TL;DR — key numbers at a glance
Salesforce integration costs range from about $5,000 for a single point-to-point connector to $600,000+ for an enterprise multi-system program on middleware in 2026. A custom API build linking one or two systems typically runs $20,000–$60,000; an iPaaS program such as MuleSoft or Boomi, $40,000–$150,000 plus licensing. Data volume, system count and compliance — not Salesforce licences — drive the number.
"How much does a Salesforce integration cost?" has the same honest answer as "how much does a renovation cost?" — it depends entirely on what you are connecting and how clean it already is. Wiring one tidy app to Salesforce and synchronising your CRM, ERP and data warehouse in real time are both "integrations," and they are two orders of magnitude apart in price. Our Salesforce integration team scopes this exact spread every week; here is the 2026 landscape upfront:
- Point-to-point connector: $5k–$20k. One system, one direction, a clean API. Fast and cheap; brittle once you have several of them.
- Custom API integration: $20k–$60k. One or two systems, bidirectional, some Apex logic and error handling. The common mid-market build.
- Middleware / iPaaS program: $40k–$150k to build. Several systems on a platform like MuleSoft or Boomi, with reusable APIs and monitoring. The 2026 default once you cross ~3 integrations.
- Enterprise multi-system: $150k–$600k+. ERP, data warehouse and multiple clouds, real-time, with governance and audit. Plus platform licensing.
- Each connected system adds roughly $10k–$50k; data migration runs $5k–$60k; regulated data and Salesforce Shield add 25–35%.
- MuleSoft licensing centres on a ~$55k/year median, with enterprise total cost of ownership reaching $250k–$600k+ in year one — discounts of 35–55% are achievable.
The four integration approaches and what each costs
Pricing only makes sense once you fix how you are integrating. There are four practical approaches in 2026, and the jump in capability between them is exactly the jump in cost.
| Approach | What it does | Typical cost | Timeline |
|---|---|---|---|
| Point-to-point connector | Links one external system to Salesforce, usually one direction, via an existing API or connector | $5k–$20k | 2–4 weeks |
| Custom API integration | Bidirectional sync across one or two systems with custom Apex, field mapping and error handling | $20k–$60k | 6–12 weeks |
| Middleware / iPaaS | Several systems orchestrated on a platform (MuleSoft, Boomi) with reusable APIs, monitoring and governance | $40k–$150k | 3–6 months |
| Enterprise multi-system | ERP, data warehouse and multiple clouds, real-time, with data governance, audit and hypercare | $150k–$600k+ | 6–12 months |
Most companies asking this question in 2026 land in the second or third row. A single point-to-point connector is fine for one app but turns into spaghetti once you have five of them; the enterprise tier is more than a first integration needs. The realistic target for a first serious CRM-to-back-office integration is a custom API build at $20k–$60k, or a middleware program at $40k–$150k if you already know several systems are coming — and where you land inside those bands is decided by the next four sections. This is squarely the territory of our Salesforce integration services, which is where the rest of this guide is grounded.
Cost by the system you're connecting
The approach bands above answer "how am I building this?" The other way people ask the question is "how much to connect Salesforce to this specific system?" — and that number depends on how tidy the other side's API is, how much data moves, and how tightly regulated it is. Here is how the common targets typically map onto the bands above.
| System you're connecting | Typical cost | Usual approach |
|---|---|---|
| Collaboration & email (Slack, Teams, Outlook, Gmail) | $500–$6k | Point-to-point / native connector |
| E-signature (DocuSign, PandaDoc) | $2k–$10k | Point-to-point |
| Helpdesk (Zendesk, Freshdesk) | $8k–$30k | Custom API |
| Marketing automation (Marketo, HubSpot, Pardot) | $5k–$25k | Custom API |
| Payment gateway (Stripe, Braintree, Adyen) | $5k–$25k | Custom API |
| E-commerce platform (Shopify, Magento, BigCommerce) | $8k–$40k | Custom API / iPaaS |
| Data warehouse (Snowflake, BigQuery) | $10k–$45k | iPaaS |
| ERP (NetSuite, SAP, Dynamics) | $15k–$100k+ | iPaaS / enterprise |
Two patterns hold across almost every project we scope. First, the cost tracks how well-governed the other system's data model already is — Slack and Outlook are cheap precisely because their APIs and object models are simple and stable, while ERPs are expensive because the data is large, inconsistent and business-critical. Second, connecting to several of these at once is exactly when the point-to-point-versus-middleware decision below stops being theoretical — three separate $10k point-to-point connectors are $30k of build and three things to maintain, while the same three systems on an iPaaS become one governed layer.
Infor, Epicor, SAP or NetSuite: does the ERP vendor change the price?
The "ERP" row in the table above hides real variance between brands, and it's the question we get asked most specifically: what does it cost to connect Salesforce to this particular ERP.
Why vendor and API maturity — not the brand — set the price
The honest answer is that the vendor logo matters less than two things underneath it — how mature and documented the ERP's API is, and whether a pre-built connector already exists for it. A cloud-native ERP with a modern REST API and an AppExchange-listed connector sits at the bottom of the ERP band; an older, heavily customised on-prem install with a thin or paid-tier API sits at the top, regardless of which vendor built it.
| ERP vendor / product line | API & connector maturity | Typical cost pattern |
|---|---|---|
| Infor (SyteLine / CloudSuite Industrial, LN, M3) | Varies by product line; thinner on older on-prem installs | Dedicated connector platforms price subscriptions from roughly $500–$600/month per integration pair; a simple sync runs about $15k–$30k, a multi-workflow build $30k–$100k, and a full managed onboarding program commonly lands near $40k–$50k in year one once setup is included |
| Epicor (Kinetic, SX.e) | Modern REST API on Kinetic; SX.e is older and more custom per install | Connector-platform licensing commonly runs $175–$400/month per connection (roughly $2k–$6k/year); a fuller bidirectional sync with mapping and testing still falls in the $20k–$60k custom-API band |
| SAP (Business One, S/4HANA) | Strong on S/4HANA's public APIs; Business One varies by add-on | $15k–$100k+, same as the general ERP band — S/4HANA's documented APIs keep straightforward syncs toward the lower end |
| NetSuite | Mature REST/SuiteTalk APIs, several pre-built connectors | $15k–$100k+, with well-governed accounts landing in the lower-middle of that range |
| Microsoft Dynamics 365 | Native Microsoft-Salesforce connectors available | Typically the lower end of the ERP band when a native connector covers the core objects |
Why a generic ERP estimate can mislead on a legacy install
A generic "ERP integration cost" estimate can mislead teams on an older Infor or Epicor install: the $15k–$100k+ band is correct, but a thin, paid-tier API on a legacy on-prem system pushes the real number toward the top of it, and toward the custom-API or iPaaS approach rather than a simple point-to-point connector — whatever the vendor's brochure says about "pre-built integrations." Scoping the actual API version and object model in discovery, not the vendor name, is what gets the estimate right.
What actually drives the cost?
Five factors move a Salesforce integration budget more than the approach label itself. Get clear on these and you can predict roughly where in the band you will land.
- Number of systems. Each system you connect is real engineering — an API or events, authentication, field mapping and error handling. Budget roughly $10k–$50k per integrated system. The count, more than anything, decides whether you stay point-to-point or move to middleware.
- Data volume and quality. Clean, well-modelled data is cheap to move; duplicated, inconsistent or undocumented data is where projects overrun. Deduplication, transformation and validation are quietly the most underestimated line in any integration.
- Real-time vs batch. A nightly batch sync is straightforward. True real-time, bidirectional sync — with conflict resolution, ordering and retry logic — is materially more engineering and testing.
- Custom Apex and Lightning work. Triggers, Apex services, Flows and custom Lightning components for the integration logic typically run $100–$200 per hour of senior Salesforce engineering, and a genuinely custom build can exceed $85,000 on the Salesforce side alone.
- Compliance. Regulated data, Salesforce Shield (platform encryption, event monitoring, field audit) and data-residency requirements add 25–35% — not as polish, but as design constraints that touch architecture, logging and testing.
Where does the money actually go?
When you open up a mid-sized Salesforce integration budget, the connector wiring is a small slice. Here is roughly how a typical build divides, and why each part matters.
- Discovery & solution architecture (10–15%): mapping the systems, the objects and the data flows, and deciding point-to-point versus middleware. Cheap to do, expensive to skip.
- Build & configuration (40–55%): the connectors, Apex, Flows, field mapping and the integration logic itself. The largest slice, and where seniority pays off.
- Data migration & quality (15–30%): cleansing, deduplicating, transforming and validating the data moving into and between systems. The work that decides whether anyone trusts the result.
- Testing & UAT (10–20%): reconciliation, edge cases, failure and retry paths, and user acceptance. Integrations fail silently; this is how you catch it before your sales team does.
- Go-live & hypercare (5–10%): cutover, monitoring and the first weeks of fixes as real data flows. Always budget for it.
This is ordinary — if demanding — enterprise software development work, and the proportions hold whether you build point-to-point or on a platform. The lesson for budgeting: a quote that is mostly "connectors" has under-scoped the data and the testing, which are where the time actually goes. For a structured way to pressure-test an estimate, see our software project estimation guide.
The ERP implementation cost breakdown in practice
For an ERP integration specifically, this breakdown skews even further toward data and testing: discovery often needs an extra pass just to document the ERP's object model, data migration and quality work (15–30%) routinely lands at the top of that band because ERP data is the largest and most inconsistent in the project, and testing (10–20%) has to cover financial or inventory reconciliation, not just a happy-path sync. A $60,000 ERP integration implementation, in other words, is rarely $60,000 of connector code — it is closer to $25,000–$35,000 of build plus $15,000–$20,000 of data work and $8,000–$12,000 of testing, with discovery and hypercare filling the rest.
Point-to-point vs middleware: when each wins
The single biggest architectural decision — and a major cost lever — is whether to connect systems directly or route them through middleware. This is the integration world's build-versus-buy question, and the answer flips with scale.
Stay point-to-point when…
You have one or two connections to make, the other systems have decent APIs, and you do not expect the count to grow quickly. There is no platform licence, less to build, and you get live value in weeks. For a single connector — say, syncing form leads or an e-commerce order into Salesforce — this is the cheapest path and the right one. The risk is purely future: every new direct link multiplies the connections you have to maintain.
Move to middleware (iPaaS) when…
You are connecting three or more systems, need reusable APIs, central monitoring, error handling and a single place to govern data flows, or you are running a real-time, bidirectional sync that point-to-point links make fragile. Platforms like MuleSoft and Boomi carry a licence cost, but they pay it back by turning a tangle of brittle one-off links into a managed, observable layer. At enterprise scale, the maintenance tax of point-to-point sprawl is exactly what makes middleware the cheaper option over the life of the system. Workato and Jitterbit come up in the same conversation, especially when SAP is one of the systems in scope — all four platforms ship native Salesforce and SAP connectors, so the real comparison between them is connector depth for your specific systems and your team's in-house iPaaS skill, not just list price.
The pragmatic path most teams take: start point-to-point to prove a single connection and learn the data, then adopt middleware once the third or fourth integration appears and the maintenance burden — not just the build cost — becomes the real number. That sequence keeps early spend small and the platform decision informed by real flows rather than a slide. The same logic applies whether your CRM is Salesforce or HubSpot; the threshold, not the vendor, is what changes.
The running costs nobody quotes
The build is a one-off; an integration is a living system that has to survive Salesforce's three annual releases and every schema change in the systems it touches. The teams that get surprised by their bill almost always priced the project and forgot the year. Plan for four ongoing lines:
- Middleware / iPaaS licensing. If you use MuleSoft or Boomi, this is usually the biggest ongoing line. MuleSoft is licensed by capacity, not per integration — the 2026 median lands near $55k/year, with enterprise programs reaching $250k–$600k+.
- Hosting & API limits. Custom integration services run somewhere, and Salesforce enforces API call limits — high-volume, chatty integrations can push you into higher tiers or force batching to stay within quota.
- Monitoring & alerting. Watching for failed syncs, data drift and silent errors, and catching regressions when an upstream system changes a field.
- Maintenance. Salesforce ships three releases a year and your other systems evolve; mappings, Apex and flows must move with them, or the integration quietly rots and starts dropping or mis-routing records.
A useful planning rule: budget ongoing costs of roughly 15–25% of the build per year for a custom integration — or $10k–$45k a year in support and maintenance on a mid-sized program — plus any platform licences on top. Model the licence cost against your real data volume before you commit; it is the number most likely to bite at scale.
Fixed fee, time & materials, or license-plus-implementation?
Once you have a cost band, the next question is how the quote is actually structured — and the three answers above are priced differently enough that it's worth checking before you compare two proposals side by side.
Fixed fee vs time and materials
A point-to-point connector or a well-scoped custom API build is usually quoted as a fixed fee, because the systems, objects and volume are known upfront — this is the lower-risk option for the buyer, and the one we default to once discovery has mapped the data. Time and materials is the honest model for work that cannot be fully scoped in advance: a messy data-quality discovery, an enterprise program where requirements will shift, or any build where the other system's API documentation turns out to be wrong once you start. A credible partner will tell you which model fits before you ask, not pick T&M by default because it's lower-risk for them.
License cost vs implementation cost — budget them separately
Every iPaaS or connector-platform quote bundles two different numbers that behave differently over time: the license (MuleSoft, Boomi or a connector subscription, billed monthly or annually, scaling with capacity) and the implementation (the one-time build, mapping and testing cost). Treating them as one figure is how teams get surprised — the implementation cost is paid once; the license cost is paid every year the integration runs, and renewal pricing is where vendors recapture the discount they gave you at signing. Ask for both numbers separately, and ask what the license renews at in year two before you sign a multi-year implementation contract.
Data migration and compliance multipliers
Two factors move a Salesforce integration budget more than almost anything else, and both are easy to underestimate at the proposal stage.
Data migration
Almost every integration project drags a one-time migration along with it: moving historical records out of a legacy CRM or spreadsheet and into Salesforce, cleansed and de-duplicated. Budget $5k–$60k depending on volume and how messy the source is. The cost is not in the move itself — it is in the cleansing, the mapping and the validation that stop you importing a decade of duplicates and broken references. Scope it as its own workstream, not a footnote to the integration.
Compliance and regulated data
Integrations that move regulated data — health, financial or personal data under HIPAA, GDPR or similar — typically cost 25–35% more. The extras are design constraints, not polish: Salesforce Shield (platform encryption, event monitoring, field audit trail), data-residency controls, PII handling, access controls and compliance reviews. Retrofitting any of these into a live integration is far more expensive than designing for them, so scope them on day one. If your data touches health records, our HIPAA software development checklist covers the obligations that apply.
How to reduce your integration cost
None of these tactics cut corners on reliability — they change the order and shape of the spend, and together they typically save 20–40% against the naive version of the same project.
- Start point-to-point, add middleware only when you cross ~3 systems. Paying for an iPaaS platform before you have the second or third integration to run on it is the single most common overspend we see (see point-to-point vs middleware above).
- Fix data quality before scoping the build, not during it. Vendors price for the mess they actually find in discovery. A week of deduplication and mapping before the quote is cheaper than the change orders that follow a mid-project data surprise.
- Reach for declarative tools before custom Apex. Flow, native AppExchange connectors and standard field mapping cover more cases than teams expect, and they are both cheaper to build and cheaper to maintain than bespoke Apex triggers.
- Negotiate multi-year middleware commitments. Discounts of 35–55% off MuleSoft or Boomi list price are routine with a volume or multi-year commitment — ask before you accept the sticker price, and size the commitment to validated volume, not a projected one.
- Ship an MVP integration first. Sync the handful of fields that actually drive a decision or a workflow before building the full bidirectional, every-field version. Most of the value of an integration comes from a small slice of the data.
How to choose an integration partner
General Salesforce admin skills are necessary but not sufficient for a production integration. This checklist separates teams who can ship a reliable, observable data layer from those who will demo a happy-path connector and stall at the first edge case.
1. Real integration and data engineering, not just admin
Ask how they handle failed syncs, ordering, retries and data reconciliation. A team that talks about idempotency, error queues and validation has shipped real integrations; a team that only talks about clicks-not-code has shipped configurations.
2. Honest point-to-point-versus-middleware advice
A partner who recommends a cheap point-to-point connector when that is genuinely the right call — rather than selling you a MuleSoft program you do not yet need — is one you can trust with the platform decision when it does arrive. Be wary of anyone who quotes enterprise middleware before counting your systems.
3. Senior Apex, not certified juniors
Integration logic lives in Apex, triggers and Flows, and it is where brittle work hides. Look for evidence of senior Salesforce engineering and a sensible approach to governor limits, bulkification and testing — not just a stack of certifications.
4. A migration and testing plan from day one
The data migration and the reconciliation testing are where projects overrun. A partner who scopes both upfront, with a cutover and rollback plan, is one who has done this before and been burned by skipping it.
5. Engagement model fit
An integration is a long-lived system that grows with your stack. A team that owns it over time usually beats a one-off handoff, and a scoped discovery should precede any fixed-price commitment — our guide on how to choose a software development company covers the full vetting process.
FAQ
How much does Salesforce integration cost in 2026?
From about $5,000 for a single point-to-point connector to $600,000+ for an enterprise multi-system program on middleware. A custom API integration linking one or two systems is typically $20k–$60k, and a middleware/iPaaS program connecting several systems usually runs $40k–$150k to build, plus platform licensing. The biggest drivers are the number of systems, data volume and quality, real-time versus batch, custom Apex and compliance — not the Salesforce licences.
What drives the cost of a Salesforce integration?
The number of systems (roughly $10k–$50k each), the volume and cleanliness of the data, whether the sync is batch or real-time, the amount of custom Apex and Lightning work ($100–$200/hour), and compliance — regulated data and Salesforce Shield add 25–35%. A connector to one tidy system is cheap; a real-time, bidirectional sync across messy enterprise systems is not.
Is point-to-point or middleware (iPaaS) cheaper for Salesforce integration?
For one or two simple connections, point-to-point is cheaper to start — no platform licence and less to build. Middleware such as MuleSoft or Boomi wins once you connect three or more systems, need reusable APIs, monitoring and governance, or run a real-time bidirectional sync. Below ~3 integrations, point-to-point usually wins on cost; above it, the maintenance tax of tangled links makes middleware cheaper over the life of the system.
How much does MuleSoft cost for Salesforce integration?
MuleSoft is licensed by capacity, not per integration. In 2026 the median buyer pays around $55,000/year; mid-market deployments are commonly $50k–$150k and large enterprise programs run $250k–$600k+ a year once implementation and specialists are included — roughly two to three times the base subscription in year one. Discounts of 35–55% are achievable. It pays off with many reusable integrations; for one or two connections it is usually overkill.
What are the ongoing running costs of a Salesforce integration?
Middleware or iPaaS licensing (the largest line if you use MuleSoft or Boomi), hosting for custom services, monitoring and alerting, and maintenance as Salesforce's three annual releases and your other systems change. Budget roughly 15–25% of the build per year, or $10k–$45k a year for support on a mid-sized program, plus platform licences. An integration is a living system, not a one-off project.
How long does a Salesforce integration take?
A single point-to-point connector to a clean system can ship in 2–4 weeks. A custom API integration across one or two systems typically takes 6–12 weeks. A middleware program with real-time sync, data migration and proper testing runs 3–6 months, and a full enterprise integration across ERP, a data warehouse and multiple clouds can take 6–12 months. Data cleansing, mapping and testing — not the connector wiring — set the timeline.
How much does it cost to integrate Salesforce with a specific system like DocuSign, Slack or NetSuite?
It depends heavily on which system. Lightweight, well-documented connections — Slack, Outlook or DocuSign — typically run $500–$10,000 using existing APIs and AppExchange packages. Mid-complexity systems — marketing automation, payment gateways or e-commerce platforms — run $5,000–$40,000 depending on data volume and custom logic. ERP systems such as NetSuite, SAP or Dynamics are the most expensive category, typically $15,000–$100,000+, because they carry the most data, the strictest data-integrity requirements and the most custom mapping. See the cost-by-system table above for the full breakdown.
How can I reduce my Salesforce integration cost?
Start point-to-point and only move to middleware once you cross roughly three systems; fix data quality before scoping the build; use declarative tools (Flow, native AppExchange connectors) before reaching for custom Apex; negotiate multi-year middleware commitments, where 35–55% discounts are common; and phase a first integration as an MVP on core fields rather than a big-bang enterprise rollout. See how to reduce your integration cost above for the full list.
How much does it cost to integrate Salesforce with Infor?
Dedicated Infor-Salesforce connector platforms typically start around $500–$600/month for a single integration pair. A simple sync — for example forwarding orders to the back office and pulling stock quantities to the front office — usually runs $15,000–$30,000, while a multi-workflow build with several endpoints can reach $30,000–$100,000. A fully managed onboarding program commonly lands near $40,000–$50,000 in the first year once setup and onboarding are included. See ERP vendor cost differences above for Epicor, SAP and NetSuite compared.
How much does it cost to integrate Salesforce with Epicor or SYTELINE?
It depends on the route. Connector-platform licensing for Epicor Kinetic commonly runs $175–$400 per connection per month (roughly $2,000–$6,000 a year) before setup fees, which fits well-documented, lower-complexity syncs. A fuller bidirectional integration with field mapping, error handling and testing still falls in the general custom-API band of $20,000–$60,000, and an older SX.e or SyteLine install on thinner APIs tends to land toward the top of that range.
Is a Salesforce integration priced as a fixed fee or time and materials?
A well-scoped point-to-point connector or custom API build is usually fixed fee, because the systems and data volume are known after discovery. Time and materials fits work that genuinely cannot be scoped upfront — messy data discovery, enterprise programs with shifting requirements, or an ERP whose API turns out not to match its documentation. Ask which model a quote uses and why before comparing two proposals on price alone — see fixed fee vs time and materials above.
Last updated 21 November 2026. Cost and timeline ranges reflect integration-complete builds for US and EU clients and will vary by scope, systems, data volume and quality, real-time requirements and compliance. Figures are planning guidance, not a quote — request a scoped estimate for your specific integration. Regulatory references (HIPAA, GDPR) are general guidance, not legal advice.