Stocky ends in four
weeks. Most brands
do not need an ERP.

Stocky stops working on 31 August 2026. If you are one of the merchants who has been running purchase orders and reorder forecasting through it, you have about four weeks to decide what replaces it, and the honest answer for most brands is not an ERP.

The trigger is real and the panic is mostly not. Shopify has moved the transactional half of what Stocky did into the admin. What it has not replaced is the thinking half: forecasting, reorder points, landed cost. That gap is what sends founders to a demo call with a system that costs six figures to implement, usually a year or two before they need one.

Here is what actually goes away, the four triggers that mean you genuinely need a system of record, and what each tier costs.

Nothing migrates
automatically, and
suppliers cannot be
exported at all.

Shopify's own migration documentation is unusually direct about this, and it is worth reading before you read any vendor's landing page about it. The dates and the data situation, verified 31 July 2026:

Three specifics that decide how much work the next month is:

Nothing moves by itself. Shopify states plainly that historical data such as old purchase orders and stocktakes will not automatically move into Shopify. The native CSV upload only adds product line items to a new draft purchase order, and cannot import past statuses, received quantities or supplier links.

Suppliers cannot be exported. This is stated twice in the migration doc and it is the item that surprises people. There is no export path for supplier records, so they get rekeyed by hand. If you have a few hundred suppliers, start that now rather than in the last week of August.

It was bundled with POS Pro. Merchants who installed after May 2020 got Stocky as part of a POS Pro subscription, published at $89 a month per location. That component is being removed with no announced price change, which is the loudest theme in Shopify's own community thread on the subject, and a fair complaint. A five-location retailer pays $5,340 a year and is losing a named part of what they bought.

What Shopify has kept and improved is the transactional layer: purchase orders with suppliers and payment terms, transfers with split shipments, adjustments with reasons, bin locations, unlimited inventory history. That covers more merchants than the community reaction suggests. What it does not cover is listed in Shopify's own gap notes: no native demand forecasting engine, no reorder points or min-max, a single static cost per item with no weighted-average or landed cost, no way to email a purchase order from the admin, and no purchase order APIs. Its answer for forecasting is to ask Sidekick what you should reorder, which is prompting rather than a forecasting system.

You need a system of
record when inventory
truth lives in more
than one place.

The question I get asked is what revenue level requires an ERP, and it is the wrong question. I have seen $40M brands run competently on a spreadsheet and a good 3PL portal, and $8M brands genuinely drowning. Revenue correlates with the real triggers without causing them.

Trigger one: inventory truth lives in more than one system and they disagree. The moment you have stock in a 3PL, stock in a retail location, stock in transit and stock at Amazon, and no single system reconciles them, you are doing manual reconciliation forever. This is the trigger, and it usually arrives with the second sales channel rather than with a revenue milestone.

Trigger two: you cannot answer what a unit actually cost. Shopify holds one static cost per variant. If your landed cost varies by shipment because of freight, duty and tariffs, then your margin reporting is an estimate and your reorder decisions are guesses. Brands importing across multiple POs at different freight rates hit this early, and it has become sharper as tariff exposure has moved around. The margin consequences are the subject of how tariffs hit unit economics.

Trigger three: the reorder decision has outgrown the person making it. One person with a good spreadsheet can forecast maybe a few hundred SKUs across one or two channels. Beyond that, or once lead times vary meaningfully by supplier, the spreadsheet stops being a forecast and starts being a record of what someone guessed. Symptoms are stockouts on your best sellers alongside a warehouse full of slow movers.

Trigger four: close takes longer than a week because of inventory. If finance cannot close the month because nobody can state the closing inventory value with confidence, that is a system-of-record problem and it will not resolve with more discipline. It is also the trigger that tends to convince a board, because it shows up as an audit finding rather than an operational complaint.

None of those is an ERP trigger specifically. Triggers one and four point toward an inventory or order management system. Triggers two and three often resolve with a forecasting or costing tool sitting on top of Shopify. A full ERP earns its place when you also have manufacturing, multi-entity accounting, or genuine multi-currency consolidation. If your inventory questions are really cash questions, the inventory cash flow model is a cheaper place to start.

Seven of the thirteen
options will not tell
you the price.

That is the most useful finding from pricing this category out, and it tells you where the tier boundaries actually are. Anything with published pricing is self-serve software. Anything quote-only is a sales-led implementation, and the quote is the smaller half of the cost.

Two cautions on that table. Katana's $299 is a floor rather than a price: manufacturing, traceability and warehouse management are separately priced modules, so a brand wanting lot tracking and bins is realistically several times the headline before any volume scaling. And its metric has shifted over time toward order counts and GMV, which penalises high-order-count, low-average-order-value sellers. That is precisely the DTC profile.

The second caution is sharper. Cogsy, which still appears in current best-of listicles and still displays a price, appears to be dead. Its marketing site is up, but the application itself returns a server error on both login and registration, its help subdomain is unreachable, and its App Store listing has no review since March 2023. There was no shutdown announcement, so I can only report the observable state rather than the cause. Do not shortlist it without checking that you can actually create an account.

The licence is the
cheap part and everyone
knows it except the
person signing.

Good independent data on ERP implementation exists, but almost none of it is segmented to ecommerce, so read the following with its sample in mind.

Panorama Consulting's 2026 ERP Report, based on 170 responses gathered between January 2025 and January 2026, found 30.0% of projects came in over budget and 22.3% over schedule, with a median timeline of nine months. Their 2025 edition put median total project cost at $450,000. The caveat that matters: the median respondent in that survey had roughly $200M of annual revenue in the 2026 edition and roughly $400M in 2025. If you are a $15M DTC brand, that median is not your number, and the gap is large enough that you should treat those figures as directional evidence about how these projects behave rather than as a budget.

The finding I would actually hold onto is from Third Stage Consulting's 2023 research, which reported that roughly half the companies studied experienced a material go-live disruption, defined as being unable to ship product or close the books, and that those disruptions increased implementation cost by fifty to three hundred percent. Their methodology disclosure is thin, so weight it accordingly, but the shape matches everything I have watched happen. The risk in these projects is not the licence fee. It is the quarter where you cannot ship.

That risk is what makes sequencing more important than selection, which is the part vendors have no incentive to tell you.

Export first. Decide
second. Implement
in January.

This week, export everything. Completed purchase order reports to CSV, stocktake history, historical cost data. Rekey your supplier list into a spreadsheet, because there is no export for it. Do this before you touch anything else, because the read-only window closes around the end of November and uninstalling early is irreversible.

Two weeks before the deadline, stop creating new purchase orders in Stocky. Receive and close everything in transit, then recreate only the remaining open quantities in Shopify. Shopify's own guidance suggests roughly a fourteen day buffer, and that matches what I would do.

Then run native for a quarter before you buy anything. This is the advice most likely to be ignored and most likely to save money. The native transactional layer covers more than the community reaction suggests, and a quarter of running it tells you precisely which of the four triggers you actually have. Buying in September, under deadline pressure, from a vendor running a migration campaign, is how brands end up with a system three tiers above what they needed.

If you must buy now, buy the layer rather than the system. A forecasting tool sitting on top of Shopify solves triggers two and three for a few hundred dollars a month and can be removed later. An OMS or ERP is a twelve month commitment with an implementation attached. The layer is reversible and the system is not, and under time pressure you should always prefer the reversible option.

Do not implement anything significant in Q4. Going live with a new system of record between October and January is how you get the disruption quarter. If the assessment says you genuinely need an OMS, scope it now, sign in November, implement in January. The reasoning is the same as everything in the peak season operating playbook: nothing structural changes while the money is being made.

The summary I would give a founder in a hurry: Stocky going away is a data export problem this month and a systems decision next quarter, and conflating the two is the expensive mistake. Export now, run native through the autumn, and let the four triggers tell you whether you need a layer, a system, or nothing at all.

The operational half of this, forecasting method and cash conversion, sits in inventory management for Shopify brands. Where these systems fit against everything else you are paying for is in the tech stack by revenue stage and the financial stack by stage.