Integration / Makro Middleware
Salesforce CRM stays the system of record for the customer. Shopify Plus becomes where the order gets placed. We build custom Salesforce CRM ⇄ Shopify Plus middleware: two-way sync, dedup and identity resolution, B2B Commerce, real-time.
Which integration this is
"Salesforce" means two different products in commerce. This page is about integrating Salesforce CRM behind your Shopify Plus storefront. It is not about Salesforce Commerce Cloud, which is a rival storefront, not a CRM. Almost no vendor page says which one it means. We do, up front.
What this page integrates
Sales Cloud, Service Cloud, and B2B Commerce on Core: the system that owns the customer, the account, the contract, and the pipeline. You keep Shopify Plus as the storefront and integrate Salesforce CRM behind it.
Sales / Service · B2B Commerce on Core · the CRM behind the storeWhat this page is not
SFCC, formerly Demandware, is a separate storefront platform that competes with Shopify. It is not a CRM. If a buyer is searching "Salesforce Shopify integration," they almost always mean CRM behind the storefront, not a replatform.
SFCC / Demandware · a storefront, not a CRMThe real problem
The work is the space between: accounts, contacts, contracts and pipeline on one side, B2B buyer self-serve on the other. Duplicate contacts, orders written twice, pricing that drifts, conflicting truth at month-end.
01 / system of record
Accounts, contacts, contracts, pipeline. The authoritative record of who the customer is and what they are owed.
02 / front door
Where B2B buyers place repeat orders without a phone call, a rep, or an email to their account manager.
03 / the gap
Duplicate contacts. Customers fragmented across emails and guest checkouts. Orders that say "fulfilled" here and "in progress" there.
We built a product around that exact gap. Salesforce plugs into it as a first-class, path-mapped adapter. NetSuite is the one already running in production.
Architecture · what syncs and which way
Accounts, contacts, products and prices are read out of Salesforce and pushed to Shopify. The storefront never writes to your CRM of record. Orders are the only thing that travels both ways: placed in Shopify, recorded in Salesforce, status returned.
System of record
read · never overwritten
The platform
map · reconcile · audit
The storefront
where buyers self-serve
PATCH /sobjects/{Object}/{extIdField}/{value}. Salesforce creates if absent, updates if present, so a re-delivered webhook is a no-op.handshakeThe console
Every sync, every failure, every reconciliation is visible. This is the same Makro Middleware console your team would watch in production. Move through the views.
| Time | Entity | Error class | Attempts | Next | Action |
|---|---|---|---|---|---|
| 08:13:04 | order · #1042 → Order | PERMANENT · validation | 3 / 3 | held | Inspect |
| 07:55:22 | contact · 003Ab00910 | TRANSIENT · timeout | 2 / 5 | 09:50 | Retry now |
| 06:40:11 | product · 01tAb44120 | REQUEST_LIMIT | 1 / 5 | paced | Inspect |
Every record carries three identities at once: its Salesforce id, a canonical id, and its Shopify gid. The middleware reads this row before every write and decides create, update, or skip.
Stamp the stable Shopify id into a custom External ID field and upsert, and a re-delivered webhook updates instead of duplicating. This, plus Salesforce matching rules, is what stops the explosion of duplicate contacts before it reaches your CRM.
Change detection runs on the stored checksum, so a catalog of 50,000 items with 12 real edits syncs 12 records, not 50,000.
entity key map · every id matched before write
Inside the platform
Ten engineered subsystems, one shared platform. This is what the Salesforce adapter plugs into, the same machine that runs NetSuite in production today.
01 / 10
Real-time vs batch, and how we keep it reliable
It is classified before it is retried, and nothing disappears into a log file. Colour carries the verdict: amber is sorted, teal retries, red is held, green replays.
A record fails to write to Salesforce or Shopify. Instead of disappearing into a log, it enters the pipeline.
capturedEvery failure is sorted before anything is retried. A REQUEST_LIMIT_EXCEEDED 403 is paced, a validation error is held.
The 403 backs off and the layer paces against both limiters at once, the Salesforce daily allocation and the Shopify leaky bucket, until it clears.
retryingA validation failure, a Person-Account Name write or a missing PricebookEntry, routes straight to the dead-letter queue for review. Never retried blindly, never lost.
held for reviewFix the cause, replay from the queue. A daily reconciliation pass catches anything the live path missed.
replayed · reconciledThe failure modes no vendor page admits
Salesforce CRM ⇄ Shopify integrations break at the same predictable points. Point connectors and most iPaaS templates handle the happy path and skip these. A custom middleware layer is built to handle all of them, explicitly.
Push Shopify customers in naively and you get an explosion of duplicate contacts: guest checkouts, the same person under multiple emails and regions, Shopify's transactional model against Salesforce's Account / Contact / Lead model. Safe sync is external-ID upsert plus DuplicateRule and matching rules and a canonical-ID map, designed up front, not last-writer-wins.
→ external-ID upsert + matching rules + canonical-ID mapShopify and Salesforce can hold conflicting truth at once: "fulfilled" in Shopify, "in progress" or "partially paid" in Salesforce. Last-writer-wins yields drift and conflicting reports. A reliable layer encodes a field-level system-of-record matrix instead: contract and price are Salesforce, fulfillment and capture are Shopify and the ERP.
→ field-level ownership matrix, not last-writer-winsSalesforce Orders enforce a Draft to Activated lifecycle, and every OrderItem must resolve to a PricebookEntry in the right Pricebook2 before the line can be written. Split shipments, partial payments and offline edits do not map to a flat field copy. They need branching mapping logic a fixed connector does not carry.
Two independent limiters drain at once. Salesforce governs by a daily request allocation and 25 concurrent long-running requests; over-limit returns a 403 REQUEST_LIMIT_EXCEEDED. Shopify governs by a leaky bucket, 1,000 points per second GraphQL on Plus. Flash sales and large backfills are where naive syncs break. The layer must read Shopify's cost and throttle status and pace itself.
Salesforce blends multiple buyer-group price books by a configurable "lowest price" or "highest-priority book" strategy. Shopify B2B assigns a Company Location to exactly one catalog and does not blend N books at runtime. Standard connectors often skip this entirely. The middleware pre-resolves the effective buyer price and writes one value per company.
→ pre-resolve via Connect REST, one price per companyThere are three different contract-pricing engines: native B2B Commerce price books, CPQ ContractedPrice (account-scoped, snapshot-at-quote, parent to child), and RLM ContractItemPrice. None of them automatically flow to the storefront, and which one the client runs dictates the entire build. You must find out which engine is in play before quoting.
Claims you will hear, against the docs
Every reality below traces to a primary Salesforce or Shopify doc, or a named vendor source. Rows one through eight bust the common myths. The rest own the ground no vendor page admits.
Account / Contact / Lead. Safe sync requires external-ID upsert plus DuplicateRule and matching rules and a canonical-ID map, designed up front, not last-writer-wins.403 REQUEST_LIMIT_EXCEEDED, a backpressure signal, not an error to retry blindly.PATCH /sobjects/{Object}/{extIdField}/{value}. Salesforce creates if absent and updates if present, so a duplicate delivery is a no-op. A connector that does INSERT on every webhook duplicates records.Person*, custom ones as __pc, and Name is read-only (write FirstName / LastName or hit a DML exception). Orders enforce a Draft to Activated lifecycle where every OrderItem must resolve to a PricebookEntry in the right Pricebook2 first. Both require branching the mapping.GET /commerce/webstores/{id}/pricing/products) and write a single value per company. Standard connectors often skip this entirely.ContractedPrice (account-scoped, snapshot-at-quote, parent to child), and RLM ContractItemPrice. None of them automatically flow to the storefront, and which one the client runs dictates the entire build. CPQ takes a price snapshot at quote time, so later changes do not reach an existing quote line. Find out which engine is in play before quoting.__c data, __r traversal, __e event, __x external, __mdt metadata). A robust middleware discovers the model at runtime via describeGlobal and sObjectDescribe (gating on createable, updateable and FLS) and re-reads the Avro schema on CDC schema-ID change. Bind to API names, never labels.The decision guide
There are four ways to connect Shopify to Salesforce: an app-store or AppExchange connector, an iPaaS or middleware platform, manual import and export, or a custom API integration. The right method depends on data complexity, real-time needs, and how much control you require.
Option 01
eShopSync, Sync Made Easy, CRM Perks. Fast, no-code, fixed scope. Competent on a simple object mapping, often Contact-first.
Right for simple stores with no contract pricing, no account hierarchy, and a single source of customer truth.
Option 02
Celigo, MuleSoft, Boomi, Workato. Flexible, configurable, recurring cost. Wires at the API level with dashboards and retries.
Right for standard multi-system sync where the B2B meaning fits a fixed recipe and you can own the reliability layer yourself.
Option 03
CSV in, CSV out. No automation, no real-time, no idempotency. A starting point, not an operating model.
Right for a one-time migration or a proof of concept, never a production B2B storefront.
Option 04 our build
Built on Shopify webhooks and Salesforce REST and Bulk APIs, full control, owned by you, operated with a support agreement. Where Plus and B2B merchants end up.
Right for B2B contract pricing, account hierarchies, multi-store or international, or high volume, where a connector overspills and you want it owned, not rented by the year.
How long it takes, and what it costs
A Salesforce CRM ⇄ Shopify Plus build lands in one of three bands depending on B2B complexity, which contract-pricing engine is in play, and how many of the break points above you carry. Bands below are market context; your figure is scoped to your build.
Band 01 · foundation
single org · standard catalog · weeks
Band 02 · B2B
B2B Commerce · price books · weeks
Band 03 · multi-cloud
multi-store · ERP in the loop · weeks
We price a Salesforce integration as an engineering engagement: a build with a support agreement, not a per-record meter or a platform seat that renews forever. For market context, app-store connectors run roughly $0 to $300 a month and generic iPaaS commonly runs into 2 to 3× TCO once implementation is counted. Your exact figure is scoped to the band and the break points in play, and we put it in writing before anything starts.
Security posture
Four cards your IT and Salesforce admin will both want answered before a single record moves.
Auth runs on OAuth 2.0 JWT Bearer or Client Credentials on an External Client App, scoped to api plus chatter_api, Permitted Users set to Admin-approved, secrets held in Named and External Credentials. No username-password flow.
Every Shopify webhook is HMAC-verified before it is trusted, then deduplicated by event id, so a replayed webhook cannot create a second order.
Salesforce and Shopify secrets are held as encrypted fields, isolated per tenant and rotatable without downtime. The JWT private key never sits in plaintext, anywhere.
An append-only audit trail logs every operation, source-tagged, so you can prove what happened during a quarter close or a peak event.
Frequently asked questions
The details that separate a team that read the docs from a team that shipped against them. Open any one.
How we build it
Every system sits behind the same contract: authenticate, fetch master data, push orders, validate, report. The platform runs in production today against a live enterprise system. Salesforce is a registered target on that same framework, its path fully mapped to REST, Pub/Sub and OAuth.
We say this plainly. NetSuite runs in production today. The Salesforce path is fully mapped. The credibility is a real platform and a precisely understood Salesforce path, not a claim of live Salesforce traffic. That distinction is exactly what your architect can verify, which is why we lead with it.
Proof · the Clarius unified account page
One named Salesforce proof, then the platform numbers your architect can verify in the repository.
Clarius · Salesforce CRM + NetSuite + Shopify PlusMakro built a unified account page for Clarius that pulls customer information, orders and invoices from Salesforce CRM, NetSuite ERP and Shopify Plus into a single customer-360 view, with a custom Membership app alongside. Combined with Shopify Flow feeding order data into the ERP, it eliminated manual data entry and cut operational costs by over $100,000 a year in third-party app fees. The Salesforce leg is the CRM spine: Account as the system of record for the customer, surfaced in one place.
Automated tests across 139 suites, run on every change.
Coverage gate enforced before any deploy.
End-to-end browser tests across Chromium, Firefox and WebKit.
Failure classes, each with its own pre-decided policy.
Credentials stored in plaintext, anywhere, ever.
A real conversation about your Salesforce org, your B2B pricing, and the three sharp questions you already have.