Build vs. Buy>ERP, Accounting & Integrations>GraphQL Throughput for Custom Sync

Should You Build or Buy Real-Time ERP Sync on Shopify Plus?

Written by Deploi EditorialReviewed by Martin Dejnicki, Director of SEO & AI SearchUpdated September 2026Pricing verified September 2026 (quote-based tiers excluded)

Real-time ERP sync on Shopify Plus is a BUY when a reviewed connector exists for your ERP, and a BUILD on the GraphQL Admin API when none does. Plus gives each app 1,000 calculated query points per second, ten times the Standard plan's 100 (verified Sep 2026). The 1,000-point budget makes a real-time sync feasible for a connector and a custom build alike. A custom sync costs $30,000–$80,000 (Deploi estimate, illustrative).

Your profile — see how the verdict shifts

VerdictBUY (a reviewed connector for your ERP) · BUILD on the GraphQL Admin API when no connector exists or the sync semantics are yours · never start a new sync on REST
Buy score
7.2
Build score
5.9
Confidence
HighRead shopify.dev's usage-limits table: the GraphQL Admin API allows 100 calculated query points per second on Standard, 200 on Advanced, 1,000 on Shopify Plus and 2,000 on Shopify for enterprise (Commerce Components), verified Sep 2026. Shopify's REST rate-limit page marks the REST Admin API legacy as of October 1, 2024 and steers all new apps and integrations to GraphQL. Searched the App Store sitemap for netsuite, pipe17, patchworks, celigo, workato, integrator, erp-connector and oms and fetched every matching listing: three NetSuite connectors (Robust NetSuite Integrator at 4.5★ across 81 reviews, Oracle's own at 4.4★ across 8, and a TechMarbles listing at 5.0★ across 32), plus Pipe17 and Patchworks as iPaaS lanes. Celigo and Workato have no App Store listing. Connector prices are verified from the listings; the build band is a Deploi estimate.
Reference scenario
$20M–$500M GMV · Shopify Plus (1,000 GraphQL points per second per app) · NetSuite or a comparable ERP with a live connector · 3,000–15,000 orders/month · real-time inventory across 3–10 locations · agency dev bench
As of
September 2026

Decision at a Glance

Your profileVerdictWhy
ERP with a reviewed connector, under 5,000 orders/monthBUYRobust NetSuite Integrator runs $80–$500/month and Oracle's own connector $199.92–$916.58/month (verified Sep 2026); a custom sync costs five to ten years of that. Plus's 1,000-point budget is spare capacity at this volume.
ERP with a connector, 5,000–50,000 orders/month, multi-location inventoryBUYThe reference scenario. Buy the connector, then measure inventory lag during your next launch. Multi-location real-time inventory is where a vendor's polling interval shows, and Plus's 1,000 points per second is what lets the connector's real-time option keep up.
Proprietary or home-grown ERP, PIM or WMS with no connectorBUILDNo listing serves your system, so the choice is a custom GraphQL app or an iPaaS you configure yourself. Build with bulk operations, webhooks and a cost-budgeted queue; $30,000–$80,000 (Deploi estimate, illustrative).
Several systems and several stores (ERP, PIM and WMS across 3+ expansion stores)CUSTOMIZEAn iPaaS such as Pipe17 ($24,000/year Standard, verified Sep 2026) or Patchworks carries the plumbing; you still build the mappings and the reconciliation jobs. Each store is a separate app install, a separate bill and a separate 1,000-point bucket.

What GraphQL Throughput for Custom Sync Actually Drives

OutcomeImpactHow it works
Operational efficiencyHighA sync that keeps pace during a launch means the fulfillment team works from truthful stock and order data instead of reconciling spreadsheets the morning after.
Revenue — directHighInventory lag oversells the fastest-moving SKUs during a drop; every oversell is a refund, an apology and a lost repeat purchase, so sync speed is a revenue line.
Customer experienceMediumOrder confirmations, fulfillment updates and returns that reflect ERP truth within minutes read as a competent store; ones that arrive hours late read as a broken one.
Data & insightMediumA custom sync writes cost, inventory and fulfillment events to a warehouse you control on the way through; that stream is the input for margin reporting and any forecasting model you attempt.

Spend ceiling: Size the spend to the cost of one oversold launch. A drop that oversells 200 units at a $60 average price (illustrative) is $12,000 in refunds (illustrative) before support time. Connectors run $80–$916.58/month (verified Sep 2026); a custom sync is $30,000–$80,000 once (Deploi estimate, illustrative). Spend the build money only when no connector maps your system.

What buying enables (top apps)

  • + A reviewed NetSuite sync live in weeks, with the vendor absorbing Shopify's quarterly API version changes
  • + Field mapping, RMA handling and multi-subsidiary support on higher tiers, from $80 to $916.58/month across the two priced NetSuite connectors (verified Sep 2026)
  • + Plus's 1,000-point-per-second bucket applies to the connector's app too, so a vendor's real-time option actually runs in real time

What building additionally unlocks

  • + Bespoke semantics no connector maps: reserved inventory, custom PIM attributes into metafields, multi-subsidiary routing rules
  • + Bulk operations and webhooks tuned to your catalog shape, in place of a vendor's fixed polling interval
  • + An event stream into your own warehouse, so margin and forecasting models read the same data the sync moves
  • + One custom app per integration, each with its own rate-limit bucket, so a catalog rebuild never starves order flow

Find Your Verdict in 3 Questions

  1. Does a reviewed App Store connector exist for your ERP (NetSuite alone has three)?

    Yes: Go to question 2.

    No: Your verdict: BUILD — a custom GraphQL sync with bulk operations, webhooks and reconciliation jobs, $30,000–$80,000 (Deploi estimate, illustrative).

  2. Do your sync semantics go beyond standard objects: reserved inventory, multi-subsidiary rules, custom PIM attributes into metafields?

    Yes: Go to question 3.

    No: Your verdict: BUY — install the connector, map the fields, and measure inventory lag during your next launch.

  3. Does the connector's top tier or its professional-services team cover those semantics?

    Yes: Your verdict: BUY — the Enterprise tier of a connector still costs a fraction of a custom build.

    No: Your verdict: CUSTOMIZE — run the connector for standard objects and build a small custom app for the bespoke flows, each in its own rate-limit bucket.

The TCC Scorecard — 12 Dimensions

TCC — Total Cost of Capability: what it actually costs to have this capability over three years, whichever way you get it. Each dimension is scored 0–5 for both paths. How we score →

DimensionBuyBuildWhy
Cost
Acquisition & implementationA NetSuite connector installs and maps in days to weeks at $80–$916.58/month (verified Sep 2026); the custom sync is an 8–14 week build at $30,000–$80,000 (Deploi estimate, illustrative).
Recurring feesConnectors bill monthly for as long as you run them, and Pipe17's Standard tier is $24,000/year (verified Sep 2026); the build's recurring line is upkeep at roughly 15–20% of build cost per year (Deploi estimate).
Maintenance & upgradesThe vendor absorbs quarterly Admin API version bumps for the connector; your team absorbs them for a custom sync, along with every ERP schema change the business makes.
Switching & exitLeaving a connector means rebuilding field mappings and re-syncing history on the next tool; leaving a custom sync costs nothing, because the code and the mappings are already yours.
Risk
Vendor riskOracle's own connector carries 8 reviews and Pipe17 21 (verified Sep 2026); only Robust NetSuite Integrator has 81. A custom app has no vendor to lose, only a bench to keep.
Security & compliance surfaceA connector holds customer and order data in a third party's cloud between the two systems; a custom sync keeps it between your ERP and Shopify, with your own key management to get right.
Platform-deprecation exposureShopify declared the REST Admin API legacy on October 1, 2024 and steers every new integration to GraphQL (verified Sep 2026); a connector vendor carries that migration for you, while a new custom sync must start on GraphQL from day one.
Value
Fit to requirementConnectors map the standard objects well and struggle with bespoke semantics: reserved inventory, custom PIM attributes into metafields, multi-subsidiary rules. A custom sync maps exactly your model.
Time to marketWeeks with a connector against 8–14 weeks for a custom build (Deploi estimate, illustrative); the ERP side's readiness usually sets the real go-live date in both lanes.
Performance & scaleBoth lanes draw on the same 1,000-point-per-second Plus bucket per app (verified Sep 2026); a custom sync can spend it on bulk operations and webhooks instead of a vendor's fixed polling interval.
Data ownership & AI-readinessA custom sync writes inventory, cost and fulfillment events into a warehouse you control on the way through; a connector's logs and mappings stay inside the vendor's platform.
Focus & opportunity costAn ERP sync is plumbing, not differentiation; spend the dev bench here only when no connector fits, because a stalled sync project blocks every other roadmap item behind it.

The App Landscape

AppStatusPricingBest for
Robust NetSuite IntegratorLive4.5★, 81 reviews; 30-day free trial. The most-reviewed independent NetSuite connector: order, inventory and fulfillment sync with advanced field mapping; higher tiers add RMA management, serialized items, two-way order edits, B2B and multi-subsidiary support.Limited $80/month; Advanced Unlimited $199.99/month; Max Unlimited $300/month; Enterprise $500/month (verified Sep 2026)NetSuite merchants who want the standard sync live in weeks from a reviewed vendor
NetSuite ERP Connector (Oracle)Live — flagged4.4★, 8 reviews; Oracle's own connector, unusual among ERP vendors in publishing fixed prices. B2C tiers cover order, coupon, gift-certificate and fulfillment sync with a real-time price and quantity option; the B2B tier adds invoice posting and standalone customer sync.B2C Standard $199.92/month; POS $249.92/month; B2C Premium $249.92/month; B2B $916.58/month (verified Sep 2026)Merchants who want the ERP vendor itself accountable for the sync
Pipe17LiveOrder-ops-focused connector layer — routing and fulfillment orchestration rather than classic iPaaS$24000/mo tiers at entry volumes; scales with volume (verified Sep 2026)Multi-channel merchants routing orders across an ERP, a 3PL and marketplaces
PatchworksLiveIPaaS alternative pitched at multi-system retail stacks, with a strong agency channelTiered platform subscriptionTeams that want to configure mappings themselves without owning the runtime
Custom GraphQL sync (bulk operations, webhooks, reconciliation)Build laneA custom app per integration with its own 1,000-point bucket: bulk operations for catalog and inventory loads, webhooks for order and fulfillment events, scheduled reconciliation jobs because Shopify documents that webhook delivery isn't always guaranteed, and a cost-budgeted queue that reads the throttle status on every response.$30,000–$80,000 one-time plus ~15–20% a year in upkeep (Deploi estimate, illustrative)A proprietary ERP, PIM or WMS with no connector, or sync semantics no vendor maps

The Build Path

  • One custom app per integration: Each custom app draws on its own 1,000-point-per-second bucket on Plus (verified Sep 2026), so the ERP sync never competes with your analytics pull or a third-party app for capacity.
  • Bulk operations for the big loads: Full-catalog and full-inventory loads run as bulk operations, one asynchronous job in place of thousands of paginated calls. On API versions 2026-01 and higher each app can run up to five bulk query operations per shop simultaneously, and a job must complete within 10 days (verified Sep 2026).
  • Webhooks plus reconciliation for the real-time path: Order, fulfillment and inventory changes arrive by webhook; a scheduled reconciliation job re-reads recent changes because Shopify documents that webhook delivery isn't always guaranteed and removes a subscription after 19 consecutive failed delivery attempts (verified Sep 2026).
  • Cost-budgeted queue with idempotent writes: Every GraphQL response carries the remaining capacity and restore rate; the queue waits before firing into an empty bucket, treats THROTTLED as a pause instruction, and keys every mutation so a replay never double-writes.
Effort band
$30,000–$80,000 build — Deploi estimate (illustrative); lands in the $25–75K contact-form band. Orders and fulfillments are the cheap half; multi-location inventory with reservations and custom attribute mapping set the number
Typical timeline
8–14 weeks (Deploi estimate, illustrative); ship orders out and fulfillments in first, then inventory, then catalog and pricing
Maintenance, honestly
~15–20% of build cost per year (Deploi estimate): roughly $4,500–$16,000/yr (Deploi estimate, illustrative) for quarterly Admin API version bumps, ERP schema changes and new fields the business asks for. There is no subscription line.
What you own — and what you take on
You own: the field mappings, the sync logs, the reconciliation logic and the event stream into your warehouse. You take on: on-call for the sync during launches, and every ERP upgrade that changes a field your mapping depends on.

3-Year Total Cost of Capability

Buy (app path)Build (custom path)
Year 0 (setup)$5,000–$15,000 (connector setup, mapping and testing)$30,000–$80,000
Years 1–3 (recurring)$7,200–$18,000 (subscription)$13,500–$48,000 (maintenance)
3-year total≈$12,000–$33,000≈$43,500–$128,000
Illustrative cumulative cost over 36 months$0$21k$43k$64k$85kMo 0Mo 12Mo 24Mo 36Buy (app path)Build (custom path)
Illustrative cumulative cost: the connector stays cheaper across the whole three-year horizon whenever one fits your ERP. The build wins only when no connector exists, or when a vendor's polling interval or field model costs you oversells that dwarf the subscription.
  • All figures illustrative samples for the reference scenario — not quotes, not verified pricing.
  • Buy path: a reviewed NetSuite connector at its mid tiers, $200–$500/month, plus a mapping and testing project at go-live; connector list prices verified Sep 2026, the setup project illustrative.
  • Build path: custom GraphQL sync covering orders, fulfillments, inventory and catalog, with reconciliation jobs; three-year horizon.

What the Sticker Price Hides

On the buy path

  • Pipe17's published Standard tier is $24,000/year for 4,000 orders/month (verified Sep 2026); tiers above it are quote-based, so growth changes the price without a listing to check
  • Connector polling intervals are the hidden lag: ask what the inventory sync interval is and whether the app subscribes to webhooks before you sign
  • Custom field mapping past the standard objects is where professional-services fees start, and a 4.9★ rating does not tell you that
  • Each expansion store is a separate app install and a separate bill; Shopify bills apps per store, not per organization (verified Sep 2026)

On the build path

  • Starting on REST because an old snippet works: the REST Admin API has been legacy since October 1, 2024 (verified Sep 2026) and the migration lands on you
  • Trusting webhooks alone; Shopify says delivery isn't always guaranteed and expects reconciliation jobs (verified Sep 2026)
  • Multi-location inventory with reservations is the fiddly third of the build and the part that oversells when wrong
  • ~$4,500–$16,000/yr upkeep (Deploi estimate, illustrative), mostly ERP schema changes and quarterly API version bumps

What Merchants Say

Dev leads describe the same launch-night failure: the connector never errored, the inventory sync fell fifteen minutes behind during the drop, and the oversells arrived as support tickets the next morning.
community-reported (2026 research corpus)
Connector reviews cluster on mapping edge cases: multi-subsidiary NetSuite setups, kit and bundle items, and custom fields that turn out to need paid professional services to sync.
app-store 1–2★ review theme

If You Change Your Mind Later

If you bought and outgrow it

Leaving a connector exports nothing useful; the mappings and sync logs live in the vendor's platform, so the exit is a re-implementation on the next connector or a custom build. Keep your own field-mapping document from day one and the exit shrinks to a few weeks of configuration.

If you built and want out

Nothing strands: the mapping tables, the reconciliation jobs and the queue are ordinary code in your integration layer and travel with you to a new ERP or a new storefront. A future retreat to a connector is an install plus re-mapping, because the model is already documented in your code.

When This Answer Changes

We're watching for:

  • Shopify raising or restructuring the per-plan GraphQL limits (100, 200, 1,000 and 2,000 calculated query points per second as of September 2026)
  • Shopify announcing a removal date for the REST Admin API, legacy since October 1, 2024
  • Shopify's B2B integrations page adding or dropping named ERP and iPaaS partners (NetSuite, Microsoft Dynamics 365, Fulfil, QuickBooks Online, Pipe17, Patchworks and OmnifiCX as of September 2026)

Verdict change log:

No changes since first publication (September 2026).

Common Questions

Does Shopify Plus have a higher API rate limit?

Yes. Shopify Plus allows 1,000 calculated query points per second per app on the GraphQL Admin API, against 100 on Standard plans, 200 on Advanced and 2,000 on Shopify for enterprise (verified Sep 2026). On the legacy REST Admin API the Plus bucket is 400 requests with a leak rate of 20 per second, ten times the standard 40 and 2. Every app installed on the store draws on its own bucket.

Should a new ERP integration use REST or GraphQL?

A new ERP integration should use the GraphQL Admin API. Shopify declared the REST Admin API legacy on October 1, 2024 and states that all apps and integrations should be built with GraphQL (verified Sep 2026). REST's Plus limits still apply at 400 requests per bucket and 20 per second, but a sync started on REST today inherits a migration before it finishes its first year.

Is 1,000 points per second enough for real-time inventory sync?

1,000 points per second covers real-time inventory for most mid-market stores when the sync uses webhooks and bulk operations instead of polling. Polling every SKU on a timer burns budget on unchanged data; inventory and order webhooks spend it only on changes. Full-catalog loads belong in bulk operations, which run as one asynchronous job and, on API versions 2026-01 and higher, allow five simultaneous bulk queries per app per shop (verified Sep 2026).

Your Next Steps

If you're going with BUY(matches your selected profile)

  1. Shortlist the connectors for your ERP and ask each vendor for its inventory sync interval and its behavior on THROTTLED
  2. Map your standard objects first: orders out, fulfillments and inventory in, then catalog and pricing
  3. Run the connector against a Plus staging store and measure lag during a simulated drop
  4. Write your own field-mapping document alongside the vendor's, so the exit stays cheap
  5. Diary a re-check at 5,000 orders/month, where polling intervals and tier prices both move

If you're going with BUILD

  1. Register one custom app per integration so each keeps its own 1,000-point bucket
  2. Move catalog and inventory loads to bulk operations before writing a single paginated query
  3. Subscribe to order, fulfillment and inventory webhooks and schedule the reconciliation job that backstops them
  4. Build the cost-budgeted queue with idempotent mutations and log every THROTTLED response
  5. Add a lag monitor comparing Shopify and ERP state, with a pause switch for non-critical syncs during launches

Official Docs & Sources

Official documentation linked for verification — our verdicts and estimates are our own.

Ready for an ERP sync that survives launch night?

We scope the connector-versus-custom call against your ERP, then build the GraphQL sync, the reconciliation jobs and the lag monitors when no connector fits.

Contact us today

API & integration development

Verdict scored for the reference scenario above. Estimates are not quotes; connector pricing and platform limits carry their verification date and get re-verified quarterly. Full scoring anchors: see the TCC methodology.

Read how we score these decisions (the TCC Framework). No affiliate links, no paid placement — no app vendor pays to appear here.

No affiliate links. No paid placement. We make money building and integrating solutions — not on referral fees.