.
Keeping them apart is the whole point. A straightforward 1040 for a client who answers nothing until October is a one on complexity and a five on effort. A trust return for a responsive client is the reverse. Both are painful, and they call for opposite decisions.
What the hours actually look like
The platform ships a standard-hours grid against those complexity bands, as a starting point every firm is expected to confirm or override with its own numbers.
- Complexity 1 — trivial: 2.0 hours preparation, 0.5 review. 2.5 hours.
- Complexity 2: 4.0 preparation, 1.0 review. 5 hours.
- Complexity 3: 7.0 preparation, 2.0 review. 9 hours.
- Complexity 4: 12.0 preparation, 3.5 review. 15.5 hours.
- Complexity 5 — hairy: 20.0 preparation, 6.0 review. 26 hours.
A five costs more than ten times what a one costs. That ratio is why counting returns tells you almost nothing about how much October you have left.
Take a pile of forty, spread six, ten, fourteen, seven and three across the bands. That is 377 hours of work. The seven fours and three fives — a quarter of the pile — account for 186 of them.
Twenty-five per cent of the returns, forty-nine per cent of the hours. Send those ten out and you have halved the work while keeping three-quarters of the client contact in-house.
Run the same arithmetic on your own bands before you accept ours. The shape holds in most firms; the exact numbers will not, and the grid is shipped as a default precisely because it is meant to be replaced.
You will not have this year's scores yet
Worth saying plainly, because it is the obvious hole in everything above. The grading happens when a return goes to review, so it describes work already done. It cannot tell you about a return nobody has opened.
Two things work instead.
Last year’s grade on the same client is the best single predictor you have. Returns rarely move band year to year. A client who was a four last October is a four this October unless something specific changed, and if something did change you usually already know what it was.
For anyone without a history, grade on sight against the bands. You do not need precision. You need to know whether a file is a two or a four, and any experienced preparer can tell that from the document set in about a minute.
Send the hard ones
The instinct is usually the opposite. Firms offer up their simplest returns first, because handing over an easy one feels like less exposure.
It is less exposure. It is also where outsourcing buys you least. Sending out a two recovers five hours. Sending out a five recovers twenty-six, and it removes the file most likely to still be open on 14 October.
There is a second reason, and it matters more over a season. A hard return is where you find out whether a desk is any good. An easy one tells you almost nothing, because nothing about it was going to go wrong. If you are running a pilot to make a decision, the pilot has to contain the work the decision is about.
That is why we suggest including at least one genuinely hard return in any first five, and it is the same argument made from the other direction in the foreign-income article.
Keep the ones where you are the product
Some returns should not leave, and complexity is not what marks them.
Keep the returns where the client is really buying your judgment rather than a filed form — the ones with a live planning conversation attached, or a position you are going to have to defend, or a relationship that is genuinely about you. Keep anything where the answer depends on facts that live in your head and have never been written down.
Keep the returns you would want to have prepared yourself if they were ever examined. Not because someone else would do them badly, but because your own memory of having built the file is worth something in that conversation.
Everything else is production, and production is the thing capacity problems are made of.
The high-effort, low-fee quadrant is a different decision
The second score exists for a reason. When you plot fee against complexity, the returns that stand out are not the hard ones — they are the ones where effort was high and the fee was not.
That quadrant is a pricing problem wearing a capacity costume. Outsourcing a return whose real cost is a client who will not answer email does not fix anything, because the chasing does not transfer. Somebody still has to get the documents, and that somebody is still you.
So triage on complexity, and put the effort scores in a different pile marked for January. We wrote about what to do with that pile in fire the client, eat the hours, or find a third option.
Doing both at once in the last three weeks of an extension season is how firms end up making a repricing decision under deadline pressure, which is the worst possible time to make one.
A triage that takes an afternoon
Pull the open list. Put last year’s complexity band against every client who has one, and grade the rest on sight.
Add the hours from your own grid and total them. That number, not the count of returns, is what you are actually facing.
Mark every four and five. Strike out the ones where you are the product and the ones with a live planning conversation. What survives is the list that should leave the building, and it is usually shorter and heavier than people expect.
Then check the total again with those removed. If the remainder still does not fit the weeks you have left, the problem was never triage and no amount of sorting will fix it.
Our desk prepares expat and domestic returns — Forms 2555, 1116, 8621, FinCEN Form 114 and Form 8938, alongside 1040, 1120-S and 1065 work. Every return is prepared by one IRS-licensed enrolled agent and reviewed by a second before it reaches your firm. Your letterhead, your client, your review — our hours. Preparing US returns since 2003 — 5,300+ tax returns filed in the last 4 seasons.