Embedded BI is the most significant analytics trend of the moment, according to a well-placed industry expert.
Embedded analytics delivers data and insights to business users within their workflows, giving them the information they need at the moment they need it to inform a decision. And that is currently the most effective means of making smarter decisions with data, said Kate Wright, senior director of product management at Google Cloud.
Google on Tuesday unveiled its "2023 Data and AI Trends Report," highlighting five current data and analytics movements.
Augmented intelligence is one of them, and perhaps the analytics trend that gets the most attention. But for all the potential AI holds to make analytics easier and break down the many barriers that prevent people from using data as an everyday part of their workflow, it isn't yet as effective as infusing data and insights -- embedding BI -- to enable workers with data, Wright said.
Instead, AI has evolved to become more of an assistant -- a means of augmenting humans throughout the decision-making process.
In addition to both AI and embedded BI, the trends the report highlighted include reducing the number of isolated data repositories, embracing an open data ecosystem, and making data more usable and discoverable with governance.
Recently, Wright discussed current analytics trends, including why embedded BI is so important.
In addition, she spoke about underrated and overrated analytics trends, and how Google -- and Looker, in particular -- is plotting its roadmap in response to key trends.
What is the most significant current analytics trend?
Kate Wright: Some of the trends in the BI space have persisted for some time. One of them is to have more people be able to benefit from data through BI tools. That has been a trend in the industry for five, six or seven years, but how we're trying to solve that problem has changed more recently. Historically, we tried to solve it by giving people access to the tooling. When that didn't work, we tried to make the tools more automated and intelligent and easy. That still didn't move things along.
Kate WrightSenior director of product management, Google Cloud
Now, the trend is more around reaching people where they actually work. Rather than asking business users -- who we were asking to be more data-driven -- to use a specific tool, now we're trying to help them by putting the data right in front of them in the tools they're using every day. Instead of asking them to go to the tool, we're bringing the data to them.
How are vendors enabling organizations to get their data to end users within their workflows and embed BI so that it can be consumed in familiar tools and applications?
Wright: Vendors are enabling developers, but they are also building integrations between BI content and the productivity tools where people are also working.
From the perspective of embedded analytics, there has been the ability to embed dashboards for a while. Most of the vendors have provided that. More recently, we're starting to see the platforms providing APIs that are more modular. They directly provide not only metrics through APIs that can be used anywhere by a developer, but also widgets and other components that can be taken off the shelf and be put into users' experiences. If we're reaching the business user where they want to live, then typically they're in an experience that has certain expectations around it, and it's often difficult to fit a dashboard into those experiences. And if it can be fit, it's often a separate page or a separate workflow. So what we're seeing is either headless BI or APIs for insights, and those widgets, like a single chart or set of filters.
You mentioned integrations with productivity tools -- what do those look like?
Wright: Going beyond developers, I see quite a few BI vendors investing and ensuring that the content in their tools becomes more readily available where a lot of knowledge workers work. We're thinking about collaboration tools, spreadsheet tools and productivity suites. So whether that's being able to ask questions or schedule information to be distributed, it still requires the business user to do the work, but it doesn't require the developer, and it's meeting the users where they are. That's something we've seen really tick up in the last few years.
What's a real-world example of an end user being provided data within their workflow to make a decision?
Wright: My doorbell rang about five minutes before we started talking, and it was the gas guy. That's a customer service relationship. I had to call in to get my meter changed, and when I did that, you can imagine that the customer service rep on the other side of the phone was automatically able to look at me as a customer and see information about me. Historically, that would have been my name and address, but these days we're seeing insights presented to customer service agents, like how happy I'm likely to be with [a net promoter score], potentially recent complaints I may have made, maybe what my potential lifetime value is to the gas company.
We're also starting to see embedded analytics become a differentiator for businesses as they're offering things and selling information to their own customers to get the analytics to their own end users. One of our customers, Wpromote, is running an ad agency and has embedded Looker's analytics into the experience for their customers so those customers can make sure they're getting the best performance for their money. The customer service agent is an example everyone can relate to, but it goes so far beyond that.
Beyond embedded BI, what is another major current analytics trend?
Wright: We've seen this oscillation where, because of self-service analytics, most departments choose their own BI tool to work with. As a result, companies have a multitude of BI products in their landscape. Five years ago, even three years ago, a lot of the conversations I was having with IT and customers were about getting down to one because it was really hard to manage all the BI products. The hard-to-manage part hasn't changed, but the desire to get down to one is slowing. It's very difficult to convince people that they have to give up something, so instead of focusing on that, organizations are focusing on how to better ensure that the data is of a high quality. They're making their investments below the BI content to govern the BI tools better, ensure the data is more trustworthy, and create best practices and a data culture.
It's a combination of process and tools.
What is an overrated analytics trend?
Wright: A lot of things that are overrated are still really good. Think about when you get a restaurant recommendation -- someone says it's the best food they've ever eaten, and even if you might have loved it, your expectations were too high.
I was a huge believer in augmented analytics and its ability to reach people -- its ability to lower the barrier to entry, its ability to reach people in different modalities, and its ability to automatically present outliers and insights. I think that's all still really valuable, but we all expected it to change our world. In fact, it just improved our world. It's part of the picture, but it didn't change the game. That continues to be an area of investment, but it's not yet a game changer. We'll see what happens in five years. The data always tells you part of the picture, but some of the data isn't in the system. It's in your expertise. I always believe that the information is something you need to make a good decision, but it isn't the only part of the picture. Having that human in the loop is a much more sound way to work. But on the flip side, just having a human and not the data is also a problem.
What is an underrated analytics trend?
Wright: We're seeing that what was old is new again with centralization and higher amounts of governance.
Semantic layers are coming back. I'm someone who started their career with a semantic layer BI product, so I have a soft spot for that. But I do think that's something that is going to become more relevant. There is such a downstream benefit to real-time, accurate data, so we need to make sure we're thinking it through and passing that value along. The opportunity is huge with governance layers, but there needs to also be balance so users have both agility and [protective guardrails]. That has such potential, and I'm not sure we've had enough conversation around that yet.
As you've seen these analytics trends evolve, what capabilities has Google developed over the past year or so to address them?
Wright: Google acquired Looker [in 2019], and Looker has always been focused on both embedded analytics and centralized BI by providing a cloud-based semantic layer. Looker was also heavily invested in APIs. What we announced last year is that we were putting together the Looker family of products, and we brought what we now call Looker Studio into the mix, and that's the self-service visualization piece of the platform. Also, it's stated that we want to be the most open BI platform, so our semantic layer has now been opened up to other tools.
There's the agility of Looker Studio to do self-service visualization, or users can connect into our semantic layer. In addition, we announced a partnership with Tableau, and we have a Power BI connector in private preview. We recognize that there are many BI tools, and we want organizations to be able to take advantage of the metrics and trusted data that Looker can provide without necessarily having to completely alter their strategy.
Looking ahead, how does Google's roadmap address ongoing analytics trends?
Wright: On the embedded analytics side, we're really leaning heavily into those investments and providing the APIs that are required to meet our customers' needs. That tends to mean ensuring that it can provide any experience -- and be embedded into any experience -- in a way that meets the design criteria of what is being asked for. Too often, a customer has to choose between the BI platform they have or the business application look and feel, and we want to make sure that no longer is a problem. That's a huge investment area for us.
We are also going to push further to ensure that our semantic layer is truly open. We want to make it more accessible to just about everyone, so there is more coming in that area, hopefully pretty soon.
Editor's note: This Q&A has been edited for clarity and conciseness.
Eric Avidon is a senior news writer for TechTarget Editorial and is a journalist with more than 25 years of experience. He covers analytics and data management.