Embedded analytics 2.0: Your secret weapon to empowering frontline workers and locking in customers loyalty

Last year, Harvard Business Review and ThoughtSpot published a groundbreaking survey on the business benefits of empowering frontline workers with data. Revenues are higher, operations more efficient, customer service better, and employees happier. And yet, few organizations deploy BI this way, historically held back by the technology, conflicting priorities, and mindset.

Enter embedded analytics 2.0, a new category that combines search with low-code development platforms to finally bring insight to action at the point of decision-making within an operational app. Further, while much of analytics and BI started as internal use cases, data savvy organizations are recognizing data apps as both a revenue stream and differentiated service to increase customer loyalty. 

Embedded analytics 1.0

When I think of the BI vendors that have differentiated on their embedded capabilities, I think primarily of Logi Analytics, Sisense, and Looker.  Others have had starts and stops such as GoodData or Izenda, but no one has been able to solve the problems of internal BI and embedded analytics use case within the same platform. The newly released ThoughtSpot Everywhere changes this.

Early embedded platforms inherited some of the same underlying weaknesses of first generation BI platforms. As reporting tools were displaced internally by visual-based data discovery, customers looked to embed visual discovery capabilities. The market became divided between report-centric vendors with SDKs for embedded reports and visual-based data discovery tools lacking in tools for developers. But in both cases, the final experience was equally limiting for users, constrained to parameterized reports or dashboards. If a manager is looking at an embedded expense dashboard that exceeds plan, they want to drill down, compare to other teams, or compare to the previous quarter. Embedded dashboards became a dead end, because whether built visually by a BI expert or programmed by a team of report developers, the analytics process would be predefined and the manager would have to go seek help from the BI team to go any further.

Another limitation: the range of functionality exposed via open APIs varied significantly from vendor to vendor. Rarely were all the components of the underlying BI platform accessible via APIs. This often resulted in both reduced functionality and a less seamless interface.Lack of a developer playground to discover and test new APIs was another limiting factor. 

Embedded analytics 2.0

From a technology perspective, APIs via low-code development platforms allow for a high degree of customization with a low degree of effort. And yet, product managers still need to remember that only what is in the core analytics platform can potentially be exposed via an API. In other words, data product managers and analytics and BI leaders should collaborate to understand differentiators of a particular analytics platform.  Here, ThoughtSpot’s key differentiator of search means “search as a service” can now be embedded in any app. Ditto for insight-to-action; developers can create and embed actionable and highly interactive capabilities in any data app. This is dramatically different from the capabilities in a dashboard-centric BI platform that requires pre-defined parameters and anticipating users' questions. ThoughtSpot Everywhere also includes a full developer playground to trial customizations and APIs.

Next generation embedded analytics is also born in the cloud, meaning the larger cloud ecosystem can be connected and leveraged throughout the app experience. This allows the insight-to-action workflow to be re-imagined, and re-engineered.

Before

Basic Embedded

Modern Embedded Analytics

Expenses exceed budget

The manager exports the report showing the variance, then emails the employee about the excess. Next month, the manual process is repeated, with no visibility into  past issues.

The manager collaborates within the data app, tagging the employee via Slack or Gmail. The manager creates an alert for this employee and adds to a list to monitor. 

Customer contact

A sales person accesses an aggregated report of customers by segment who have not engaged via webinars or opened emails in the last 30 days. A custom report with email addresses is sent to an SDR to manually email or call these customers.

A sales person accesses a personalized pinboard of their customers with low engagement and generates a personalized outreach to schedule a check-in with full calendar integration.

The approaches to embedding and white labeling have improved significantly but the unique selling point of search and AI in embedded is also key here.  Some of ThoughtSpot’s largest customers and highest impact use cases were for customer-facing analytics that bring the power of search to external customers. This is where customers like British Telecom, Carlson Wagonlit Travel have received industry awards.  DRC, TPICAP, Just Eat, and Verisk tout both large user bases as well as measurable business benefits.  At the 2021 Gartner Data and Analytics summit, TP ICAP, Parameta cited increased revenues from its data app. Verisk in a recent webinar cited increased revenues, improved customer experience and a 10x increase in their proof of technology requests for their data app.

Conflicting priorities

Embedded analytics, customer facing analytics, and internal self-service analytics often compete for the same limited resources.  Analytics and BI teams naturally will start by focusing on the most pressing needs – this includes responding to who shouts the loudest. These are the hungry data analysts. Rolling out self-service analytics to non-technical business users is both a push (let me enable you and upskill you) as well as a pull (when a business user is fed up with slow access to data and impatient about the dashboard back log.) Are operational users and frontline workers even asking for embedded analytics? They often muddle through with slow operational reports. And yet giving access to embedded analytics that empower the micro decisions may yield a bigger ROI.

Just as every organization is aspiring to be data-driven, this view of sharing data has to be re-imagined across the full value chain. Data monetization is certainly part of this, but I suspect the bigger value proposition is in winning customer loyalty by optimizing insights across the full value chain. This is where organizations like Just Eat and OpenTable share data with their customers to advise them of the busiest times to better adjust staffing levels, or what are the most popular food products and more. Retailers who share sales data with their suppliers ensure a more efficient supply chain and higher sales by having the right inventory available. 

In this regard, companies need to rethink how they fund and prioritize analytics projects. Very few do a formal ROI for funding or to assess if value was delivered. However, even a back-of-the envelope ROI is a useful guide in prioritizing competing projects. Further, analytics needs to move from being viewed primarily as a cost center to more a value-driver.  This is where a shift to more line of businesses funding initiatives will force value-based prioritization.

Mindset: Build vs. buy

Build versus buy is a constant debate in analytical applications and it reaches a crescendo when it comes to customer facing data apps. The arguments for building a customer facing data app:

  • Full control over the user experience

  • Consistent branding within a larger app

  • No vendor lock-in

I would say the first two items are largely addressed with embedded 2.0.  This third bullet is more difficult to overcome.  Pick your embedded partner wisely and know the product roadmap.

The arguments for “buying” an analytics platform to deliver a customer-facing data app:

  • Faster time to market

  • Robust, out-of-the-box functionality

  • Ability to focus on core competencies that are differentiators

I recall a few years ago working with an embedded customer who wanted to add an export to PDF option for their analytics app. It would take the development team nine months to get to this enhancement request that was widely available from BI vendors and not in the least differentiated. This was the a-ha moment when they realized they needed to buy instead. In terms of robust, out-of-the box functionality, be sure to evaluate the degree that all the functionality of the core analytics platform is indeed available via APIs.

Your game plan for capitalizing on embedded analytics

As you look to broaden your reach of analytics to front-line decision makers and customers and suppliers:

  • Understand the range of use cases where embedded can have an impact to democratize data, link insight to actions, and enhance customer loyalty.

  • Think beyond simple data monetization to creating a value added and differentiated analytics-as-a-service.

  • Evaluate analytics vendors embedded functionality including open APIs and developer environment.

  • Investigate what functionality from the core analytics platform can be deployed in an embedded data app.

  • Address objections from developers who prefer to build apps from scratch, being realistic about pride of development and weighing the risks and benefits of buy against the faster-time-to-market and robustness of a build approach.

Thank you to Dave Eyler, Senior Director, Product Management, Bryant Howell, Technical Architect, Andrew Yeung, Senior Director, Product Marketing for contributing to this blog.