TapCub is live — analytics, insights and live chat, free on one platform
Glossary

Embedded analytics: definition, levels and permissions

Embedded analytics puts reports and metrics inside the software people already use. How it differs from self-service BI, when an iframe is enough, and why permissions and metric definitions — not charts — are the hard part.

Lin Zhou Reviewed by Chen Yu Also called embedded BI Last updated Nov 16, 2026 Glossary

Quick answer

Embedded analytics is the practice of putting reports, dashboards and metrics inside the software people already use — a product, customer portal or internal workflow — instead of sending them to a separate BI tool, with access controlled per user or customer.

What embedded analytics means

Embedded analytics is the practice of putting reports, dashboards and metrics inside the software people already use — a product, a customer portal or an internal workflow — instead of sending them to a separate BI tool. Microsoft's Power BI embedded documentation defines it the same way: embedding reports, dashboards and tiles into websites and applications, in two modes — embed for your organization (internal users) and embed for your customers (external users who never touch the BI tool itself).

A setup counts as embedded analytics when it passes a three-part test:

  1. The data appears where the work already happens. A client checks campaign results inside your portal, not in a spreadsheet you emailed last Friday.
  2. Visible scope is controlled per user or customer. Each viewer sees only the rows and modules they are entitled to — not everything the dataset holds.
  3. Metrics share one definition. "Conversion" means the same calculation in every tenant's view, because it is defined once, centrally.

That test also states the thesis of this page: putting data where users already are is the easy half. The hard half is permissions and metric definitions, not charts. Sharing numbers more widely is a data democratization decision; embedding is how you do it without losing control.

Embedded analytics vs self-service BI vs customer-facing analytics

These three terms overlap, and vendors use them loosely. The distinguishing questions are who looks at the data, where, and who decides what they can see.

ApproachWho uses itWhere they see itWho controls accessTypical example
Embedded analyticsInternal users or external customersInside your product, portal or workflowThe host application, per user or tenantA usage dashboard inside a SaaS admin panel
Self-service BIInternal analysts and business teamsA separate BI tool they log intoBI admins, by workspace or dataset roleA marketing team exploring campaign data in a BI workspace
Customer-facing analyticsYour customers (external)Inside your product or a shared reportYou, scoped per customer accountAn agency client portal showing per-client traffic reports

Two rules keep the terms straight. Customer-facing analytics is a subset of embedded analytics: every customer-facing dashboard is embedded somewhere, but an operations tool with internal metrics is embedded without being customer-facing. Self-service BI describes the opposite direction — instead of bringing data to the user's workflow, it invites internal users into the data tool to explore on their own. If your readers are clients rather than colleagues, you are designing an analytics dashboard for external consumption, and access scoping matters more than chart variety.

Embedding levels: iframe, SDK/API, white-label

"Embedded" covers a wide range of engineering commitments. The three common levels differ less in chart quality than in identity handling and tenant control.

LevelIntegration effortIdentity and authenticationMulti-tenancyBranding control
iframe / embed codeLow: copy an embed code, as in Tableau's Embed Views into Webpages or Power BI secure embedViewer needs a license or a signed URL; little per-user contextWeak: one view per link unless you add URL filtersVendor toolbar and styling remain
SDK / APIMedium to high: Power BI app-owns-data, Tableau Embedding API, Looker's embed SDKHost app authenticates users; SSO, embed tokens, connected appsRow-level security applied per embedded sessionComponents can be styled; filters and events exposed
White-labelHighest: SDK plus design-system workSame as SDK, plus full identity integrationTenant isolation and audit logging expectedVendor branding removed; the view looks native to your product

Analyst checklists treat these as thresholds, not options: Constellation Research's ShortList for embedded analytics requires iframe, SDK and API-first embedding, white-label support and multi-tenant deployment, alongside role-based access and row-level security.

Choose with if-then rules:

  • If the audience is small and read-only, then an iframe or a hosted share link is enough; do not build an SDK integration you will not maintain.
  • If each logged-in user must see only their own rows, then you need SDK/API embedding with row-level security — an iframe alone cannot do it.
  • If analytics is part of the product you sell, then plan for white-label and budget ongoing maintenance, not just the initial build.
  • If you have no developer capacity, then use your analytics tool's read-only sharing and be honest that it is sharing, not per-user embedding.

Permissions and multi-tenancy: the hard part

The engineering problem behind customer-facing analytics is isolation: customer A must never see customer B's data, even though both look at "the same" dashboard. Two mechanisms cover most designs, as Netguru's embedded analytics guide describes: tenant isolation (a separate project, dataset or schema per customer) and row-level security (RLS) (one dataset, with every query filtered by tenant ID or user identity). Constellation's threshold criteria list both, plus audit logging, as baseline requirements.

Worked example (Sample data). An agency sends monthly traffic reports to 20 clients. Today an account manager exports numbers, pastes screenshots and writes an email per client: 45 minutes each, so 20 × 45 = 900 minutes (15 hours) per month.

Switching to per-client read-only links:

  • One-time setup: 10 minutes per client to create the link, pick modules and set a password → 20 × 10 = 200 minutes.
  • Monthly upkeep: 5 minutes per client to check scope and password → 20 × 5 = 100 minutes per month.
  • Savings: 900 − 100 = 800 minutes ≈ 13.3 hours per month.
  • Payback: 200 ÷ 800 = 0.25 months — the setup pays for itself in about a week.

The time math is the easy part. The design work is the access-scope table you fill in before sending anything:

AudienceWhat they getModules openedProtectionVisitor list and sessions
Clients (×20)Read-only link to their own project onlyOverview, Pages, SourcesPassword + 90-day windowNever visible
Internal teamConsole loginAll modulesAccount permissionsVisible in the console
Public marketing link (optional)Read-only link, no passwordOverview only30-day windowNever visible

Count the decisions: 20 links × 3 open modules = 60 module-level scope decisions, each one a chance to expose something you meant to keep internal. Each link points at exactly one client project, so client A cannot reach client B; the password and the day-window limit shrink the exposure if a link gets forwarded. Finally, define "conversion" once — the same purchase goal in every project — so all 20 reports run the same calculation. That is the thesis in practice: the charts took minutes; the permissions and the metric definition are the actual project.

Common mistakes

  • One link for every client. If you send all clients the same dashboard link, then anyone can browse everyone's data. Fix: one project or dataset per client, or RLS that filters every query by tenant.
  • Treating the iframe as a security boundary. If you assume "it's embedded, so it's safe," remember links get forwarded. Fix: password-protect, limit the day window, and filter by tenant on the server — hiding a toolbar filters nothing.
  • Letting metric definitions drift per tenant. If each client's report computes "conversion" slightly differently, then numbers stop being comparable and trust erodes. Fix: define the goal or metric once, centrally, and reuse it everywhere.
  • Confusing a read-only link with an embedding SDK. If your requirement is "each logged-in user sees only their rows," then a shared link cannot deliver it — everyone with the link sees the same scope. Fix: SDK/API embedding with row-level security, or scope down to per-client links and say so.

How TapCub shares analytics today

First, the boundary: TapCub is not an embedded-analytics or white-label BI product. What it offers for sharing is a public read-only link, with no per-login-user, row-level view: everyone with the link (and its password, if set) sees the same scoped data. For the 20-client agency above that is still enough, with one TapCub project per client website:

  1. Create one read-only link per client project. In each client's project, open the Public sharing tab under Configuration › Settings and generate the link; see share a dashboard. Because the link belongs to that client's project, client A's link cannot show client B's data. Only owners and admins can turn sharing on or off.
  2. Apply the scope table. Allow the Overview, Pages and Sources modules (Audience and Bots are the other two you could open), add a password and cap the visible window at 90 days. The visitor list and sessions are not shareable at all; they stay in the console for your team.
  3. Rely on the scope, not on a hidden toolbar. Filters for a module you did not allow are dropped, so a client who edits the link cannot pull a slice that was not shared. The page is read-only and cannot change goals or filter definitions.
  4. Keep one metric definition. Define the purchase goal the same way in each project under Web analytics › Goals, so the conversion numbers you report to all 20 clients come from one rule.
  5. Give your own AI agent read-only access, if you use one. On the MCP tab of All-platform analytics › Operations › AI Assistant you copy the endpoint, and the key needs only the Read stats scope; the tools read analytics and do not write events or edit definitions. See connect AI agents with MCP.

Send each client the link and the password through separate channels, and the 800-minutes-a-month saving from the worked example holds. As the thesis says, the charts were never the hard part: the five decisions above are all about who may see what, and which definition they see.

Sources

Was this explanation clear?

We use your answers to fix and expand this page.

Lin Zhou

Product Experience Specialist

Writes about product experience, practical workflows, and growth, turning complex processes into clear steps for everyday decisions.

FAQ

Questions people ask

Short answers to the follow-up questions that come up most.

Still have a question?

Chat with the team behind TapCub — we usually reply within a few hours on business days.

What is embedded analytics?

Embedded analytics means placing reports, dashboards and metrics inside the application people already use — your product, a customer portal or an internal tool — instead of a separate BI login. Access is scoped per user or customer, and metrics share one central definition. Power BI documents both internal and customer-facing embedding modes.

What is the difference between embedded analytics and traditional BI?

Traditional BI asks users to leave their workflow and log into a separate tool. Embedded analytics brings reports and metrics into the application people already use, with access scoped to each user or customer. Power BI documents both modes: embed for your organization versus embed for your customers.

Iframe or SDK — which should you use to embed analytics?

Use an iframe or embed code for small, read-only audiences: Tableau and Power BI both offer copy-and-paste embed codes, but viewers still need a license or signed URL. Choose SDK/API embedding when each logged-in user must see only their own rows — that requires row-level security an iframe alone cannot apply.

Is embedded analytics secure for customer data?

It is, when isolation is designed in: a separate project or dataset per tenant, or row-level security filtering every query by tenant ID, plus audit logging. An embed is not itself a security boundary — links get forwarded — so add passwords, time windows and server-side tenant filtering rather than relying on the frame.

See this number on your own site

Add one script tag for web and apps, and the first reports fill in within minutes.

See your data clearly. Find your growth.

Every click, backed by data. Install one line of code and see your first numbers in a minute.

No credit card · Free plans for analytics and chat · Cookieless analytics