inventory & order sync
Products, prices and stock in from your ERP or POS; orders, refunds and customers out; fulfilments back. One stock figure across online and stores.
$ storesmith build --automation
StoreSmith automates the manual steps behind a Shopify Plus store: ERP, PIM and WMS sync, B2B price tiers, order routing, Flow and scheduled jobs. Integrations that retry, reconcile and tell you when something didn't land — instead of a spreadsheet someone fixes by hand.
01 — what we build
Each integration has an owner, a source of truth per field and a record of every sync. You know what moved, when, and what failed.
Products, prices and stock in from your ERP or POS; orders, refunds and customers out; fulfilments back. One stock figure across online and stores.
Product data, variants, media and metafields from your PIM, pushed with bulk operations so large catalogues update without throttling the store.
Companies and locations, catalogues and price lists per customer segment, payment terms and quantity rules — kept in step with the ERP that owns them.
Orders sent to the right location, warehouse or 3PL by your rules, with tracking and status written back to the order.
Flow workflows for tagging, holds, alerts and approvals — plus custom triggers and actions when Flow needs to reach a system it doesn't know.
Nightly reconciliations, price updates, feed exports and the reports your finance team rebuilds every week.
02 — typical architecture
Webhooks tell us something changed. A queue absorbs bursts. Workers apply the change through the Admin GraphQL API or your system's API, and a scheduled pass catches anything a webhook missed.
The same event applied twice changes nothing twice. Retries are safe, so failures can be retried automatically.
Every sync is logged with what was sent and what came back. Failures alert a person, not just a log file.
Price from the ERP, copy from the PIM, stock from the WMS. Written down at scoping so two systems never fight over a field.
03 — when it's right
Not every integration needs custom code. We'll tell you which tier your problem sits in before we quote.
| option | what it covers | choose it when |
|---|---|---|
| shopify flow | Rules that live inside Shopify: tags, holds, notifications. | The rule never leaves Shopify. Often no app is needed at all. |
| a connector | A maintained connector for your system. Our own connectors are listed below. | It covers your system and your rules. |
| a custom build | Several systems, non-standard data, B2B pricing logic. | Volumes or rules make an off-the-shelf sync fall over. This is where we spend most of our time. |
04 — proof
Work is listed by industry and pattern, with figures rounded. The integrations we kept rebuilding are now apps of our own.
wholesale distribution · b2b automation
Several legacy storefronts merged into one Shopify Plus store, 10k+ SKUs, automated wholesale price tiers by customer segment. Real-time inventory sync took fulfilment errors to zero.
Orders, customers, fulfilment, returns and delivery updates synced to Salesforce through your MuleSoft layer. One install for every store and brand.
see crm connect →Products, prices and stock in from your ERP or store POS; orders, customers, payments and refunds out; fulfilments back. Online and in-store on one stock figure.
details →05 — questions
If the rule lives inside Shopify (tags, holds, notifications), Shopify Flow often does it with no app at all. If a maintained connector covers your system and your rules, use it. A custom build is for several systems, non-standard data, B2B pricing logic or volumes where an off-the-shelf sync falls over.
Webhooks tell us something changed and a queue absorbs bursts. Workers apply each change through the Admin GraphQL API or your system’s API, and a scheduled pass reconciles against the source to catch anything a webhook missed. Every event is idempotent, so retries are safe, and every sync is logged with what was sent and what came back.
One owner per field, written down at scoping: for example price from the ERP, copy from the PIM and stock from the WMS. That way two systems never fight over a field.
Failures alert a person, not just a log file. Drift found by the scheduled reconciliation is flagged, not hidden, and failed events can be retried automatically because applying the same event twice changes nothing twice.
06 — start
List the systems and the steps. An engineer replies with what we'd automate, what we'd leave alone, how long and roughly what it costs.