A Stockphoto - stock.adobe.com
Data governance, AI among the past year's analytics trends
Organizations recognized data governance as an enabler of self-service BI and a way to keep data safe as well as that AI is best used to augment humans rather than replace them.
Data governance was one of the major trends that shaped analytics in 2022.
Data governance isn't glamorous like augmented intelligence, machine learning or natural language processing. It's the grunt work of analytics.
But after many organizations suddenly realized the importance of data-informed decision-making at the start of the COVID-19 pandemic -- and have continued to recognize its value as world events like the war in Ukraine and repeated supply chain disruptions have resulted in ongoing economic uncertainty -- data governance is starting to be a critical need.
Some organizations hastily deployed analytics operations over the past few years. 2022 was the year they emphasized governance of those operations.
But data governance wasn't the only significant analytics trend in 2022. The evolution of AI and a market correction for tech vendors all played prominent roles in shaping the past year.
Data governance is an organization's rules for data.
It includes guidelines such as naming conventions so that data is consistently labeled and easy to find as well as access controls to determine which employees can have access to what data. It also includes methods for determining how an organization goes about processes such as data integration, data preparation and data analysis.
Ultimately, the aim of data governance is to simultaneously protect an organization from risk of violating regulations for data while also serving as an enabler for end users who need to safely and confidently work with data.
So with many organizations hastily launching analytics programs to deal with the pandemic and the tumultuous events that followed, data governance was a critical analytics trend in 2022.
But it wasn't just organizations implementing guidelines after getting started with analytics that made data governance a big analytics trend. It was also organizations that have been using analytics to inform decision-making for years evolving their data governance to be less about protecting the organization and more about enabling the business user, according to David Menninger, an analyst at Ventana Research.
"Data governance is really skyrocketing in terms of awareness and popularity," he said. "The reason is that organizations have changed their approach to governance from a disabling process to an enabling process. It's now about helping people get their jobs done and, in doing so, creating a framework that creates governance."
Those organizations adopting self-service analytics need data governance frameworks that enable data exploration and analysis while protecting against compliance risk.
Self-service analytics is about giving business users access to data with easy-to-use tools. But because those business users aren't data experts, and because they aren't high-level executives who have the right to access sensitive data, companies must put specific limits on their data usage.
Meanwhile, self-service analytics is growing, noted Ritesh Ramesh, COO of healthcare consulting firm MDAudit and a customer of analytics vendor ThoughtSpot.
"There is pressure on all organizations to do more with less due to shrinking profits -- hence the need to [for business users] to have total autonomy," he said.
With more organizations deploying analytics tools and self-service analytics growing, one means of organizing data, putting data governance measures in place and monitoring data's use is by implementing a data catalog. Data catalogs are indexes of organizations' data and data products that incorporate governance measures to ensure data is used properly.
"This year, a lot of effort has been put into data observability, from linking it to data catalogs to completely automating data governance," said Donald Farmer, founder and principal of TreeHive Strategy. "Specifically, there has been a focus on having the capacity to detect, keep track of and assess data flows throughout a business to recognize and solve data-related problems as they come up."
Most analytics vendors now have data governance capabilities built into their platforms with Tableau and MicroStrategy, among those that made data governance a priority in 2022. Vendors such as Alation, Collibra and Informatica specialize in data catalogs.
Just as 2022 marked the year that organizations realized data governance can be an enabler of analytics, it was the year analytics vendors started to view augmented intelligence as a facilitator of human analysis rather than a replacement for it.
AI is nothing new. But it has often been viewed as a means of machines replacing people.
Instead, the concept of decision intelligence became an analytics trend in 2022 -- one in which AI and machine learning augment human analysis.
Organizations possess far more data than any one person -- or even a team of people -- can sift through and observe for the changes and anomalies that can affect their business. Tools built with AI and machine learning, however, can comb through hundreds of thousands of rows of data in seconds and constantly observe key organizational metrics for changes.
They can also alert employees when something needs attention so the human, who may have taken months to discover the change or anomaly -- or never found it all -- can investigate and take action.
"Decision intelligence was a big trend this year," said Krishna Roy, an analyst at 451 Research. "We saw it establish after much buzz in 2021. Decision intelligence platforms aim to address the long-held holy grail of data and analysis democratization so that everyone who needs to be can be a data-driven decision-maker."
David MenningerAnalyst, Ventana Research
Vendors specializing in decision intelligence include Pyramid Analytics, Tellius and Sisu.
The need for AI and machine learning goes beyond just decision intelligence, according to Bryan Harris, executive vice president and CTO at SAS.
Organizations need help dealing with the exponential growth of data. The amount of data captured and consumed globally amounted to two zettabytes in 2010, according to Statista. By 2020, that grew to 64.2 zettabytes. By 2025, Statista forecasts an increase to 181 zettabytes.
Much of that growth is due to the growing influence of the cloud, which enables organizations to quickly ingest data from an ever-increasing number of sources and store it in cloud data warehouses and lakes with far more capacity than an on-premises data warehouse.
To make use of that data and not just leave it sitting untouched in cloud repositories, such as Azure or Redshift, organizations need AI and machine learning to manage and model that data to ready it for analysis.
"To make sense of the increasing volumes of data, organizations are looking to apply the right analytic and modeling techniques to support transparent and explainable decisions," Harris said. "If businesses can successfully transition to the cloud of their choice while also accelerating the adoption of AI, they can ultimately become more resilient and agile and grow their businesses during disruption."
As a result of the growing need for AI and machine learning, many analytics vendors have now added such capabilities in an attempt to serve the needs of their customers, Menninger noted.
AI and ML tools are difficult for organizations to build on their own, so vendors recognized an opportunity to provide the capabilities as part of their platforms.
"Augmented intelligence has really come to the fore," Menninger said. "We're continuing to recognize that AI and ML are hard, so to the extent vendors can provide a subset of those capabilities and do them in a way that makes them broadly accessible, they've found a lot of success."
While technology advanced in 2022, one analytics trend was a slowdown in the investments that enabled startup vendors to grow quickly and operate despite a lack of profitability.
In February 2021, data lakehouse vendor Databricks raised $1 billion in venture capital funding. In addition, vendors including Confluent, Couchbase, Neo4j and Sigma Computing all executed funding rounds of at least $200 million in 2021.
Meanwhile, cloud data platform vendor Snowflake set a record for tech companies by raising $3.4 billion in its initial public stock offering in September 2020.
But since early 2022, when the stock market began to slide, inflation rates spiked and recession fears continued to grow, data and analytics vendors have found venture capital hard to come by and the stock prices of publicly traded companies have slid precipitously.
"It feels like we're at an inflection point right now," said Dan Sommer, senior director and global market intelligence lead at Qlik.
He noted that the pandemic and subsequent worldwide events sparked a desire for new technology that could help organizations manage uncertainty. As a result, cash flowed freely to vendors finding new and inventive way to help organizations survive the uncertainty and thrive compared to peers not using analytics to inform decisions.
But when the broader stock market slide took tech stocks with it, and when fears of a recession became more real, the stream of cash slowed.
According to Crunchbase, total VC funding was down 53% during the third quarter of 2022 compared to the same three months of 2021. That continued a yearlong trend with VC funding declining each quarter so far in 2022.
In the stock market, though the Dow Jones Industrials Average is down only about 9% to date this year, the more tech-heavy Nasdaq Composite Index is off more than 30%.
Among publicly traded data and analytics vendors, Snowflake's shares are down over 50% year-to-date, and both Domo and MicroStrategy are down more than 60%.
Meanwhile, vendors such as Pyramid, Qlik, SingleStore and ThoughtSpot have all expressed interest in going public but have not yet proceeded with IPOs. Qlik went so far as filing initial paperwork with the Securities and Exchange Commission in January 2022 but remains privately held as the year comes to a close.
"In January 2022, right when the year started, the tone changed drastically," Sommer said. "We had geopolitical events that took place, so the macro backdrop is what changed [and] the tech landscape was affected."
The result has been a shift among investors from potential to proof, with vendors needing to show sustainable growth and, at minimum, a path to profitability.
"The micro backdrop is that there's been a shift from growth to value," Sommer said. "Technology needs to prove itself, and it needs to be sustainable versus previously when people were buying hype. It's shifted from hype to value."
Beyond profitability, Sommer said that data and analytics vendors can demonstrate their value by focusing on more than just one feature and offering a full-featured platform that includes automation, data integration, data management, data science and analytics.
"That provides value [for organizations]," Sommer said. "And it goes back to that idea of hype versus value."
Additional 2022 analytics trends
Beyond some of the major analytics trends in 2022, numerous other trends also played a role in how the year played out.
Ramesh noted that governance extended beyond just data in 2022 to AI governance and governance of machine learning models.
Just as organizations want to put parameters on the use of their data and understand how it's being used, organizations with more advanced analytics that delve deep into data science want to know as much possible about how AI and machine learning models are being used.
"People want to understand what are the variables and decisions behind algorithms and models making decisions," he said.
Tracy Daugherty, general manager of the Amazon QuickSight analytics platform, noted that two ongoing trends -- cloud migration and embedded analytics -- kept gaining momentum in 2022.
"The big thing in 2022 was the continued acceleration of people moving to the cloud," he said. "It's not just moving the BI tools but also moving the data with it. That has grown in a big way. And we've seen it with some substantially sized deployments, which is exciting."
Regarding embedded analytics, Daugherty added that as time passes, he expects people to increasingly get their data and perform their analysis in applications rather than traditional BI environments.
"There's just more context in applications -- it's easier to understand," he said. "And most vertical industries are driven by applications as opposed to standardized tools, so I think we'll continue to see growth in that area."
Finally, Harris cited a rise in the use of digital twins to do scenario planning as an ongoing analytics trend.
Digital twins are copies of data that can be used to play out different scenarios, to repeatedly ask "what if?" of data to inform how an organization should react if one scenario or another plays out.
Vendors like Anaplan specialize in scenario planning, but such capabilities are not yet common among BI vendors.
"The next generation of the analytics life cycle, or ModelOps life cycle, will be simulating complex systems to help prepare for any possible scenario or disruptive event," Harris said. "From there, businesses can make rapid and resilient decisions to minimize risk and maximize profits."