ERP and the API Economy

Shopify's market cap just crossed 100 billion dollars. Stripe is the most valuable fintech company in the US. Twilio went public three years ago and is still growing. These companies don't have traditional products. They have platforms built on APIs. They won because they were designed API-first.

The ERP industry is watching this and panicking.

SAP, Oracle, NetSuite—these are monoliths. You implement them on-premise or in the cloud. You customize them heavily. You integrate them with other systems through middleware that costs as much as the ERP itself. The architecture is: ERP is the center of the universe. Everything else connects to it. This model is starting to look fragile.

What's happening now is what happened to databases. Twenty years ago, there were hundreds of database companies. They died because a few platforms (relational databases) became the default, and the economics shifted to software-as-a-service. Today there are thousands of database systems, but the successful ones are API-first: provide access to your data through well-designed interfaces, and you win.

ERP is following the same arc. The companies that will dominate ERP in the next decade won't be the ones with the fanciest UIs or the most modules. They'll be the ones that have built APIs first, and made it easier for customers to extend and integrate them than it is to go somewhere else.

Why This Matters Immediately

Let me ground this in a real problem I see constantly. A mid-market manufacturing company uses NetSuite. They have good financial reporting. They have decent CRM integration. But their warehouse management system is disconnected. Their quality management system is disconnected. Their production scheduling is disconnected. These are critical systems for their business, but they're data silos.

The company wants a unified view. They want production schedules to impact material planning in the ERP. They want real-time inventory visibility. They want quality issues to auto-create purchase orders for rework.

With a traditional ERP, this means hiring systems integrators. It means custom middleware. It means ongoing maintenance. The integration costs two to three million dollars, takes a year, and breaks whenever either system updates.

With an API-first ERP, the warehouse system talks to the ERP through clean APIs. Production scheduling updates the ERP directly. Quality issues trigger workflows in the ERP through webhooks. The integration is maintained by the vendors, not your IT team.

The difference in cost, time, and ongoing complexity is massive.

What API-First Actually Means

This isn't just "having an API available." NetSuite has an API. Oracle has APIs. But having an API and being API-first are different things.

API-first means:

• The API is the primary way the system is used, not an afterthought
• The UI is built on top of the same API that external developers use
• Rate limits and authentication are designed for thousands of external integrations
• Documentation is comprehensive and actively maintained
• The business model accounts for API consumption

Most legacy ERPs fail here. The API is a second-class citizen. The UI is built in ways that bypass the API entirely. The rate limits are low. The documentation is sparse. The business model is per-user licensing, not consumption-based.

Stripe gets this right. Everything Stripe does programmatically, you can do through their API. Everything you can see in the dashboard, you can access via API. They rate-limit based on usage, not on whether you're a "premium partner." The docs are world-class. The business model aligns with API consumption.

NetSuite sort of gets this. You can do a lot via API. But the UI doesn't use the API. Custom UX extensions are harder than they should be. Rate limits are restrictive unless you negotiate. The business model hasn't adapted.

Workday doesn't really get this. Workday has APIs, but they're designed for system integrators, not for customers building their own extensions. The culture is still "buy the package, use it as-is."

What's Changing

The arrival of best-of-breed SaaS is forcing this issue. Twenty years ago, you'd implement NetSuite and accept that it was your entire back office. Today, you implement NetSuite for financial accounting, Shopify for e-commerce, Stripe for payments, HubSpot for CRM, Guidepoint for HR, Fishbowl for inventory. These systems are best-in-class in their categories.

For this to work, you need the core financial system (the ERP or whatever you're calling it now) to be a hub that connects easily to everything else. An API-first architecture makes that possible.

The companies recognizing this are building new ERPs. Accrual, Airbase, Stripe's own financial platform—these are all designed API-first. They're not trying to replace NetSuite in every dimension. They're trying to be great at financial operations and integrate deeply with the best tools in each category.

Some of the legacy vendors are adapting. SAP bought Concur and is building more integration-friendly interfaces. Oracle is improving their API story. But the cultural shift is slow. When your revenue comes from implementation fees and per-user licensing, API-first integration actually threatens your business model.

The Real Opportunity

What this means for businesses right now: if your ERP isn't API-first, think about that as a long-term risk. Not immediately. But in three to five years, the cost of integrating systems will be a competitive factor. The companies with systems that integrate cleanly will move faster and have lower total cost of ownership.

If you're evaluating ERPs, ask about APIs. Not "do you have an API?" but "can I build my entire workflow through your API?" Can your system be extended without touching the vendor's code? Can you integrate with other best-of-breed platforms easily? How does the vendor price API usage?

If you're building a business around ERP—implementing it, extending it, building integrations—the winners in the next decade will be the ones who can orchestrate multiple systems as if they were one platform. That's only possible if the underlying systems are API-first.

The API economy has already remade databases, infrastructure, and payments. ERP is next. The platforms that adapted first will have a huge advantage. The ones that resist will slowly become specialized systems for companies that can afford to run legacy software.

I think we're at the inflection point right now. By 2021, API-first will be table stakes for any new ERP. The interesting question will be how quickly the incumbents adapt before they become dinosaurs.