Switching costing tools without losing your numbers

Sooner or later the tool you keep your costs in stops being the right one. It gets retired, or repriced, or it simply never understood that you make what you sell. Whatever the reason, the move itself is the same move, and the thing at risk is always the same thing: not your stock count, which you can redo in an afternoon, but your cost history, which is the only record of what each material actually cost you over months of buying it.

This is that move, written for any tool you are leaving. It is five jobs in a fixed order, and the order is what keeps a half-finished migration from being worse than no migration at all.

Coming from Stocky? Shopify retires Stocky on August 31, 2026, and there is a version of this page written around exactly that move: the Stocky migration checklist. Start there if Stocky is the tool you are leaving.

The five jobs, and why the order is fixed

1. Freeze. Stop creating new records in the old tool, so your final export is a complete one.
2. Export. Pull the files, priority first, and check each one is real before you move on.
3. Copy out by hand. Capture the records that have no export path at all.
4. Rebuild. Enter things in the order the new tool computes them in, not the order you remember them in.
5. Reconcile. Prove the numbers landed, against a real invoice, before you price anything from them.

People reverse jobs two and four all the time: they start typing materials into the shiny new tool while the old one is still running, then export later and find the two disagree. Now they have two partial truths and no way to tell which is right. Export first. The file is the fact.

Job 1 — Freeze, about two weeks out

Pick a date roughly two weeks before you intend to switch and stop opening new purchase orders, new counts, or new products in the old tool from that day. Record anything you buy after the freeze somewhere you control: a plain spreadsheet with date, vendor, item, quantity, and unit price is enough. Those few rows go into the new tool by hand later, and they are far easier to handle than an export that was taken while the underlying data was still moving.

Put the export session in your calendar, and put it before the final week. If the tool is shutting down rather than merely being left, the last week is when export buttons get slow and support queues get long.

Job 2 — Export, in priority order

Take these in this order, because if you get interrupted after the first one you have still saved the part that cannot be reconstructed from anywhere else.

1. Purchase history: one row per purchase line. Date, vendor, item, quantity, unit cost, line total. This is your cost history and it is the priority file. From it, a new tool can rebuild not only what a material costs but how it got there.
2. Stocktake or count history. Less irreplaceable, and the fastest way to seed accurate starting quantities wherever you land.
3. Any per-item cost or valuation report. Between this and the purchase history you want a clear answer to "what did each material cost me, over time" sitting in a file you own.
4. Your product and variant list. This is the bridge between old item names and the material names you will use next, which turns matching rows later into a lookup instead of a memory test.

Ask for CSV wherever you are offered a choice. If the only option is a spreadsheet file, take it and save a CSV copy yourself, in UTF-8, because that is the conversion step where accented supplier names get mangled if you skip it.

The click-by-click detail differs per tool and we are deliberately not inventing it. What does not differ is the two-minute check you run on every file the moment it lands. The tool-agnostic export guide walks that check and the four data shapes in full.

Job 3 — Copy out what will never export

Every tool has a corner it does not let you take with you, and you find out which corner on the day you can no longer get in. Open a fresh spreadsheet, while the old screens still load, and copy across:

• Your supplier list. Vendor name, contact, website or login, what you usually buy there, typical order size, any price break you have earned, and rough lead time. This is the single most commonly unexportable thing in the category, and rebuilding it from memory is a month of small annoyances.
• Purchase orders still in flight at the freeze, with what you expect to receive and pay.
• Notes, attachments and photos hanging off records. These almost never come out through an export and they are usually the reason a strange price made sense.
• Anything you looked at every week that is in none of your export files. That gap is the whole list for this job.

Job 4 — Rebuild in the order the tool computes

A costing tool computes downstream from materials, so enter them upstream first. Do it in this order and each step has what it needs:

1. Materials and their costs. Everything else is derived from these, so one wrong material cost quietly poisons every product that uses it.
2. Decide the unit for each material once, and never mix. Grams or ounces, each or case, but one per material, forever. Unit drift is the most common way an imported cost history ends up wrong by a factor of twelve or sixteen.
3. Products and their recipes. The wax, the fragrance, the jar, the wick, the label. If your old tool was a retail stock tool, it never held this at all, so there is nothing to bring and nothing lost. You build it once and the cost maintains itself from then on.
4. On-hand counts. A fresh physical count on the day you switch beats an imported number from three months ago.
5. Vendors, from the spreadsheet you copied out in job three.
6. Then a week of ordinary purchases, entered as they happen, before you trust the numbers for a repricing decision.

The file-level mechanics of step one, the columns that decide whether your numbers survive and the five breakages that quietly inflate every product cost, are in the import guide. That running order is also, not coincidentally, the order any costing tool wants a brand-new account set up in: getting started is the same six steps for someone arriving with nothing to migrate.

Job 5 — Reconcile, and prove it landed

This is the job people skip, and it is the one that decides whether you are running a business on real numbers or on a confident-looking screen. All of it together is about twenty minutes.

1. Count rows. Distinct items in your export against materials created. A gap means something was skipped or silently merged.
2. Reconcile one whole invoice. Add up its line totals and find the same total in the new tool. This one check catches currency, decimal and quantity errors at once.
3. Recompute one finished product by hand. Add its parts up on paper. When paper and screen disagree, it is nearly always a unit.
4. Read both extremes. Sort materials by unit cost. A wick that costs more than a jar is a decimal or pack-size error announcing itself.
5. Hunt zero and blank costs. Every one of them makes some product look more profitable than it is.
6. Check the date span. Earliest and latest purchase dates should match your file. A short span means part of the history did not arrive.
7. Run the leave-test. Export from the new tool on day one and open the result. If your data cannot come back out, you have traded one dead end for another.

What this honestly costs you

Budget an hour for the exports, an evening for the rebuild, and a week of ordinary use before you trust a margin. Those are three separate sittings and only the first one has any urgency attached to it. Nobody finishes a tool migration in one evening, and a plan that assumes otherwise is how the reconcile job gets skipped.

The part that is genuinely unrecoverable is small: cost history you never exported, records with no export path that you never copied out, and anything still sitting only in a tool you can no longer open. Everything else is inconvenience.

Where Batchnook fits, and where it does not

Plainly, and with nothing oversold:

• A generic column mapper. Paste your CSV or upload the file, then match your columns to item, unit, quantity and cost. There is no per-tool preset for anything, so expect an honest mapping step per file rather than one button.
• A dry run before anything is written. It tells you how many materials and products will land and flags conflicts, and nothing saves until you approve it.
• Moving-average cost from then on. Every new purchase re-averages the material, so a product's cost reflects what you have been paying lately, and the method is disclosed on every report. Snap a supplier receipt and the scan reads the lines; you approve them before anything counts.
• A 24-hour undo, so a bad mapping is a reversible afternoon rather than a lost weekend.
• Export free on every tier, forever. Materials, products, batches, orders and cost history, as JSON or CSV. You are reading this because a tool did not work out. It should never be hard to leave the next one either.

And who should not move to us:

• Retailers who buy finished stock and mainly need counts synced across a point of sale and an online store. A retail inventory tool fits you better.
• Makers whose hardest problem is labeling or food compliance. That is a specialist's job and we deliberately do not build it. Rules differ by state, so check your own state's requirements, and see labeling and regulations for the honest starting shape of that question.
• Anyone happy on a spreadsheet. If a sheet is working and updating it does not hurt, keep it.

The full set of questions to ask any candidate, including us, is in what to look for in a maker costing tool.

FAQ

Do the exports first, check every file the moment it downloads, and leave the reconcile job on the calendar rather than to memory. The rest of this cluster covers each half in detail: getting the data out, and getting it in without breaking it.