Self-Service AND Embedded Analytics: Is It Possible?
We all know the phrase “opposites attract,” but in reality, getting opposites to work together is often pretty hard. When it comes to self-service business intelligence (BI) and embedded analytics, it’s hard to accomplish both in a single system. Software-as-a-service (SaaS) providers have felt this pain more than most other industries.
In the early days of business intelligence software, end users were rarely able to create their own dashboards. They were forced to rely upon either pre-built reports or enlist system experts to create reports on their behalf. Then the day came when users could finally create a customized view of their data, and self-service BI was born. But, for those that remember, this was not a simple transition. It created a window where new vendors emerged to solve this challenge.
As I’ve said before, reinvention is always risky and incumbents often struggle to evolve. This phenomenon led to the emergence of new companies like Tableau. It was no small feat for Tableau to compete with the established players, but its founders knew users needed a platform built from the ground-up for self-service.
Tableau’s success gave rise to several other me-too BI players, all of which are fighting for a piece of the internal analytics market within enterprises. While these companies all claim to do something different, they are, in fact, still built for internal, self-service usage. All of this competition has led the BI industry to another significant shift, one that promotes embedded analytics from a feature to an architecture.
The need for customized reports and dashboards is no longer a requirement that is limited to the average internal analyst. With the rise of SaaS applications over the last decade, a whole new class of end user has risen. These end users of SaaS applications are tired of static reports and they now have the same demands for customization that analysts had in the mid-2000s. The problem: all of the existing BI tools weren’t built to be embedded into multi-tenant SaaS applications. They have struggled to pivot, reinvent themselves or even truly understand the needs of product and development teams. Embedded analytics, as it turns out, tends to be the “third rail” of traditional internal analytics.
Providing quality embedded analytics takes focus. SaaS providers require embeddable software that is extremely flexible. They can have hundreds or even thousands of customers, each with potentially hundreds or thousands of end users themselves. Each of those customers has different requirements for data and user security, licensing, data models, and customizability. This creates sets of unique requirements that simply cannot be met with traditional BI software. In a previous post, I talked about why SaaS Providers Have Unique Analytics Needs. That article takes a deeper look at the unique requirements of SaaS providers.
Each of these major capabilities takes years and multi-millions of dollars to accomplish, refine and perfect. Building each of these capabilities also requires very different architectural decisions. It is extremely difficult to build software that targets both internal and embedded analytics use cases at the same time. This is why so many BI vendors went down the path that was their highest payoff at the time — providing self-service capabilities. They treated embedded analytics merely as a feature of a system that was built for something else. What SaaS providers need, however, is a system with an architecture built for SaaS.
For the longest time, SaaS providers have been forced to make compromises. They’ve had to provide self-service analytics to their customers while internally struggling to efficiently embed those capabilities into their own software. While BI vendors were certainly aware that SaaS providers had (and still have) a need to offer advanced analytics capabilities with multi-tenant abilities, their offerings simply missed the mark. SaaS application providers paid the price with subpar user experiences and overly cumbersome development integration.
When Qrvey started in 2016, we made the architectural decision to provide self-service capabilities AND embedded analytics from day one. This was a risky decision for us, both on the engineering side, where it meant risking an extremely long time-to-market, and on the marketing side, where we worked on the assumption that the SaaS provider market would grow significantly. But in the end, a platform based on today’s newest cloud-native technologies, such as serverless and microservices, has proven to be a winning architectural bet that has paid off. Now SaaS providers really can have their cake and eat it too. I invite you to learn more about Qrvey and see what modern analytics can offer.
