Top analytics trends include cloud, AI/ML and embedded BI

Cloud migration, embedded BI and the incorporation of augmented intelligence and machine learning are three of the top analytics trends.

Doug Henschen

That's according to a new report by Constellation Research analyst Doug Henschen titled the "2023 Analytics and BI Market Overview" report.

For decades, analytics was the centralized domain of a team of data experts who lorded over their organization's data, doing all the analysis themselves and delivering reports when requested. Then came the era of self-service analytics, with vendors such as Tableau and Qlik enabling developers to build dazzling dashboards that let business users interpret data.

However, the vast majority of data was still kept on premises, and analysis was limited to those organizations that could afford to set up users with licenses for their BI tool of choice and equip those users with data literacy training so that they could interact with data.

Over the last decade -- and in particular the past few years -- that has changed. Vendors have pushed to enable customers to become more data-driven by making analytics tools usable by a broader audience.

For more than two decades, BI adoption within organizations has been relatively stagnant, stuck at about 25%. Now, however, vendors are offering cloud-native versions of their platforms that don't limit the number of users with licenses. In addition, they are enabling customers to embed BI into other applications so that users don't have to search for relevant data. And they are incorporating AI and machine learning (ML) that reduce the required level of data literacy it used to take to interact with data.

Recently, Henschen discussed not only the findings of his report, which focuses on how well vendors are reacting to major analytics trends, but also other ongoing BI trends. He also talked about overrated and underrated trends, and which trends might be most important a few years from now.

In your report, you identify three ways BI/analytics platforms are evolving, so let's start by looking at them in the aggregate. When combined, what do cloud migration, embedded analytics, and AI and ML technology enable organizations to do that's so critical in today's business landscape?

As for what these three trends do for customers, one common theme is enabling more users to take advantage of data-driven insight.

Doug Henschen: These three trends are interrelated and supportive of each other in many ways. For example, cloud computing has supported embedded approaches, particularly services-based architectures and granular APIs. Cloud has also enabled AI, ML and augmented capabilities with scalable compute and storage.

As for what these three trends do for customers, one common theme is enabling more users to take advantage of data-driven insight. Cloud approaches make it possible to deliver bigger and broader deployments more quickly and easily. Embedded approaches make it possible to deliver concise insights where people work within applications, productivity and collaboration tools, and workflows and business processes. Augmented capabilities, such as natural language querying and explanations, make it easier for all users, including novice business users, to take advantage of analytics and BI.

Now looking at each of the top analytics trends individually, what are they doing differently from a few years ago to better enable cloud migration?

Henschen: Just a few years ago, many vendors supported cloud deployment with cloud marketplace offerings that made it easier to deploy their software, but it was still up to the customer to manage the deployment. Software as a service options were multiplying three years ago. Today, all 17 vendors in our report have a SaaS offering on at least one public cloud.

Many vendors have also developed containerized images that ease deployment on public clouds using so-called cloud-native services. That's helpful, but in the absence of a services-based offering, it assumes that customers are willing to manage the software themselves. Our new 'Multicloud Analytics and Business Intelligence Platforms' shortlist includes eight vendors that offer their platforms as SaaS or as a managed service on two or more public clouds.

Similar question regarding embedded analytics -- how are vendors enabling customers to embed BI and analytics, and how has that evolved over the years since we first started hearing about embedded BI?

Henschen: Embedded analytics options have long been used by independent software vendors and SaaS companies to accelerate development and data-driven software and services. Today, innovative organizations of every description -- banks, insurance companies, you name it -- are developing software and services, and they are using embedded capabilities to weave analytics into custom apps, enterprise apps, productivity and collaboration apps, and workflows and business processes.

The nine leading vendors in our 'Embedded Analytics' shortlist are supporting embedded approaches harnessing microservices architectures, fine-grained application programming interfaces, RESTful interfaces and software development kits that support flexible embedding of data, metrics, visualizations and dashboards into a range of external destinations. They're also pursuing automated DevOps-style deployment approaches, low-code/no-code development options, event architecture, and in some cases, workflow and process integration to automate actions and trigger human review based on analytic insights and thresholds.

And how are vendors incorporating AI and ML into their platforms to better enable customers -- and again, how is what they're doing now different from what they were doing just a few years ago?

Henschen: Our 2019 analytics and BI market overview focused on then-emerging augmented capabilities. Data-prep options included recommended joins. Discovery and analysis features included outlier detection and influencer/root cause analysis. Natural language interactions included NL query and NL explanations.

The breakthrough here in 2023 is harnessing generative AI, and vendors including Microsoft, Salesforce and ThoughtSpot are now previewing generative AI features. Generative AI has an amazing potential to put natural language interaction, sentiment analysis, data visualization and query code generation on steroids, as detailed in our report. But it's important to note that all the generative features we've seen in the analytics and BI space are in private or limited preview at this time.

Beyond those key analytics trends, what are some other ways vendors are now trying to help customers become better and more efficient with their analytics and ultimately make better decisions?

Henschen: Many vendors are introducing or improving upon their data cataloging, data modeling, metadata management and governance capabilities. Some are pushing into predictive analytics and data science with new augmented and automated ML features. I'm also seeing incremental improvements and refinements of more mature capabilities, such as data preparation, dashboarding, storytelling and reporting.

Delving more deeply into generative AI, what do you see as the role of generative AI in analytics in 2023? And beyond 2023, how might the relationship between analytics and generative AI evolve?

Henschen: Possible generative AI applications are numerous and potentially powerful. Generative AI promises to put natural language interaction on steroids, which could truly bring data-driven insight to the masses -- both internally and to external partners and customers.

But here's where caution is advised. Every vendor that is previewing generative AI features has made a point of incorporating human review steps. Data privacy concerns must be addressed, and the accuracy of AI-generated content demands ongoing monitoring and human review. Analytics and BI is just one potential use case for generative AI, and there's a vigorous tech industry and public debate raging over the proper use of generative capabilities and the need for regulation and safeguards.

If cloud deployment, embedded analytics and AI/ML augmentation are the three most important analytics trends as of May 2023, what do you think will be most important a year or two from now?

Henschen: It's easy to guess that generative AI breakthroughs will still be rippling through all technology markets and society in the coming years. Despite talk of six-month pauses, geo-specific bans and possible heavy-handed regulation, I think it's going to be difficult to hold such powerful technology back. I'm optimistic that the good will outweigh the bad and that it will be a net job creator. In the analytics and BI space, old-school reports and dashboards just might seem like stone knives and bearskin rugs five years down the road if it all moves to conversational analysis, prediction, what-if scenario planning and optimization.

Staying with analytics trends for one more question -- what is the most underrated trend and what is the most overrated trend?

Henschen: I'd say the most underrated trend has been the emergence of data catalogs, collaboration features, and ML-driven data and content recommendations, all of which have helped users to find existing and trusted data sources, measures, reports and dashboards. I'd say the augmented trend didn't quite deliver as promised because too many of these features have been too complex for broad adoption.

NL query and NL explanations have been exceptions seeing broad adoption and use because they are user-friendly.

Moving past trends, which platforms do you think are doing the best overall job of meeting the modern needs of their customers?

Henschen: I really must push back on one-size-fits-all analysis, because every customer has different needs. Can one of those choices be on your shortlist if it isn't available on the cloud that you use primarily or exclusively? We think not. If an organization is deeply invested in a particular enterprise applications suite, it should definitely put the integrated analytics and BI offering from that vendor on its shortlist. The consistent data models and pre-built data integrations, dashboard and reports, and embedded analytical capabilities will accelerate their deployment and help to deliver a rapid return on investment.

If you're exclusively on one public cloud and expect to stay that way, I'd definitely look at the tightly integrated analytics and BI offering available on that cloud. Large and technology-diverse organizations tend to embrace the independent offerings that help them make sense of information across their tech landscape and, very often, on multiple clouds. Many organizations want to move from insight to action, and the [embedded analytics] leaders can help them do just that.

Lastly, are there vendors whose capabilities are lagging well behind those of their peers when it comes to modern customer needs?

Henschen: There are circumstances when vendors that aren't on any of our shortlists should be considered by specific types of customers. For example, customers that operate exclusively on AWS and that are contemplating particularly large deployments should consider Amazon QuickSight. Infor customers should definitely consider Infor Birst. And cost-conscious organizations and, certainly, any customer using any applications from Zoho should look at Zoho Analytics.

This is where it pays to [do due diligence] rather than just looking at plots in a box and forming a superficial opinion about which analytics and BI products an organization should consider. Reports are just a starting point or steppingstone to an organization's shortlist and final selection.

Editor's note: This Q&A has been edited for clarity and conciseness.

Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.