Mar 27, 2026

Build vs Buy: What a Custom Power BI Embedded Portal Really Costs (2026)

ByPascal Kiefer

Build vs Buy: What a Custom Power BI Embedded Portal Really Costs

If you've decided to deliver Power BI analytics to external users, partners, or customers, you're facing a fundamental question: do you build the infrastructure yourself, or do you use a platform like Embedsy? The API is free and well documented, which makes building look cheap. This post walks through what building actually involves, puts realistic numbers on it, and gives you a checklist to decide which path makes sense for your situation.

What building your own solution actually involves

Power BI Embedded is Microsoft's API for embedding Power BI reports. It's powerful, but the API is just the starting point. A working sample that renders one report with a service principal is an afternoon's work. A portal that real customers log in to is a product, and you need to build everything around the API.

Authentication and access control

You need to handle how users log in, how their identity maps to Power BI access, and how you enforce Row Level Security (RLS) at scale. This typically means integrating Microsoft Entra ID (and often a second provider for customers outside the Microsoft world), managing service principals, and writing custom logic to generate embed tokens per user with the right RLS roles. Tokens expire, so you also need silent refresh. It's not trivial, and it needs to be secure from day one.

Multi-tenancy

If you're delivering analytics to multiple customers or partners, each needs to see only their own data. Building proper tenant isolation, at the data, report, and access level, is one of the more complex engineering challenges in this space. A misconfiguration can expose one customer's data to another, and it is the kind of bug that surfaces in a security questionnaire rather than in testing.

Branding and portal experience

Reports embedded via API appear inside whatever interface you build. That means designing and building a full frontend: navigation, user management, branding controls, responsive layouts, a custom domain with certificates. If you want it to look professional and feel like a real product, expect significant frontend effort. Our guide to white-label Power BI portals lists everything a customer-facing portal has to cover.

Administration

Someone needs to manage users, assign roles, control which reports are visible to whom, and handle onboarding. Without an admin interface, that means direct database edits or custom tooling, which becomes a support burden as you scale. Admin screens are rarely in the first estimate and always in the second.

Capacity management

Power BI Embedded runs on Azure capacity, which bills for every second it runs. Managing that capacity (pausing it when idle, starting it when a user arrives, scheduling refresh windows, monitoring load, avoiding a surprise invoice) requires its own tooling and ongoing attention. Our pricing guide shows how much this alone can move the bill.

Ongoing maintenance

Microsoft updates Power BI and Azure regularly. Your custom solution needs to keep up: API changes, security patches, dependency updates, and feature requests from your users. This is an ongoing engineering cost that's easy to underestimate upfront.

What it costs to build, in practice

Every project is different, but after several years of building these systems for customers we see the same shape. The numbers below assume a small, competent team that has not built a Power BI Embedded portal before.

Phase	Typical effort	What's in it
Proof of concept	1 to 2 weeks	One report, one user, service principal, embed token. This is the part that makes building look easy.
Minimum viable portal	3 to 6 months, 1 to 2 developers	Login, roles, RLS per user, multi-tenant separation, branding, admin screens, capacity pause and resume, custom domain, testing.
Hardening	1 to 2 months	Security review, penetration test findings, logging, monitoring, backup, the customer's IT questionnaire.
Ongoing	15 to 25% of the build, per year	Platform updates, dependency upgrades, feature requests (scheduling, exports, public embeds), support.

Ranges from our own project experience. Teams with prior Power BI Embedded experience land at the low end; teams also building an application around the analytics land above it.

At typical European or North American developer rates, the minimum viable portal alone lands somewhere between the mid five figures and low six figures before anyone has used it, and the ongoing line never goes away. That is the number to compare against a platform subscription, not the price of the API.

When building makes sense

Building your own solution is the right call in specific situations:

You have very specific requirements that no off-the-shelf platform can meet: deep integrations, unique UX patterns, or proprietary infrastructure constraints.
You have a dedicated engineering team with Power BI Embedded expertise and the capacity to maintain the solution long term.
You're building at very large scale where the economics of a platform subscription don't make sense compared to in-house development costs.
Analytics is a core part of your product and the embedded experience needs to be deeply integrated with your existing application logic.

When using a platform like Embedsy makes more sense

For most teams delivering Power BI analytics, the build path is slower and more expensive than it first appears. A platform makes sense when:

You want to go live quickly, days rather than months. Embedsy is production-ready from day one, with authentication, multi-tenancy, branding and administration already built.
Your team's strength is analytics, not engineering. You want to focus on the reports and insights, not the infrastructure around them.
You need to keep costs predictable. Platform pricing is transparent; custom development costs are not.
You're a consultancy or partner delivering analytics to multiple clients. Embedsy lets you productize your offering without building and maintaining separate infrastructure per client.
You want your data to stay in your environment. Embedsy deploys inside your Azure tenant, so you keep full control and ownership.

The third option: buy the platform, build the edge

Build versus buy is rarely all or nothing. The pattern that works for most of our customers is to take the platform for the 90% that is the same in every portal (login, roles, tenancy, branding, capacity, administration) and build only the part that is genuinely theirs. In the Embedsy Portal that happens in two ways. Embedsy Studio lets your own team add custom pages, input forms, writeback, D3 and Vega visuals and AI agents on top of your semantic models without leaving the portal. And where a requirement doesn't fit, our custom development team builds it, either as a core feature or as an addition for your installation. You get the speed and cost profile of buying with the differentiation of building.

The full comparison

	Build Internally	Embedsy Portal
Time to launch	Months of development	Go live in days
Upfront cost	Mid five to low six figures	None; from $80 per month
Authentication & access control	Custom implementation required	Built-in and production ready
Multi-tenant architecture	Complex to design and maintain	Included out of the box
White-label analytics portal	Requires frontend development	Ready immediately
Capacity cost control	Build your own pause and schedule logic	Auto-pause and schedules included
Maintenance & upgrades	Ongoing engineering effort	Managed platform layer
Security & governance	Must be designed internally	Proven deployment model
Custom functionality	Unlimited, at full cost	Embedsy Studio and custom development

A quick decision checklist

Answer honestly. Three or more "yes" answers in the first group and building is defensible; otherwise buy, and build only the edge.

Build if:

Analytics is the product, not a feature of it.
You have developers who have shipped Power BI Embedded before.
Those developers will still be assigned to this in two years.
You need UX that no portal product can express.
You expect a scale where per-user platform pricing exceeds a full-time engineer.

Buy if:

You need to be live this quarter.
Your team's strength is data and reporting.
Your customers will ask for a security questionnaire and you'd rather answer with a proven deployment.
You want the capacity bill managed for you.
You'd rather pay a predictable monthly fee than an unpredictable engineering one.

Bottom line

Building your own Power BI Embedded solution gives you maximum flexibility, but comes with real costs in time, engineering effort and ongoing maintenance. For most organizations, the right answer is to use a platform for the infrastructure layer and focus their energy on the analytics itself. If you're unsure which path fits your situation, we're happy to talk it through.

Book a free 20-minute demo