Moving a Winery Off Spreadsheets: What Breaks and When
Spreadsheets are legitimate winery records. Federal wine regulations specify what information a record must contain rather than what software holds it, so a well kept spreadsheet satisfies the requirement. What spreadsheets cannot do is preserve relationships: a blend of three lots into one is easy to write and nearly impossible to trace backward, and a cell edited six months later is indistinguishable from one entered correctly. The crossover point is driven by lot count rather than case count, and it usually arrives when a second person starts entering data. The migration itself should be timed to a vintage boundary, with complete history migrated for wine still in bond and archived spreadsheets retained for wine already sold.
Are spreadsheets even allowed?
Yes. Federal wine regulations prescribe what information a record must contain, not what form it takes, and records kept in the ordinary course of business are acceptable as long as they are complete, retained for the required period, and producible for inspection.
So the case for moving is not that spreadsheets are prohibited. It is that they are weak evidence. A spreadsheet has no audit trail: a cell edited last March looks exactly like one entered correctly in September, and there is no way for the winery or a reviewer to tell the difference. When the requirement is a contemporaneous record of what happened, a document that can be silently rewritten satisfies the letter and not much of the purpose.
That distinction matters most in exactly the situation you least want to be in. A winery whose records are unambiguous has a short conversation with a reviewer. A winery whose records are a workbook with unknown edit history has a longer one, and the length of that conversation is the real cost of the spreadsheet, deferred.
What breaks first, and why
Lot lineage, every time. Everything else fails afterward and for the same underlying reason: a spreadsheet stores results, and winemaking is made of relationships.
| What breaks | How it shows up |
|---|---|
| Lot lineage through blending | A blend is entered as a new row. Six months later nobody can say what went into it in what proportion, which also destroys the evidence for appellation and vintage label claims. |
| Volume reconciliation | Transfers are recorded as destinations rather than as movements with volumes, so losses are inferred at year end instead of measured, and the inventory drifts. |
| Additions history | Additions are logged per tank per day rather than per lot, so when the lot moves vessels its history stays behind in the old tank's column. |
| Concurrent editing | Two people work during harvest. One copy becomes "final v3 KN edit", and the other has the Tuesday numbers. |
| Formula drift | A row inserted outside a summation range silently drops a lot from a total that everyone continues to trust. |
| Field capture | Nobody carries a laptop into a cellar, so readings are written on tape and transcribed later, or not. |
| Cost accumulation | Costs land in a bookkeeping system by vendor and date, never attaching to the lot, so cost per case is an annual estimate rather than a fact. |
The blending failure is worth dwelling on because it is the one that cannot be repaired later. Volume drift can be corrected at the next physical inventory. A missing addition can sometimes be reconstructed from an invoice. But once a blend has been recorded as a result rather than as a relationship, and the people who made it have moved on to the next vintage, the proportions are genuinely gone. Any label claim that depends on them is now an assertion rather than a documented fact.
When does the crossover actually arrive?
When lot count exceeds what one person can hold in their head, which is usually far earlier than case count would suggest. A winery making 2,000 cases from eighteen separate lots is a harder tracking problem than one making 8,000 cases from four.
Three specific triggers show up repeatedly. The first is a second person entering data, because a spreadsheet has no concept of who wrote what and no mechanism for two people to work at once. The second is the first serious blend, when a wine stops having a single history. The third is a question nobody can answer quickly: how many gallons of the 2025 Cabernet are actually in bond right now, what did lot 24-03 cost per gallon, when was tank 7 last sulfured. If answering takes an afternoon of reconstruction, the spreadsheet has already stopped being a record and become an archive.
A useful self test: pick a lot at random and try to produce its complete history, fruit source through current volume, including every addition and every transfer, in ten minutes. Wineries that can do this are fine on spreadsheets for now. Wineries that cannot have a records problem that predates any software decision.
When should the migration happen?
At a vintage boundary, in the gap between the end of one harvest cycle and the start of the next. This is the only window where the winery is not actively generating the kind of data that a half migrated system handles worst.
Migrating mid-vintage is the source of most bad migration stories. During harvest, capture matters more than at any other time and attention is scarcer than at any other time, and asking a cellar crew to record the same event in two places for six weeks reliably produces it being recorded in neither. In the northern hemisphere the practical window is winter through early spring: after the previous vintage is through primary and malolactic, before ripeness sampling starts.
Give it more runway than the vendor's implementation estimate suggests. The software configuration is rarely the long pole. Deciding how your lot naming should work, agreeing what a "lot" even is across people who have used the word differently for years, and getting a crew comfortable with a new capture method are all human timelines, and they do not compress.
What data has to move
Everything about wine you still hold, and enough about wine you have sold to satisfy retention. Splitting the job that way keeps it finite.
| Data | Migrate fully | Why |
|---|---|---|
| Lots in bond: volume, vessel, vintage, variety | Yes | The opening inventory the new system starts from |
| Lab history for current lots | Yes | Trend lines and current chemistry drive live decisions |
| Additions history for current lots | Yes | Required records, and needed for total SO2 and allergen questions |
| Barrel inventory with fill dates and cooperage | Yes | Fill number and time in oak are otherwise unrecoverable |
| Finished goods by SKU | Yes | Opening balance for inventory and removals |
| Accumulated cost per lot | Yes, if it exists | Cost per case is wrong without it and cannot be reconstructed later |
| Fruit source and contract records | Yes, for current lots | Appellation and vintage label claims depend on them |
| Sold vintages | Summary only | Retention satisfied by archived originals if they stay readable |
Two things get forgotten reliably. Accumulated cost is one: if it was never attached to lots in the first place, migration is the moment to decide whether to reconstruct it or accept that pre-migration vintages will never have accurate cost per case. Barrel fill history is the other, and it is worse, because barrel age is not derivable from anything else. A cellar that migrates without it has permanently lost track of which barrels are on which fill.
Whatever you leave behind, keep the original files and make sure they remain readable for the full retention period. Federal wine records must be kept for not less than three years, extendable by up to three more, and that obligation follows the winery rather than the software.
How to sequence the move
Six steps, in this order. The ordering matters more than the speed.
| Step | What happens |
|---|---|
| 1. Inventory what you actually have | Find every workbook, tab, and side document in use, including the ones on someone's laptop. Migration scope is always larger than the official spreadsheet. |
| 2. Agree the vocabulary | Define what a lot is, how lots are named, and what constitutes a new lot versus a continuation. Do this before importing anything; it is nearly impossible to change afterward. |
| 3. Take a physical inventory | Migrate measured volumes, not book volumes. Starting a new system from numbers you already doubt guarantees the doubt persists. |
| 4. Import and reconcile | Load lots, vessels, barrels, lab and additions history, and finished goods. Reconcile the imported totals against the physical count before proceeding. |
| 5. Train on real work | Have the crew record actual operations in the new system, not a sandbox. The questions that matter only surface against real lots. |
| 6. Set a hard cutover date | On that date the spreadsheet becomes read only and archived. Announce it, hold it, and remove edit access rather than relying on discipline. |
Step six is the one most often softened, and softening it is how migrations fail slowly. A spreadsheet that stays editable as a safety net stays in use, the two records diverge within weeks, and the winery ends up maintaining both with no clarity about which is authoritative, which is strictly worse than either alone.
Step two is the one most often skipped, and skipping it is how migrations fail immediately. If three people mean three different things by "lot," no software will reconcile them, and the disagreement will surface as data that looks wrong to everyone for reasons nobody can articulate.
Frequently asked questions
Are spreadsheets acceptable for TTB winery records?
Yes in principle. Federal wine regulations prescribe what information a record must contain, not what software it lives in, and records kept in ordinary business format are acceptable. The practical problem is not legality, it is that a spreadsheet has no audit trail, so a cell edited last March is indistinguishable from one entered correctly, and that weakens the record as evidence.
When does a winery outgrow spreadsheets?
When lot count exceeds what one person can hold in their head, usually well before case count suggests. The trigger is rarely volume. It is a second person entering data, a blend that makes lot history ambiguous, or a compliance question nobody can answer without an afternoon of reconstruction.
What breaks first in a winery spreadsheet?
Lot lineage through blending. A blend of three lots into one is trivial to write in a cell and nearly impossible to trace backward six months later, because the spreadsheet records the result rather than the relationship. Volume reconciliation and additions history fail next, for the same reason: the events were summarized instead of recorded.
When should a winery migrate to new software?
At a vintage boundary, in the window between the end of one harvest cycle and the start of the next. Migrating mid-vintage means running two systems during the weeks when data capture matters most and attention is scarcest, which is where most difficult migrations come from.
How much historical winery data should you migrate?
Complete data for wine still in inventory or in bond, and summary data for anything already sold. Wines still in your cellar need full history because you will keep making decisions about them. Wines already sold need enough to satisfy the retention obligation, which the archived spreadsheets themselves can often serve, provided they remain readable.
What data does a winery need to migrate?
Lots with current volume and vessel, vintage and variety, lab history, additions history, barrel inventory with fill dates, finished goods inventory by SKU, and any accumulated cost per lot. Customer and club data migrates separately if the sales side is moving too. Fruit source and contract records matter for appellation and vintage label claims.
Should you run a spreadsheet and new software in parallel?
Briefly and deliberately, or not at all. A short parallel period to build confidence is reasonable; an open ended one guarantees the two diverge and nobody knows which is authoritative. Set a date on which the spreadsheet becomes read only and hold it, because a fallback that stays available is a fallback people keep using.
The spreadsheet is not the problem, the missing relationships are
Wineries do not leave spreadsheets because someone told them spreadsheets are unprofessional. They leave when a question they need answered turns out to be unanswerable: what actually went into that blend, why the cellar total and the book total disagree by 90 gallons, what the reserve program really costs per case. Those answers were never in the workbook, because a grid stores values and winemaking is made of relationships between them. The migration is worth doing when the questions start mattering more than the convenience.
Solera's Data Migration Hub is built for that specific starting point, importing spreadsheet and CSV exports with assisted column mapping so the workbook you already keep becomes the opening inventory, with lot lineage, additions, and cost carried forward from there rather than started from scratch.