QuickBooks vs Custom ERP: Which to Choose for Your US SMB in 2026

Choose QuickBooks Online when your business has standard accounting, sub-$5M revenue, single legal entity, no manufacturing, no multi-warehouse with batch tracking, and no commission structure beyond flat percentages. Choose a custom ERP when QuickBooks forces you to maintain 3+ side spreadsheets, when you pay 4-6 paid integrations to fill gaps, or when manual reconciliation eats more than $30,000/year of staff time. QuickBooks Online: $30-235/month. Custom ERP for a US SMB: $25,000-90,000 initial build plus $300-1,500/month infrastructure.

By Pedro Corgnati -- founder of SystemForge, full-stack developer with 8+ years building custom software for SMBs. I have shipped custom operational layers on top of QuickBooks Online via the Intuit API and full QuickBooks replacements for US clients in distribution and light manufacturing.

I'm not anti-QuickBooks. For 80% of US SMBs, QuickBooks Online is the right call. The argument for a custom ERP is not about cost -- it is about functionality QuickBooks does not deliver and the operational hidden costs of working around it.

Direct comparison summary

QuickBooks Online is the dominant US SMB accounting platform. Strong: tax handling (1099, W-2, sales tax via integration), accountant network, ecosystem (Shopify, Amazon, Square, Bill.com, Avalara). Weak: inventory beyond basic, no real manufacturing, role-based access only on Advanced, anemic custom reporting.

Custom ERP is software designed around your specific business rules: production scheduling, lot/serial tracking, multi-warehouse, complex commissioning, custom pricing, multi-entity consolidation, embedded BI. Built once, owned forever, evolves with the business.

The right question: does QuickBooks resolve 100% of what you need (stick with it), or do you maintain spreadsheets, hire integrations and reconcile manually to fill its gaps (evaluate custom)?

What QuickBooks Online actually does

Plans, features, real limits per tier

Plan	Cost/month	Users	Strengths	Real limits
Simple Start	$30	1	Basic accounting, invoicing, expense tracking, mileage	No bill pay, no inventory, no multi-user
Essentials	$60	3	+ bill management, time tracking, multi-currency	No inventory, no project profitability
Plus	$90	5	+ inventory tracking, project profitability, class/location	Inventory is FIFO only, no batch/serial, no multi-warehouse
Advanced	$235	25	+ role-based access, custom fields, workflow automation, smart reporting	Still no real manufacturing, weak multi-entity

QuickBooks for retail and e-commerce (Shopify, Amazon, eBay)

The integration story is solid for standard retail. Shopify, Amazon, eBay and Square sync orders, fees, refunds and inventory adjustments. Most US SMB retailers under $5M run this stack happily.

Friction starts when you need: cross-channel inventory truth (Cin7, Linnworks or custom), sales tax automation across multiple states (Avalara at $50-200/month), 3PL integration (ShipStation at $30-200/month), abandoned-cart and email automation outside Shopify (Klaviyo at $45-1,000/month).

By the time you stack 4-5 integrations on top of QuickBooks, your "free" accounting platform is a $500-1,500/month operational stack with reconciliation gaps.

Real QuickBooks limits SMBs hit at scale

The honest list of where QuickBooks Online breaks at scale:

• No real MRP -- bill of materials is rudimentary; no routing, no work orders, no shop-floor scheduling.
• No lot/serial/expiration tracking -- a hard stop for food, pharma, cosmetics.
• No multi-warehouse with transfers -- the "Advanced" version has location tracking but it is not a real WMS.
• Weak multi-entity -- you run one QBO file per legal entity and consolidate manually or with QuickBooks Enterprise (different product, different price).
• Custom reports are weak -- Smart Reporting (Advanced plan) is better but still falls short of Power BI / Metabase / Looker for cross-cutting analysis.
• API rate limits -- the Intuit API is throttled (500 req/minute, with daily caps); high-volume e-commerce hits this.
• Sales tax across states -- requires Avalara, TaxJar, or manual filing; QBO does not handle Wayfair nexus tracking natively.
• Complex commissioning -- tiered commissions, splits, claw-backs are spreadsheet land.

When 3+ of these matter to your operation, the hidden cost of working around them adds up.

What a custom ERP gives you

Specific business rules QuickBooks can't model

Examples from real US client engagements:

• Distributor with per-customer, per-product pricing matrices, EDI to 12 retailers and weekly truckload allocation rules.
• Light manufacturer of contract food packaging with lot tracking, allergen segregation, batch certificates of analysis, and per-customer minimum-order quantities.
• Multi-location service business (6 dental clinics) with shared patient records, per-clinic P&L, per-dentist commission and claims integration with payer clearinghouses.
• B2B + DTC hybrid with separate price lists, shared inventory pool, and channel-level allocation rules.

QuickBooks can hold the accounting layer for any of these. Operations need their own system.

Multi-entity, multi-warehouse, lot/serial, MRP-light

A custom ERP delivers exactly the operational depth you need -- not a generic feature trying to fit every industry. For most SMBs that means: real multi-warehouse with transfer logic, lot/serial with expiration, MRP-light for assembly or kitting, multi-entity with intercompany transactions, custom pricing with rule-based discounts.

Custom reports and embedded BI

Postgres + Metabase or Postgres + Power BI gives you analytics depth that QuickBooks reporting cannot match. Cohort analysis, custom margin breakdowns, vendor scorecards, customer profitability with allocated overhead -- all become trivial when you own the data layer.

Real cost comparison over 36 months

A representative comparison for an $8M revenue light manufacturer:

Item	QuickBooks Plus + integrations	Custom ERP
Software license / build	$90/month QBO + $1,400/month integrations	$65,000 build (year 1)
Infrastructure	Included	$700/month
Reconciliation FTE	0.6 FTE = $36,000/year	0.1 FTE = $6,000/year
Customization	$200-400/hr per change	Internal, ongoing
36-month total	~$162,000	~$120,200 (incl. amortized build)

The custom build pays back inside 24 months, then continues to widen the lead because the FTE cost compounds.

3 switching scenarios

Multi-state distributor ($12M revenue)

QuickBooks + Cin7 + ShipStation + Avalara + Excel for allocation = $1,800/month + 0.8 FTE on reconciliation. Custom ERP scope: orders, allocation rules, EDI to top 5 retailers, custom pricing per customer, real-time inventory across 3 warehouses. Build cost: $75,000. Year-1 break-even: 16 months including FTE savings.

Light manufacturer ($8M revenue)

Detailed above. Custom build $65,000, payback 14 months.

Multi-location service business ($4M revenue, 6 clinics)

Considered NetSuite ($42,000 year-one quote). Built custom on Next.js + Postgres + Stripe + claims clearinghouse API: $55,000 build + $700/month infra. Equivalent to 1.3 years of NetSuite, with the platform being theirs.

Hybrid path: keep QuickBooks for accounting, build operational layer

This is the path I recommend most often. Keep QuickBooks Online for accounting (tax filings, GL, AP, AR, your accountant likes it). Build a custom operational layer for orders, inventory, production, custom pricing, BI. Sync between the two via the Intuit API:

• Customer master: source of truth in custom ERP, sync to QBO when first invoice generated.
• Invoices: created in custom ERP from completed orders, posted to QBO via API.
• Bills: received in custom ERP if AP is operational, posted to QBO; otherwise stay in QBO.
• Inventory adjustments: source of truth in custom ERP, summarized journal entry to QBO monthly.
• Bank feeds: stay in QBO, accountant reconciles there.

Build cost: $25,000-50,000 versus $50,000-90,000 for full replacement. Faster payback (8-14 months), preserves the accountant's workflow, lower risk.

For more depth, see the build vs buy analysis for ERP, the SMB ERP cost comparison, the QuickBooks/Zoho vs custom ERP guide, the ERP integration playbook, and the ERP by industry chooser.

How to test custom ERP fit before committing

Before you commit $50k-90k to a custom build, validate with a 30-60 day pilot.

1. Document the 5 most painful workflows with screenshots and time-spent measurements over 2 weeks. This becomes your scope baseline.
2. Build a 2-week throwaway prototype of the highest-pain workflow (Properties + Inventory adjustments, or Orders + Custom Pricing). Real cost: $6,000-12,000. Real outcome: you know if the team adopts the workflow before scaling.
3. Run it parallel to QuickBooks for 30 days with 2-3 users. Measure time saved, errors caught, satisfaction.
4. Decide: scale up the prototype (typical), abandon (rare but valid), pivot scope (common).

This is how you avoid the $80,000 build that nobody uses. The prototype dollars are not wasted -- they become the foundation of the full build or they save you 10x in avoided sunk cost.

What to demand from any custom ERP vendor

• TypeScript strict mode plus end-to-end tests on critical paths.
• Schema-first development with proper foreign keys, indexes and constraints.
• CI/CD with preview deploys per pull request.
• Sentry + uptime monitoring from day one.
• Code ownership -- you own the GitHub repo, not the vendor.
• Documentation in repo: README, .env.example, ARCHITECTURE.md, RUNBOOK.md.
• Onboarding plan for the next developer -- this is the single most important deliverable for long-term success.

Without these, you have a $80,000 black box. With them, you have an asset that survives any developer turnover.

FAQ

At what revenue does QuickBooks genuinely break? Around $5M with operational complexity (multi-warehouse, lot tracking, manufacturing) or around $10-15M with simple operations but multi-entity needs. The early signals are 3+ side spreadsheets and a paid integration stack costing $800+/month.

Can I keep QuickBooks for accounting and build a custom operational layer? Yes -- this is the hybrid path and the right call for most $5-15M US SMBs. Custom ERP for operations, QuickBooks for accounting, sync via the Intuit API. Cheaper, faster, lower risk than full replacement.

How much does the average custom ERP cost in the US in 2026? $25,000-50,000 for an operational layer on top of QuickBooks. $50,000-90,000 for a fuller replacement. $90,000-180,000 for multi-entity/multi-location with light manufacturing. Add 15-20% annual maintenance and $400-1,800/month infrastructure.

What integrations does QuickBooks Online API support reliably? Customers, vendors, items, invoices, bills, payments, journal entries, bank feeds, attached files, classes, locations. Throttled at 500 requests/minute. Webhooks available for change notifications. Reliable enough to be the back-end of a hybrid ERP.

Is NetSuite a better step than custom? For $20M+ multi-entity businesses with structured operations and a $50k+ implementation budget, possibly. For $5-15M businesses, NetSuite is usually overkill and overpriced versus a focused custom build that solves your specific business rules.

How long does a custom ERP take to ship? Operational layer: 3-4 months. Full replacement: 5-7 months. Multi-entity / multi-location with manufacturing: 7-10 months. Phased delivery (MVP in 2-3 months, then iterations) is the right approach for most.

---

If you're hitting QuickBooks limits and want a second opinion on the hybrid path versus full custom before you commit, message me on WhatsApp -- no pitch, no commitment. Or see the custom systems service.