Stocky Alternative for Shopify: Purchase Orders and AI Reordering

Stocky stopped on 31 August 2026. Its job now sits in Shopify admin: purchase orders live there, and Shopify points merchants to Sidekick, its AI assistant, for reorder ideas. Old Stocky history cannot be imported, so rebuild open orders and suppliers by hand and keep past data as a spreadsheet.

Do the export first. Shopify's transition page promises read-only access for exporting Stocky data for at least 90 days after 31 August, which runs to about the end of November 2026, and suppliers cannot be exported at all, so copy them by hand. Then treat AI reorder ideas with care: Sidekick or any assistant is only as good as your stock counts and supplier lead times, and a small backtest on an independent bookshop with roughly 1,200 products shows how to check them.

Follow me on Instagram@sagnikteaches

What stopped, what replaced it, and what did not come across

Shopify's transition page says Stocky has not been available since 31 August 2026 and that its APIs stopped the same day. Facts I could confirm from Shopify's pages: purchase orders are available in Shopify admin, where you can save a draft, mark an order as Ordered once the supplier confirms, create an inventory transfer to receive stock, and keep suppliers with contact details, payment terms and currencies. Stocktakes and cycle counts move to stock counts in admin. For reordering, the same page says Sidekick recommends items based on sales data and drafts the purchase order or transfer for you. Sidekick is included with your Shopify plan, although Shopify says features and usage limits vary by plan, and it can write and run its own code to forecast, compare and summarise store data.

Connect on LinkedInSagnik Bhattacharya

What I could not confirm from the pages I read: whether every purchase-order feature is available on every plan, and exactly what Sidekick returns when you ask about reordering on a given store. Check both in your own admin before you plan around them. And one fact is firm: Shopify says historical purchase orders cannot be imported, nor their statuses, received quantities or supplier links.

Subscribe on YouTube@sagnikteaches

So the migration is a rebuild, not a transfer. That sounds worse than it is: most small shops only have a handful of open purchase orders and 10 to 30 active suppliers at any time.

Before you rebuild: take what you can from Stocky

1. Export open purchase orders. Use Stocky's built-in reports while the read-only window lasts and save a list of every order not yet fully received: supplier, products, quantities ordered and received, cost, expected date.
2. Copy the supplier list by hand. Shopify says suppliers cannot be exported from Stocky, so type or screenshot name, contact, payment terms, lead time and minimum order value. If you never recorded lead time in Stocky, write it from memory now: how many days from order to delivery, typically.
3. Keep a closed-order archive. Shopify suggests exporting completed purchase order reports as CSV files, plus anything you need for tax or audit records. Twelve months of purchase orders is enough for forecasting later. You will not rebuild old orders in admin.
4. Take a stock count. The old numbers may have drifted, and admin's stock counts now handle full counts and cycle counts. The tutorial on running a stocktake faster with a phone scanner shows how to do it without closing the shop.

Rebuilding purchase orders in Shopify admin

The route, based on Shopify's purchase-order documentation:

1. Create each supplier with contact, payment terms and currency.
2. For every open order, create a purchase order and save it as a draft while you check quantities. A long order can be filled by bulk-importing products from a CSV file; a row that repeats a variant already on the order fails, so edit that quantity on the order instead.
3. Mark it Ordered when the supplier has confirmed. Shopify describes that as the point when you can create an inventory transfer to track the shipment.
4. When goods arrive, receive them against the transfer, adding any cost adjustments, and note any short deliveries.
5. Track completion on the inventory transfer, not the purchase order: Shopify's help centre says receiving, partial receipt, completion and closure all live on the linked transfer. Archive the purchase order once you no longer need it on screen; archiving does not close or cancel the transfer.

A bookshop, illustrative: an open order of 120 books across 18 titles was part-delivered, with 80 received. Rebuild it as a purchase order for the 40 still outstanding rather than the original 120, with a note "balance of order placed 12 October". Otherwise admin will show 120 on order and your reorder checks will think you have more coming than you do.

Mistakes to avoid in the rebuild

Double counting incoming stock. If a part-delivered order is rebuilt in full, incoming stock is overstated and any reorder suggestion will say "no need". Check each rebuilt order against what has physically arrived.

Losing supplier lead times. Admin holds payment terms and currency, but a reorder check needs a lead time. Keep lead times in a column of your own sheet, per supplier, because that number drives every reorder point. For a wider look at your options if admin alone is not enough, see best AI inventory tools for small retailers compared.

The reorder arithmetic that AI should reproduce

Before you test any AI suggestion, you need a rule it can be checked against. The classic one is a reorder point: average daily sales multiplied by supplier lead time in days, plus a safety stock for surprises. Order up to a target that covers a chosen number of weeks.

Take one title, illustrative: a bestselling paperback sells 21 copies in 7 weeks, so 3 a week or about 0.43 a day. The supplier delivers in 10 days. Lead-time demand is 0.43 x 10 = 4.3 copies. Add a safety stock of 3 and the reorder point is 7.3, so you order when stock hits 8. To cover four weeks you want about 12 copies on hand after delivery, so the order quantity is 12 minus what will be left when the delivery lands. You order at 8 and sell about 4.3 while you wait, leaving roughly 4, so the order is about 8 copies.

The seasonal cookbook is the one a simple rule gets wrong, and the one a sensible assistant should flag. Keep that in mind for the test below. For a plain-English introduction to the idea, how small shops can use AI to forecast stock and reorder goes through the basics.

Testing AI reorder suggestions before you trust them

You cannot judge a suggestion tool from one good answer. Run a backtest: ask it to suggest what to reorder as of a date in the past, then compare with what actually sold.

1. Pick a past date, say eight weeks ago, and export sales by product up to that date and the stock levels on it.
2. Ask the assistant for reorder suggestions as of that date, giving it lead times.
3. Compare with the next eight weeks of sales. Which suggested items ran out? Which suggestions were never needed? Which sellers did it miss?
4. Score three numbers: stock-outs it would have caught, money it would have tied up in surplus, and items it missed.

Below are 12 weeks of unit sales by product and today's stock on
hand, with each supplier's lead time in days. Using only this data,
suggest which products to reorder and how many, targeting four weeks
of cover after delivery. For each, show: weekly average, lead-time
demand, safety stock used, and the order quantity. Flag any product
whose sales look seasonal or whose data is too thin to judge.

Sample output, illustrative: "Paperback bestseller: average 3 a week, lead-time demand 4.3, safety stock 3, reorder now, order 8. Seasonal cookbook: sales jump from 0.5 to 4 a week in the last three weeks, so an average understates demand; flagged, suggest checking last year's November sales before ordering." What you fix in it: ask for the working every time, spot-check three calculations by hand, and overrule any number that ignores a promotion or a supplier minimum. The assistant cannot see a pending author event in your diary.

What Sidekick adds, and what to test in it

Shopify describes Sidekick as an assistant that can write and run its own code to forecast, compare or summarise your store data, build Flow workflows for repeated tasks, and present changes for your review before applying them. Its Stocky transition page goes further: Sidekick recommends items based on sales data, drafts the purchase order or transfer, and can create purchase orders for you. Shopify also notes that in some stores Sidekick cannot run code, which limits the forecasting part. Run the same backtest through it by asking in plain words what to reorder and why, then compare its answer with your own arithmetic. Ask three follow-ups: how it handled seasonality, what lead time it assumed, and whether it counted incoming stock on open purchase orders. If it cannot tell you the last one, its suggestions are unsafe for a part-delivered order. The tutorial on using Shopify Sidekick to run your store faster covers its other jobs.

Cut-over day: a checklist for the switch

Pick a quiet day, ideally just after a delivery has been received and before the next order goes out, so open orders are few. Work through this list in order and tick each line as you go.

• Every supplier created in admin with contact, payment terms and currency, and a lead-time column kept in your own sheet.
• Every open Stocky order rebuilt for the outstanding balance only, with the original order date in the note.
• Stock counts matched to the shelf for your top 50 sellers, and a plan to count the rest over the next month.
• A closed-order archive saved as a spreadsheet, with 12 months of supplier, product, quantity and unit cost.
• Each reorder point written down per product or per group of products, with the date you last reviewed it.
• One person named as the buyer, and one as the person who receives stock, so the two jobs do not blur.

The last item is the one small shops skip. When the same person orders and receives, mistakes slip through unnoticed; when two people do, a short delivery is spotted at the door.

A supplier sheet worth keeping next to admin

Admin gives you the supplier record, but the numbers that drive reordering sit better in a sheet you control. Here is a filled-in sheet for three bookshop suppliers, illustrative:

This sheet is what you give your assistant when you ask for reorder suggestions. Without lead times and minimums, it will suggest an order the supplier will not accept or a delivery that arrives after the sell-out.

A backtest with numbers

A finished backtest looks something like this imagined run on 20 titles. You ask the assistant, as of eight weeks ago, what to reorder. It names 9 titles. Over the next eight weeks, 6 of those 9 sold out or fell to zero; 3 would have been unnecessary orders, tying up roughly $160 of stock that sold slowly. It missed 2 titles that did sell out, both of which had a sudden jump in demand after a local event.

Score it: 6 catches, 3 false alarms, 2 misses. That is a reasonable assistant for a weekly check, as long as someone adds the knowledge it cannot have, such as events. If it had named 9 titles and only 2 had sold out, with 5 misses, you would not use it for ordering at all. Repeat with a second date, because a single window can flatter or punish. If the results are similar on both dates, you have a basis for trust; if they swing wildly, the data is too thin or too seasonal for suggestions.

Setting the safety stock from your own misses

The misses from the backtest tell you where safety stock is too low. For each missed title, count how many copies short you were at the worst point, and add that to the safety stock for similar titles. Do it by group, for example "event-driven local history titles: add 3", rather than for 1,200 products one by one. It is a rule you can keep in your head and give to the assistant in the next prompt.

Before and after: a stock-out that a routine would have caught

Before. In Stocky, a children's picture book had a reorder point of 5. A school visit boosted sales to 9 a week, and nobody was watching. The title hit zero on a Wednesday, the order went in on Friday, and the 5-day supplier lead time meant two weeks without stock. About 12 sales were lost.

After. With the Monday routine, the export on the following Monday shows the title at 3 against a reorder point of 7. The assistant flags that sales over the last two weeks average 9 a week against 5 the month before, works out that four weeks of cover at the new rate would be about 36 rather than 20, and asks whether a one-off event explains it. The buyer confirms the spike came from a school visit, with a second visit booked next month, and settles on ordering up to 24 as a middle course. The sum: at a $4 margin, 12 lost sales is $48, which is small per title, but three or four such titles a quarter add up, and the customers who found another shop are worth more than the margin.

Where the AI should never decide alone

Several parts of purchasing should stay human, however good the suggestions look. Pricing negotiations and supplier minimums need judgement about the relationship. Anything that commits money, such as marking an order as Ordered or paying a deposit, needs a person's name against it. Products with an expiry date, a safety regulation or a recall risk need a person who knows the rules. And new products with no sales history have nothing for a forecast to work on, so an AI suggestion for them is a guess with decimal places.

Think of the assistant as a fast analyst who has never met your customers. It will list what the numbers say; you decide what the numbers are missing. That split also keeps you safe from the worst failure, a confident order of 200 copies of something that sold three, so give it a standing instruction to flag any suggested order above, say, twice the previous largest order for human review.

A weekly ten-minute reorder routine

1. Monday morning, export products with stock at or below their reorder point, including incoming quantities from open purchase orders.
2. Paste the list into your assistant with last month's sales and ask for a draft order per supplier. Check the arithmetic on the three largest lines.
3. Create draft purchase orders in admin and review each against supplier minimums and payment terms.
4. Mark Ordered only after the supplier confirms, and put the expected date in the note.
5. Friday: receive what has arrived, and chase anything past its expected date.

The time cost is about ten minutes for the list and one supplier call, much less than the hours lost to a bestseller that sold out for two weeks. A rough sum: if one popular title sells three copies a week at a $9 margin and is out of stock for two weeks, that is $54 of lost margin from one title, before counting the customer who goes elsewhere.

When admin is enough and when a third-party app is worth it

For most small shops with a few suppliers and under a few thousand products, admin purchase orders plus a weekly routine covers it. Look at a third-party inventory app when you have many suppliers with different currencies and terms, when several people place orders and need approvals, when you sell on more than one channel and need synced stock, or when you want automated reorder points per variant. Check each app's own page for pricing and whether it imports purchase orders, and ask whether it replaces any of the routine above. The tutorial on syncing stock across Shopify, Amazon and eBay covers the multi-channel case.

Wherever you land, record the decision: what you use, who orders, and where lead times live. Stocky disappeared with its history, and that is the lesson: keep your own copy of the numbers you depend on.

Questions from Shopify stores that used Stocky

Can I import my old Stocky purchase orders?

No. Shopify's transition page says historical purchase orders, their statuses, received quantities and supplier links cannot be imported, and suppliers cannot be exported from Stocky at all. Shopify promises read-only access for exporting reports for at least 90 days after 31 August 2026, so save the CSVs now and rebuild open orders and suppliers by hand.

Do I need a paid app to replace Stocky?

Not necessarily. Shopify admin now holds purchase orders with suppliers, draft and ordered states and receiving. A third-party inventory app is worth considering when you need forecasting, multi-supplier automation or reports that admin does not give. Check any app's import options and pricing on its own page.

Can Sidekick place orders with my suppliers?

Shopify says Sidekick recommends items from sales data and drafts the purchase order or transfer for you, and that it presents changes for your review before applying them. Treat those drafts as an adviser's work, not a buyer's. Anything that commits money, such as marking an order as Ordered or sending it to a supplier, should stay a human decision.

How often should I review reorder suggestions?

Weekly for fast sellers and monthly for the long tail. Suggestions go stale quickly after a promotion or a supplier delay, so rerun them before each ordering day and compare against what you actually sold, not what was predicted.

Further reads

• How to Use AI to Set Reorder Points and Prevent Stockouts — Set reorder points per product before trusting suggestions.
• Why AI Stock Forecasts Go Wrong: 8 Mistakes Small Businesses Make — Eight ways AI stock forecasts go wrong in small shops.
• Does AI Stock Ordering Pay Off for an Independent Pharmacy? — Does AI stock ordering pay off in another shop type?
• How Much Does It Cost to Add AI to a Shopify Store? — Add up the cost of AI tools on a Shopify shop.
• Shopify Inbox vs Tidio Lyro: Which AI Chat Tool Works Better? — Another AI choice for a small Shopify store.
• Best AI Chatbots for Small Shopify Stores (2026) — More chat options for a small Shopify shop.
• Boutique Inventory Management: AI for Size Runs and Reorders — Size-run reorder rules, a worked AI forecast for one dress in two colours, and a markdown ladder with margin floors for a small boutique.
• Supply Chain AI Software: What a Small Importer Needs First — A buying checklist for small importers covering data, lead times, supplier emails and stock sync, with one worked reorder to test any vendor against.
• How Will AI Affect Retail? The Jobs a Small Shop Hands Over First — A four-week walk-through of the first AI jobs in a small shop, with the set-up list, prompts, sample outputs and the mistakes to catch.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Shopify help centre pages on transitioning from Stocky, Stocky, purchase orders, creating purchase orders and Sidekick, and Shopify's Sidekick page, read 4 and 5 November 2026.