Your team has 200 company accounts. Each has three company locations. Each location is tied to a custom price list, each price list is tied to a catalog assignment, and the whole thing took 18 months to build, negotiate, and sign off on.
Then on a Tuesday afternoon, someone saves a CSV import with the wrong column mapping. Six months of pricing logic flattens. Quantity rules misfire against products they were never scoped to. Three of your top accounts log in and see no prices.
You do not need a refresher on what a catalog is. You need to know what happens next, because a store-level rollback will not give you back the 40 buyer relationships that just broke while leaving the other 160 alone.
The shape of the problem
You already know the shape of your B2B operation. What matters for this piece is that every relationship in it is a node with edges. On Shopify Plus, B2B catalogs are unlimited and assignable down to the specific company location, with companies, company accounts, company locations, price lists, contracts, quantity rules, and payment terms all carrying their own state. Lower Shopify tiers cap at three active catalogs.
The edges are where contract state lives. Which catalog governs which location. Which price list applies to which account on which catalog. Which quantity rule fires for which SKU under which contract. When a bulk operation, an import, or an agent workflow writes to the wrong edge, the blast is not one record. It is every B2B customer whose purchasing experience depends on that edge.
How the break happens on a real Tuesday
Three patterns show up in the support queue again and again.
A buyer runs a bulk update to raise prices 3% on one product family. The operation targets the wrong column. Every price list entry for that family writes a null. B2B customers hit the storefront and see no prices, only a request-a-quote fallback that nobody on the sales team is staffed to handle at volume.
A scheduled job is supposed to move Catalog A onto 50 newly onboarded company locations. A filter misfires. Catalog A lands on 500 locations instead. Half the book of business is suddenly priced as if they were net-new prospects, not contract customers with negotiated terms.
An automation writes a quantity rule update meant for one product line. A condition evaluates too broadly. The rule applies to every product. Every open B2B cart is now blocked by a minimum the customer cannot meet.
The common factor is not malice. Verizon’s 2024 Data Breach Investigations Report found that 68% of breaches involve a non-malicious human element, a person making an error or falling for social engineering. Your operation is exactly where an honest mistake gets multiplied by 200 accounts in under a minute.
Why store-level rollback is the wrong tool
Shopify provides platform-level rollback capability for some store settings, and that is real. The problem is scope. A store-level rollback treats the store as a single unit. Roll the store back to yesterday and you have also rolled back every legitimate change since yesterday, including the onboardings, the new price lists, the contract updates your team actually wants to keep.
A B2B graph does not want a snapshot of the whole store. It wants one catalog-to-location assignment reset to the state it was in at 2:47pm Tuesday while leaving every other assignment, price list, and contract untouched. A rollback that cannot express “this node, this edge, this timestamp” forces you to choose between losing recent legitimate work or eating the damage of the incident.
Where Rewind fits
Rewind is a SaaS resilience platform built on independent architecture, a platform (not a plugin) that ensures data is accessible even if the SaaS vendor is compromised. Rewind is a Shopify Certified Technology Partner, and more than 25,000 organizations trust Rewind today.
For B2B specifically, the capability that matters is item-level, granular restore, restoring individual items (a single ticket, page, file, or configuration) without affecting the rest of the system. Non-destructive recovery.
Translated into B2B vocabulary you already speak:
- A single company’s catalog assignment brought back to its pre-incident state while the other 199 sit untouched.
- A single price list’s entries restored to the values they held before the CSV import, with no change to the other price lists updated legitimately in the same window.
- A single quantity rule’s configuration reset to its pre-automation state, leaving the rules that are working as intended in place.
- A single company account’s permissions restored without disturbing the other accounts inside that company.
Each is a node-level operation. None force you to roll back legitimate work to recover illegitimate work.
A worked example
A Plus merchant runs a bulk edit intended to move 12 company accounts from net-30 to net-45. The edit runs correctly on those 12. A condition also misfires against 40 accounts it should not have touched. 52 accounts now show net-45, instead of the mix of net-30 and net-60 they are contractually entitled to.
The store-level response is to manually identify the 40 incorrect accounts and reset their terms by hand, while a billing exec writes apology emails to every finance contact affected. It works, but it is slow, visible, and error-prone under pressure.
The Rewind response is an item-level restore on the 40 incorrect accounts to their pre-edit payment term state. The 12 correct accounts stay exactly as they are.
AI, automations, and the new write surface
Rewind does not have AI products. Rewind’s backup and restore product protects your SaaS data from errors introduced by your own AI agents, automations, and AI-assisted workflows.
For B2B merchants, the AI conversation is less about what AI can do and more about what AI is allowed to write to. A pricing agent that recalculates tier breakpoints. An ops automation that updates payment terms on renewal. A content workflow rewriting metafield descriptions across a catalog. Each is a write surface. Each can misfire. The matching response is a restore surface at the same resolution, node-level and edge-level, not “the whole store as of last night.”
A 30-day readiness checkpoint
- Days 1 to 10: inventory your write surfaces. Not the graph itself, since you already know that. The automations, scheduled jobs, AI agents, bulk ops, and CSV import paths that can write to it, and who owns each one.
- Days 11 to 20: tabletop the three scenarios above against your real operational patterns. The one that makes your ops lead flinch is your priority.
- Days 21 to 30: confirm restore paths at the graph-node level. Run a test restore in a sandbox so the first real restore is not your first.
B2B buyers expect their pricing, terms, and contracts to hold steady. A contract graph that cannot be restored at the node level cannot make that promise.
Learn more about Rewind for Shopify at rewind.com/shopify.
Rewind">
