AWS on Tuesday unveiled a series of new features for Amazon QuickSight, the cloud computing giant's analytics platform.
AWS revealed the new capabilities during re:Invent 2022, a user conference held both in-person in Las Vegas and digitally for remote attendees.
Included in the update are new query, forecasting and data preparation features that add functionality to QuickSight Q, a natural language query (NLQ) tool Amazon first launched in September 2021. In addition, the update includes enhanced paginated reporting capabilities, a more powerful in-memory engine and tools to aid migration from on premises to the cloud.
Taken together, the new features add up to a significant update, according to David Menninger, an analyst at Ventana Research.
He noted that many vendors -- including Amazon -- issue monthly updates, which makes most of their updates relatively incremental. But amid the many monthly updates, the QuickSight features being revealed at re:Invent stand out.
"Most software vendors are making frequent updates to their products, so most releases end up being incremental," Menninger said. "If you look at the release history, Amazon has been issuing updates nearly every month. [But] in that context, this collection of new features is fairly significant."
All of the new QuickSight capabilities are part of Amazon's quest to help organizations expand their use of analytics, according to Tracy Daugherty, general manager of QuickSight.
Depending on the study, it's estimated that only a quarter to a third of employees in organizations use analytics as part of their workflow. And despite the evolution of business intelligence platforms to include embedded analytics and no-code/low-code capabilities, the percentage of analytics users within organizations has remained static for years.
"There's this relentless desire to get more than 20% to 30% of an organization using BI tools," Daugherty said. "The question is how to get it to 50% or 60%."
One approach Amazon is taking to increasing analytics usage is to focus on the end user rather than the developer, he continued.
"That's where NLQ came from," Daugherty said. "Because we want to go to a broader percentage of people in the organization, we have to think about more than just that core technical group."
QuickSight Q is a no-code tool that enables users to ask questions of their data using natural language and responds with relevant data visualizations that help them derive insights leading to data-informed decisions.
Other vendors offering NLQ capabilities include Tableau, ThoughtSpot and Yellowfin, among others.
Beyond basic queries -- such as what sales volume was for a certain month in a given region -- QuickSight Q now enables users to ask "why" questions of their data, such as why sales increased or declined during that month in that specific region.
They can also do forecasting with QuickSight Q, asking the tool to predict what might happen given previous patterns.
Both new querying capabilities are part of AWS' effort to increase the effectiveness of NLQ, which is not yet as robust as traditional queries written in code.
"The nirvana is asking a 'why' question and getting everything back, and these are efforts along that path," Daugherty said.
In addition to the new NLQ features, QuickSight Q now includes automated data preparation capabilities.
Since the launch of QuickSight, Amazon discovered that data preparation for NLQ is different than preparing data for dashboards, according to Daugherty.
Generally, when preparing data for a dashboard, the developer has predetermined both the questions and answers.
Data preparation for NLQ, however, requires preparation for queries that are more exploratory because the answer isn't yet known. In addition, given the vagaries of natural language, the same question can be asked in myriad ways, whereas coded queries can be asked in only one way.
To address those differences -- and at the behest of customers, according to Daugherty -- Amazon developed data preparation capabilities tailored for NLQ.
The feature uses machine learning -- both pre-trained models built by AWS as well as a user's query history -- to automatically add semantic information to datasets. The result is a reduction in the time it takes to prepare those datasets for natural language queries from potentially days down to minutes, according to AWS.
That attempt to reduce the time it takes to prepare data is critical, according to Menninger, who called data preparation the most time-consuming aspect of analytics.
Meanwhile, with its continued focus on enhancing QuickSight Q, Amazon is turning QuickSight into a platform that can compete with peers such as Qlik, Tableau and Microsoft Power BI after it initially lacked functionality following its launch in 2016, Menninger continued.
"Amazon continues to invest [in QuickSight]. And with some of its features -- such as Quicksight Q natural language query -- it could be considered a leader," he said.
Additional new features Amazon added to QuickSight include the following:
- QuickSight Paginated Reports, a paginated reporting tool for the cloud that provides formatted summaries of data, such as daily and weekly transactions that can be broadly shared to inform data-driven decisions;
- added power for QuickSight's Super-fast, Parallel, In-memory Calculation Engine (SPICE) that doubles previous support for datasets of up to 500 million rows to support for datasets of up to 1 billion rows; and
- new application programming interfaces designed to help customers migrate data and analytics assets like reports and dashboards from on-premises databases to the cloud.
The new reporting capabilities could be of particular benefit, according to Menninger.
"Reporting is a feature often overlooked by many of the products in the market, yet our research shows it's used by 87% of organizations," he said.
Similarly, Daugherty highlighted the QuickSight Paginated Reports.
He noted that Paginated Reports helps QuickSight bring reporting, data visualization, embedded analytics and augmented intelligence capabilities like NLQ together in one environment.
"Some users love dashboards, some love reports, some love embedding their analytics and some love asking questions," Daugherty said. "The reality is that most people use a combination of those things and don't like using different tools for each of them."
As Amazon continues to add capabilities to QuickSight, Daugherty and Menninger noted that one of the platform's strengths is its foundation in the cloud.
Unlike many analytics platforms that began as on-premises tools that were later re-architected for the cloud, QuickSight is one of the few that was cloud native from the start; Domo is another.
David MenningerAnalyst, Ventana Research
With its cloud foundation, QuickSight can scale to meet the needs of large customers.
Pricing has also been a differentiator for the platform, and was so even before Amazon added features like QuickSight Q and embedded analytics capabilities that brought it more in line with the capabilities of competitors.
Plans for authors -- those who create data products like dashboards -- start at $18 per month for an annual commitment. For readers who analyze data but don't develop any products, the cost is as little at 30 cents per session with a maximum of $5 per month.
And for organizations with a high volume of users, Amazon offers capacity-based pricing for QuickSight that can range anywhere from $250 per month to over $250,000 for an annual subscription.
Going forward, two targets for further improvement are data preparation and data governance, according to Daugherty.
"We need to make our data prep experience stronger," he said. "We wanted to get other end user-facing capabilities rock-solid first, so data prep is an area of investment. The second one is governance. We don't have [governance] all in right now, but we will this time next year."
The roadmap for QuickSight, meanwhile, includes features for the three personas the platform aims to serve: authors, readers and administrators.
Menninger, meanwhile, noted that QuickSight is dependent on the larger AWS ecosystem for multi-cloud capabilities. Additionally, there remain some capabilities the platform doesn't have that he'd like to see.
"From a functionality standpoint, there are some user interface elements that lag others, such as data storytelling and Microsoft Office integration," he said.