A serverless architecture that enables a unique pricing model differentiates Amazon QuickSight from other business intelligence platforms.
Four years after its introduction, however, it's still not considered an innovative analytics tool.
Amazon first made QuickSight generally available in November 2016 after more than a year of beta testing with 1,500 customers. At the time, it featured basic BI functionalities built around dashboards. In the years since then, Amazon has added machine learning and other augmented intelligence capabilities such as natural language processing.
And this week on Nov. 30 as Amazon kicked off AWS re:Invent, the tech giant's virtual user conference, the vendor added new embedded analytics capabilities and an additional pricing option to the QuickSight platform.
Analysts, however, say QuickSight's capabilities trail those offered by the more innovative BI specialists and other tech giants with more full-featured platforms.
"In terms of its user interface, it's a year or two behind in natural language capabilities," said Dave Menninger, research director of data and analytics research at Ventana Research. "It's lagging there. And the state-of-the-art is also ahead of it in terms of its application of machine learning to the analytics process. It's got good basic capabilities, but it doesn't seem like it's trying to lead."
Similarly, Boris Evelson, principal analyst at Forrester Research, said the capabilities of QuickSight, while improving, are not yet at the level of its competitors.
"It fell behind competition in augmented analytics, data preparation, richness of data visualizations and other key components of large enterprise BI platforms," he said. "Since then, QuickSight has been catching up, but competitors have not been resting on their laurels either so I suspect the gaps between QuickSight and market leaders remain."
Where Amazon QuickSight is unique, however, is in its pricing.
QuickSight is among the only BI platforms built natively for the cloud from the outset. Domo, among major vendors, is another, while Oracle and SAP are among those that began on premises and have since been redesigned to become cloud-first.
Dave MenningerResearch director of data and analytics research, Ventana Research
Because QuickSight is cloud-native and doesn't require a server, Amazon is able to charge far less for its use than competitors do for similar offerings and does so with both pay-per-use and bulk use options.
QuickSight Enterprise is $18 per month on an annual subscription basis for users defined as authors -- those who connect to data sources, create dashboards and other visuals and analyze data -- and $24 per month without an annual commitment. For "readers," those who only consume the data curated by authors, the cost is just $5 per user, per month.
In addition to individual per user pricing, on Nov. 30 Amazon introduced Session capacity pricing, an option for large enterprises working with big data. Starting at $57,600 per year, subscribers get up to 200,000 user sessions.
"This provides predictable BI spend for organizations," Jose Kunnackal, principal product manager for Amazon QuickSight, said during a presentation at re:Invent. "In a nutshell, what this model allows you to do is roll out QuickSight at scale to a very large number of users but benefit from scale with low prices."
The standard edition of QuickSight, meanwhile, is available to authors for $9 per month with an annual subscription and $12 if they pay month to month.
"It's the only truly serverless BI product on the market offering a pay-per-use business model," Evelson said. "While some cloud database management systems have been offering pay-per-use options, no other BI platforms have re-architected on a serverless model and therefore are not positioned to offer that pricing model."
Regarding the new embedded analytics capabilities, developers can now build and embed dashboards designed for three separate types of end users without having to subsequently provision and manage end users. The first is for those embedded dashboards that require no user provisioning, the second is for those designed only for authenticated users, and the last is for those aimed only at power users of apps.
In addition, Amazon unveiled an embedded developer portal for QuickSight where application developers can easily access tools for creating embedded applications.
Despite its growing capabilities, however, Menninger said QuickSight is limited.
"It's good at the two ends of the spectrum, very basic BI/analytics and then integration with ML," he said. "If you have basic needs it's got decent stuff, and if you're a sophisticated user you can get something. But if you're looking for a flexible business department tool, it's missing the middle section."
He added that while many BI platforms are known for excellence in a given area -- Tableau is associated with data visualizations, for example, and SAP and Yellowfin offer pixel-perfect reporting- -- QuickSight lacks a standout capability that gives it an identity.
Nevertheless, Menninger said that what Amazon has thus far added to QuickSight has it headed in the right direction, just not quickly.
"I'm impressed with how much is there," he said. "It seems like they're investing in it."