Let’s Talk

Thank you, you're submission has been received.

There was a problem submitting your request.

What are your primary business content and marketing goals?

Tell us more

Let’s Collaborate

#39 The Top Trends in 2022 for Data Leaders from DataRobot, Databricks, and Google

Play episode

At the end of every year, you’re probably asking the same questions we are. What are the big changes coming next year? How do I stay ahead of them? And what’s separating real trends from the hype?

To answer these questions, we are excited to bring together some of the top minds in the industry. In this special episode, we’ll pick their brains and dig into what you need to know to thrive in the year ahead. 

You’ll hear from three incredible guests — all of whom are building and shaping the future of data and analytics:

  • First, Ben Taylor, the Chief AI Evangelist at DataRobot
  • Then, the Global Field CTO of Databricks, Chris D’Agostino.
  • And finally, Bruno Aziza, the Head of Data & Analytics at Google Cloud.

Nothing is off the table. So whether you want to hear about augmented everything, dig into the debate around different cloud platforms, or learn why analytics has become more impactful than ever, this is the episode for you.

Key Takeaways

  • CDOs must deliver simplicity but contend with complexity: As the data ecosystem continues to introduce new innovation at an ever increasing rate, data leaders must grapple with all these new capabilities. At the same time, however, the rising need for access to this innovation from nontechnical, business professionals means CDOs must simultaneously deliver simple, intuitive experiences that empower the rest of the business
  • Is the data warehouse on the way out? D’Agostino makes a bold prediction that within 10 years, the traditional data warehouse won’t exist. That begs the question: what will replace it? The lakehouse, data mesh, and data fabric are all contenders, but require organizational changes, not just the introduction of new technologies, as Aziza points out. 
  • Preventing bias within models: A consistent problem in the industry – one that we’ve touched on several times this year – is the potential for machine learning and AI to scale bias in unprecedented ways. As we enter 2022, it will become even more imperative that you and your team are able to answer questions like “how will this model potentially amplify basis,” “how can we prevent biases,” or “what biases exist in our data sets?” 
  • Creating an ecosystem of data sharing: The rise of analytics exchanges creates massive opportunity for businesses for two reasons. First, it allows users to share data across platforms at a faster rate. And second, users are now able to share more than just data, but actual assets at an improved rate.
  • In 2022, AI, ML, and data products must prove value: For years, companies have experimented with AI and ML, but as Taylor points out, the disillusionment with the impact of these experiments is at an all time high. So whether you’re building data products or launching new AI use cases, data leaders need to lead with the value they will deliver, not only imagine the art of what’s possible.

Ben Taylor Key Quotes

“If you don’t see the path to value, then you’re wasting time. And sometimes with A.I. That’s a common miss… Imagine what you could do, and I chuckle now when I hear that line, ‘Imagine what you could do with A.I.,’ because where’s the value in that statement? Why not just direct people straight to value, and focus on a proof of value?”

“Your creativity is the new bottleneck. And someone said, ‘How do I increase my creativity?’ You can talk about trust and delivering value, but you really do need to inspire people as well at the same time.”

“The most successful teams I’ve seen treat [data science] like sales, where you can ask them ‘What’s your data science attribution number?’ And I bet if you asked most organizations what that number was, they’re not prepared to answer that. The most impressive organizations have that number.”

“If you look into the details, how did this happen? No one is being proactive. Anyone who’s doing anything with data and algorithms, you should be thinking about this. How could this model amplify a bias? And what are the biases that are top of mind with the data sets that we have? We have unconscious bias. Everything we do as humans will introduce bias.”

Chris D’Agostino Key Quotes

“Taking a distributed computing approach to data…The question becomes how you actually start executing on machine learning models and training those models with vast amounts of data. And as Google published a large amount of high-quality data with a less sophisticated model, it will outperform a really sophisticated model with poor quality data or small amounts of data.”

“We’re presenting the Lakehouse architecture and the light bulb is going off. [Customers are] seeing that they can actually simplify the architecture. They can have fewer running systems. The more systems you have, the more complexity you have, the more DevOps support you need. If you can reduce that complexity, you can actually do more with data because you’re reducing the number of copies. You have fewer running systems that you need to keep up. Data just flows and the curation steps happen and the data becomes more consumable.”

“The Lakehouse will indeed become the future architecture. We’re seeing it today. We’re seeing enterprise data warehouses, vendors rebranding themselves as data platforms and kind of moving away from that moniker, the EDW moniker.”

Bruno Aziza Key Quotes

“There’s two trends that are going to change the next decade, next year. One is the move to the cloud. If you look at Gartner’s prediction, 75% of databases by 2022, are going to be in the cloud. So there’s a gigantic move to the cloud and this transformation of not just moving what you’re doing on premise to the cloud, but also using it as an opportunity to transform your processes, excel rates, your path to production.”

“We asked our customers, ‘how do you think about the cloud opportunity?’ And often, they’re looking at multiple clouds. [Customers are thinking] ‘What is my cloud strategy?’ You should be thinking about, ‘What is my cloud strategy? And how you mature through it.’

“Nobody wakes up in the morning, and says they have a data warehousing problem, or they have a data lake or data lakehouse problem. They wake up and they’re worried about innovation. They’re worried about, how to create data products?”

“The best way that we’ve seen customers really succeed is, what is your endgame? What are you trying to do ultimately? You’re trying to build data products and how do we get you there? You’re probably going to need a construct that looks like a data lake. You’re probably going to need a construct that looks like a data warehouse. Does this have to be in the same product? I’ll let you make the decision.” 

“How do we help the CDO deliver simplicity for their users, while in the background, we still conserve this ability to handle the sophistication? Because the questions they’re asking are not simple questions. The interface needs to be simple, but the complexity behind it needs to be supported. So that’s where integration, openness… The ability to have an analytics exchange and the rise of analytics exchange is really going to be a big trend in the next year and next 10 years, where people now need more than just sharing data. They need to be able to share assets.”

Episode 41