Self-Service Analytics

The Power of Dataset Customization

⏱ 6 min read

Self-service embedded analytics is a powerful promise: give your customers the ability to explore, visualize, and act on their data on their own terms. But that promise is only as strong as the data experience behind it.

While many platforms advertise self-service, they quietly limit what users can actually do—especially when it comes to customizing the datasets. And that’s a problem. In SaaS, where every customer has unique metrics, naming conventions, and workflows, the ability to tailor datasets isn’t optional—it’s essential. Yet many vendors don’t offer this capability because their multi-tenant architecture wasn’t built to support it. That’s not just a technical limitation—it’s a strategic failure.

The heart of self-service: customizing the datasets

Most analytics platforms offer some surface-level interactivity—filtering, drilling down, maybe even rearranging widgets in dashboards. But true self-service goes deeper. It means giving users the ability to:

Rename fields to match their business language
Create calculated fields or metrics
Build entirely new datasets from existing ones
Tailor data structures to reflect their workflows

This level of control transforms analytics from a static reporting layer into a dynamic decision-making engine. It’s what separates a good embedded analytics experience from a great one.

As outlined in our Embedded Analytics Evaluation Guide, dataset customization is a powerful differentiator. It enables customers to shape your application around their business—not the other way around—enhancing both flexibility and satisfaction.

Why most vendors cannot offer dataset customizations

Here’s the uncomfortable truth: many analytics vendors don’t support dataset customization in multi-tenant environments. And it’s not because they don’t want to—it’s because their architecture won’t let them.

Most general-purpose BI platforms were designed for internal, single-tenant use. Multi-tenancy was bolted on later, often through workarounds like:

Creating a separate data model (e.g., Elasticube) per tenant
Using rigid namespace structures tied to cloud accounts
Limiting tenant access to pre-built dashboards only

These approaches create trade-offs. For example, Sisense requires a separate Elasticube per tenant to enable dataset customization, which significantly increases management complexity and cost. Looker and ThoughtSpot don’t offer embedded dashboard builders at all in multi-tenant mode.

As a result, SaaS companies are forced to choose between operational efficiency and customer empowerment—a choice they shouldn’t have to make.

The consequences of limiting dataset customization are not trivial

When dataset customization is restricted, the impact goes far beyond a missing feature—it undermines the entire self-service promise. Customers are forced to work with rigid, pre-defined data models that don’t reflect their business logic, terminology, or KPIs. This leads to frustration, disengagement, and a reliance on your support or engineering teams for every minor change. Over time, this bottleneck slows down customer insights, inflates operational costs, and erodes trust in your analytics offering. Worse, it signals to customers that your product isn’t built for their growth or complexity. In a competitive SaaS market, that’s a risk you can’t afford.

What to ask when evaluating platforms

If you’re evaluating embedded analytics solutions, ask these questions:

Can tenant users create and manage their own datasets?

Are custom fields, metrics, and naming supported at the tenant level?

Can users customize datasets directly within the embedded experience—or are they forced to switch to a separate vendor platform to do so?

Does the platform support co-mingled data models with secure, isolated customization?

If the answer to any of these is “no,” you’re likely looking at a platform that wasn’t built for SaaS.

Dataset customization isn’t a bonus – it’s the baseline

Dataset customization isn’t a bonus—it’s a baseline for modern SaaS analytics. If your platform can’t support it, your customers will feel it. And so will your team.

Platforms that support tenant-level dataset customization unlock a more scalable, satisfying, and differentiated analytics experience. They empower your customers to move faster, reduce your support burden, and help your product stand out in a crowded market.

To help you evaluate platforms with this in mind, download our Embedded Analytics Evaluation Guide. It outlines the four critical areas to assess—self-service, data management, deployment, and embedding—and includes specific questions to uncover architectural limitations before they become your problem.