The Barrel Dump Sheet Is Written in Pencil and TTB Form 5110.40 Is Due on the 15th. 2026's Readers Finally Read Pencil

Two thousand five hundred and twenty numbers. That is roughly what a fourteen-tank cellar generates in a month if your cellarman writes down six figures per vessel per day — gravity, temperature, pH, dissolved oxygen at transfer, tank level, and whatever he noted about the krausen. All of it in pencil, on a clipboard that hangs on a nail by the glycol manifold, in handwriting that got worse as the shift went on.

Every one of those numbers has to end up somewhere else. In Ekos or Beer30 or whatever your brewery runs. On the quarterly Brewer's Report of Operations. And in the distillery next door — if you run a combined operation, and more of you do every year — on TTB Form 5110.40, 5110.11 and 5110.28, all three due by the fifteenth of the following month, every month, forever.

Fourteen pages of pencil, due on the fifteenth

Watch what actually happens between the clipboard and the filing. Somebody keys it. Usually the cellar lead at the end of shift, tired, standing at the office computer with beer on his boots, or the office manager on Monday from a pile of pages she cannot always read. The barrel dump sheet is worse: proof, temperature, wine gallons, the correction from the gauging tables, the proof gallons that come out the other end, entry proof on the refill, barrel numbers copied off chalk on the head.

Then there is the paper that arrives from outside. The certificate of analysis from your hop supplier, emailed as a scanned PDF, carrying the alpha acid percentage for that specific lot — the number your bittering calculation depends on, and the number that gets typed into the recipe by hand every single time. The malt analysis from the grain supplier. The bill of lading for the pallet of cans, with a handwritten note from the driver about the two crushed layers on the bottom. The distributor purchase order that a small independent house in a rural market still sends by fax, because they always have.

Nobody thinks this is a good system. Everybody has decided it is not worth fixing, because the previous generation of scanning software wanted a flatbed scanner and a clean printed form, and a barrel dump sheet is neither.

What actually got better this year

The thing that changed in 2026 is not that software can read documents. It is that it reads your documents — the ugly ones. A phone photo of a clipboard taken at an angle under the tank deck lights. A carbon copy where the second sheet is faint. A fax that came through at half resolution. Initials in the margin next to a crossed-out gravity reading with the corrected number written above it. Pencil.

Reading the artifact means the software works from a photograph of the actual page — the crossed-out line, the margin note, the smudge — instead of requiring a person to retype what it says. That is the whole shift, and it is why this matters now and did not in 2024. Your cellar crew is not going to walk a wet clipboard to a scanner. They will take a photo, because they already photograph things twenty times a shift.

It also handles the mixed page. A hop certificate of analysis has a printed table, a lot number, a supplier stamp and often a handwritten note from your own receiving lead about which pallet it came in on. Older tools read one of those and dropped the rest.

flowchart TD
  A["Cellar log, pencil on a clipboard"] --> E["Reader pulls the figures off the photo"]
  B["Barrel dump and gauge sheet"] --> E
  C["Hop certificate of analysis, scanned"] --> E
  D["Faxed distributor purchase order"] --> E
  E --> F{"Does the gauge math tie out?"}
  F -->|"Ties out"| G["Posts to Ekos, drafts the TTB monthly report"]
  F -->|"Does not"| H["Line flagged for the head distiller, both numbers shown"]

The error rate nobody measures because it never looks like an error

Here is the uncomfortable part. Keying errors in a brewery do not announce themselves. A transposed gravity reading does not throw an alert; it quietly makes a batch look like it finished at 1.014 instead of 1.041, and six weeks later your head brewer is trying to explain an attenuation trend that never happened. A proof gallon miskeyed on a dump sheet does not bounce; it goes onto the monthly report and stays there until TTB or your own year-end reconciliation finds a gap between what you produced, what you bottled, and what you paid excise on.

Assume, conservatively, a half-percent error rate on hand-keyed handwritten figures — lower than most published estimates for manual data entry, and deliberately kind. On 2,520 cellar figures a month that is about thirteen wrong numbers, every month, permanently in your records. Add forty barrel dumps a month at eight figures each and you get another sixteen. Twenty-nine bad numbers a month, none of which look bad.

The ones that hurt are concentrated in three places: proof gallons on the dump sheet, alpha acid on the hop certificate, and the loss and destruction lines that feed your excise return. Those are the ones worth checking twice regardless of what software you use.

What the swap actually saves, in hours and in mistakes

Assume a combined operation: fourteen fermenters, forty barrels dumped a month, a taproom, and one office manager who owns compliance. Illustrative figures — run yours.

About $915 a month of labor, which for most small breweries is the difference between the office manager finishing the excise return on the 12th and finishing it at 11 p.m. on the 14th. But the labor is the smaller prize. The bigger one is that the disagreements surface at the moment they happen instead of at year end, when reconciling a nine-month-old gap costs a week and possibly a conversation with a regulator.

Where you keep a person in the loop

Do not let any of this file anything. The reader's job ends at a draft. Your monthly TTB reports and your excise return get signed by a human who has looked at the numbers, because that signature carries real personal weight and no software should be standing behind it.

Second, gauging stays with the person licensed to do it. Reading a hydrometer, correcting for temperature, applying the tables — that is a trained distiller's job and the reader is only transcribing what they wrote down. If it thinks the arithmetic on the sheet is wrong, the right behavior is to flag the line, show both numbers side by side, and wait. Never to correct it silently.

Third, watch the handwriting it struggles with. Every cellar has one person whose sevens look like ones. When you see the same field flagged for the same person three weeks running, the fix is a five-minute conversation, not more software.

Start with one form, not the whole cellar

Pick the single worst document you handle. For most combined operations that is the barrel dump and gauge sheet, because it has handwriting, arithmetic, and a regulator on the other end. Photograph one month of them — you probably have them in a binder already — and compare what comes back against what your office manager keyed at the time.

You will find disagreements. Some will be the reader; more will be the keying. That comparison is the whole business case, it costs you an afternoon, and it tells you the real error rate in your own operation instead of an illustrative one from a trade column.

Frequently asked questions

Our cellar has no Wi-Fi past the bright tanks. Does this still work?

Yes, because the phone holds the photos and uploads when it walks back into range. Photography does not need a connection. If your crew is used to Ekos timing out on the tank deck, this will feel different for exactly that reason — the capture happens where the work happens, and the reading happens later.

Can it handle our hop supplier's certificate of analysis, or do we still key the alpha acid?

Those are among the easier documents, because they are structured and consistent per supplier even when they arrive as a scan. Lot number, alpha and beta acids, moisture, crop year, harvest date. The practical win is not the two minutes of typing; it is that the alpha acid actually used in the batch calculation matches the lot in the cooler, which is the sort of thing you find out you got wrong when a flagship suddenly tastes different in October after you switch crop years.

What about the old binders? Ten years of dump sheets in the file room.

Worth doing selectively, not wholesale. Photographing a decade of paper has a cost and most of it will never be looked at again. The exception is your barrel inventory: if you have barrels aging that were entered before your current system, getting those dump and entry records into a searchable form pays for itself the first time somebody asks which barrels from the 2021 fill are ready and you can answer in a minute rather than a morning.

Will TTB accept records kept this way?

TTB cares that your records are accurate, complete, kept at the premises and available on demand — not what you typed them with. Keep the original paper as well; it costs nothing to keep a binder. Confirm your own record retention practice with whoever handles your permit work, because the details differ between a brewer's notice and a distilled spirits plant.

The same reading ability is why the phone stops being a dead end. A distributor rep who texts a photo of a damaged pallet, a customer who sends a picture of a can with a smeared date code — those arrive as pictures and used to sit until someone had time. CallSphere builds voice and chat agents that answer your brewery's line and web chat when the taproom is slammed, take the tour and private-event bookings, and route the ones that need your packaging lead to a person rather than to a voicemail box nobody empties.