Skip to main content

What are you looking for?

Explore our services and discover how we can help you achieve your goals

Client Workspace

Embedded analytics in a SaaS product: build, buy or hybrid

Most SaaS teams should buy or split embedded analytics and build it fully in-house only when analytics is the product itself; the deciding test is where tenant isolation is enforced, which white-label limits customers will notice and who absorbs heavy queries. Netbase scopes the choice from those three tests, not from a vendor demo.

Book a solution review See the related solution

By Netbase Commerce and Platform Engineering Desk · Reviewed by David (CEO) · Updated · 11 min read

star

This guide is for CTOs and product leads of B2B SaaS companies who need customer-facing dashboards inside their product. It sits under Netbase's data engineering and platforms service and builds on the guide to multi-tenant SaaS architecture, which covers isolation for the product as a whole; this page applies it to analytics.

In this guide

What does embedded analytics have to do in a multi-tenant product?

Embedded analytics in a multi-tenant product has three jobs: show each customer only its own data, look and behave like the rest of the product, and answer queries fast without slowing other tenants. Failing any one of these is a security, support or performance problem.

The harder work sits under the charts: the shared data model, the rule that filters it per tenant, the identity handoff between your product and the analytics layer, and the load a large customer's report puts on shared infrastructure. Those parts decide the option.

Where must tenant isolation be enforced?

Tenant isolation for analytics must be enforced below the user interface, in the data layer or the analytics engine, so that a missing filter in one dashboard cannot expose another tenant's rows. A filter added in the front end is a display choice, not a security control.

The ladder runs from weakest to strongest; each step up adds cost, so pick the lowest step your contracts accept.

Isolation level How it works Known weak point Fits when
Interface filter The app passes a tenant filter to each chart One missed filter leaks data Never on its own for customer data
Row-level security in one shared model The database or analytics engine filters every query by tenant identity Privileged roles can bypass it; it must be tested per query path Many small tenants with the same schema
Schema or workspace per tenant Each tenant gets its own schema, semantic model or workspace More objects to deploy, migrate and monitor Fewer, larger tenants or custom fields per tenant
Dedicated store per tenant Each tenant's analytics run on separate storage or compute Highest running and release cost Regulated or enterprise tenants who require it

Two documented bypasses show why. In PostgreSQL row security, superusers and roles with the bypass attribute always skip the policies, and table owners skip them unless row security is forced on the table. In Power BI's embed-for-your-customers model, assigning users to roles in the Power BI service does not affect row-level security when an embed token is used; the identity supplied with the token decides what is filtered.

Which white-label limits will your customers notice?

White-label limits are the places where an embedded tool still behaves like someone else's product: sign-in, branding, exports and self-service editing. Test each one in a proof of concept with your own data and identity provider, not in the vendor's sandbox.

  • Identity handoff

    What to test
    Can your product sign the user in without a second login, and pass tenant and role attributes?
    Why it matters
    A second login breaks the product experience and invites shared accounts
  • Branding and domain

    What to test
    Can fonts, colours, empty states and error messages match the product; does any vendor name or domain show?
    Why it matters
    Error pages and loading states are where third-party branding leaks
  • Exports and scheduled reports

    What to test
    Do PDF, CSV and e-mailed reports respect the same tenant rule and branding?
    Why it matters
    Exports often run on a different path from the live dashboard
  • Self-service authoring

    What to test
    Can a customer build its own report, and on which fields?
    Why it matters
    Self-service widens the query paths that isolation must cover
  • AI query features

    What to test
    Does a natural-language query feature respect the same row rule?
    Why it matters
    A new query path is a new place for isolation to fail

Looker's signed embedding is one documented pattern for the identity handoff: the embed URL carries the user identity, permissions, models and user attributes, and a closed system keeps tenants from discovering each other's content.

Build, buy or hybrid?

Criterion Build in-house Buy an embedded platform Hybrid (buy the engine, build the interface)
Tenant isolation Yours to design and test on every path The vendor's model; you configure and test it The engine's model, called from your code
White-label fit Complete, at the cost of building every screen Limited to what the vendor exposes Complete on screens; the engine stays hidden
Query performance You own caching, pre-aggregation and limits The vendor's capacity model; plan for heavy tenants You own the cache; the engine owns the query plan
Ongoing work Every new chart type and export is backlog Vendor upgrades and contract terms Two layers to maintain and upgrade
Choose it when Analytics is the product, or needs are a few fixed reports Standard dashboards for many tenants, little front-end capacity Analytics must feel native but the semantic and security layer should not be rebuilt

The reverse split buys a platform for standard dashboards and builds only the views that differentiate the product. The Cube build-or-buy article is one vendor's case for the first split; treat vendor cost estimates as marketing until your team has priced its own build.

How do heavy tenants affect performance?

Heavy tenants slow everyone when analytics queries run on the same database as the transactional product. The AWS SaaS Lens calls this the noisy neighbour problem: one tenant's load on shared resources degrades the experience of others, and multi-tenant workloads are hard to predict.

Three controls keep it in check. First, run analytics on a replica, a warehouse or a separate store, never on the primary product database. Second, pre-aggregate the metrics most tenants open first and decide their freshness per metric, as the guide to streaming or batch latency per business metric explains. Third, set per-tenant query limits and timeouts, and move a tenant that keeps hitting them up the isolation ladder.

When should a SaaS team not build embedded analytics?

Do not build embedded analytics in-house when customers need standard dashboards, the team has no capacity to own a semantic layer and an export pipeline over the long term, or the main ask is "let us get our data out". In that last case an export or a data share into the customer's own BI tool may answer the need better than any dashboard.

Do not buy a platform either when a few fixed reports cover what customers use: server-rendered report pages with the product's own tenant rule are cheaper to run and easier to secure until customers ask for self-service reports.

How to decide and roll out

  1. List what customers ask for

    Collect requested reports and exports; mark which are standard and which differentiate the product.

  2. Pick the isolation level

    Choose the lowest step on the isolation ladder that contracts and regulators accept, and write down which layer enforces it.

  3. Design the shared data model

    Model metrics once, with tenant identity on every row, outside the transactional database.

  4. Run a proof of concept on two tenants

    Test identity handoff, row rules, exports and branding with real data from two tenants, including one with an unusually large data set.

  5. Test isolation on every query path

    Try to read another tenant's data through dashboards, exports, scheduled reports, self-service queries and any AI query feature, and keep those tests in continuous integration.

  6. Set query limits

    Add per-tenant timeouts, cache rules and slow-query alerts before release.

  7. Review the choice yearly

    Revisit it when tenant count, self-service demand or vendor terms change.

Where AI fits

AI helps around embedded analytics, not inside the isolation rule. An assistant can draft metric definitions, generate isolation test cases for each query path and summarise a dashboard in plain language. Netbase works with the major commercial and open-source AI tools and models, chosen per project.

Engineers still decide the isolation level and approve every row rule and query path; no model changes an access rule on its own.

What delivery record exists, and what does not

Netbase's ERP work includes cloud ERP platform development and API sync of product catalogues, inventory, orders and customer data between e-commerce stores and ERP systems. It has delivered AI projects for clients that are not named, including MLOps pipelines, and its named AI delivery is the 4over4 recommendation engine, which suggests products from browsing history and previous purchases.

Its security practices include role-based access control, MFA for admin dashboards, TLS in transit and AES at rest, and vulnerability scanning. Delivery runs in a six-step lifecycle from discovery and architecture planning to rollout and ongoing support, and most Netbase projects are delivered on fixed-price contracts. No published Netbase record describes an embedded analytics platform rollout for a client, and no performance, cost or time figure is claimed here.

Alternatives: other ways to give customers their data

  • Scheduled reports by e-mail. A fixed report per tenant; choose it when customers rarely explore.
  • Data export or data share. Let each tenant pull its own data into its own BI tool or warehouse; choose it when customers already run analytics teams and ask for raw data.
  • In-product report pages. A few server-rendered pages with the product's own tenant rule; choose it when needs are stable and few.
  • An API for metrics. Expose aggregated metrics through the product API; choose it when partners or customers build their own views.

The data catalog and lineage guide shows how to record which customer-facing reports read which tables.

Limits of this guide

  • The isolation ladder and the white-label checklist are planning aids; your contracts, regulators and penetration tests decide what is enough.
  • No tool is recommended here, and no vendor is scored; tools are named only where their public documentation illustrates a pattern.

Plan the next step with a Netbase consultant

Frequently asked questions

Buy, or buy the engine and build the interface, when customers need standard dashboards across many tenants. Build fully in-house only when analytics is what customers pay for, or when a few fixed reports cover the need and an embedded tool would add more risk than value.

Row-level security in one shared model is enough for many small tenants with the same schema, provided the query account cannot bypass it and every query path is tested. Larger or regulated tenants may need a schema, workspace or store of their own.

Usually the identity handoff, the exports and the error or loading screens, because they run on different paths from the live dashboard.

Run analytics on a replica, warehouse or separate store, pre-aggregate the metrics most tenants open, and set per-tenant query limits and timeouts. A tenant that keeps hitting the limits moves to a more isolated tier.

Next step

Netbase can review your tenant model and the reports customers ask for, and propose an isolation level and a build, buy or hybrid path: book a solution review.

Data engineering platforms for reporting and AI Data engineering platforms for reporting and AI

Netbase builds data pipelines and analytics foundations for companies whose data is spread across a store, an ERP, a CRM and spreadsheets. We collect it, clean it, model it and check its quality, so reports agree with each other and AI projects start from data people trust. Data engineering is a capability we are growing alongside the platforms we deliver.

Learn More
line
AI-ready SaaS development from a team that runs its own SaaS AI-ready SaaS development from a team that runs its own SaaS

Netbase provides SaaS development for founders and product teams to launch and scale multi-tenant subscription software with AI features customers will pay for. We build and operate our own SaaS platforms, Printcart and the AI-powered Cloodo workplace, and bring that experience to client products. Discovery fixes the MVP scope and schedule.

Learn More
line
SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules

A SaaS product accelerator is a set of reusable Netbase modules for accounts, billing, roles and integrations that helps founders and product teams launch subscription software faster, with room for in-product AI from the first release. Reusing these modules can shorten development time, and the approach is proven on Printcart, Netbase's own web-to-print SaaS.

Learn More
line

Work with Netbase remotely

Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.

Contact Netbase

Discuss a project

Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.
Project enquiries

project@netbasejsc.com

WhatsApp

+84 937 869 689

Office address

91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam

Get in touch

Tell us what you want to build, modernize, or operate.

Tell us what you want to build, modernize, or operate.

Contact Netbase