Re-Thinking Tomorrow’s Data-Driven Applications
In the newest installment of the Data Unlocked Podcast, Jason Davis, Simon Data’s co-founder and CEO links up with two thought leaders: Tom Tunguz, Managing Director at Redpoint Ventures and Bill Stratton, Head of Media, Entertainment and Advertising at Snowflake.
The Big Question
As always, Jason asks Tom and Bill how marketers could unlock opportunities with better data capabilities.
Tom’s Reply Summarized – The Data Supply Chain
So Redpoint, we’ve been calling 2020 the decade of data. There’s been lots of innovation in data for an awfully long time. But the cloud data warehouses and the modern data stack have really changed the way that people use data. It’s brought the benefits of insights to almost every single employee. Data is created in some place, and then it needs to be processed and it needs to go where it’s going to create value. And over the last couple of years, we’ve seen meaningful innovation in parts of that data supply chain.
So the first is moving the data from where it’s being produced. And so you have ETL vendors there, Extract, Transformation and Load, who are doing quite well. They’re piping that to a data warehouse where it lives. And you’ve got, obviously, Snowflake, which has become just a monster business doing that. You have modeling layers and that’s a place where data engineers create single definitions for metrics.
And then the last part is really making use of the data once it’s inside of the data warehouse. There are machine learning applications and ways of building predictive models.
Then there are pieces around the data supply chain. There’s data security, making sure that people are using data in a compliant way and only the people who actually have access to it are accessing it. And then you have data observability. It’s really understanding through my data supply chain, is everything working well? And if it’s not, why? And enabling people to go fix it.
One that strikes me is this notion of Reverse ETL (read more about Reverse ETL here or here). Reverse ETL comes in, sits on top of the data warehouse and the SaaS application and then allows you to analyze the data very simply. Now those are, at a high level, sort of the biggest categories that we see today.
Bill’s Reply Summarized – An Industry View on the Data Supply Chain
My role is within the media, entertainment and advertising verticals. And so I certainly agree with Tom and Redpoint that the decade of data is certainly upon us. I think this data wave is probably going to extend longer than just the decade. It’s a great frame of reference because the next seven, eight years are going to determine a lot of new things.
And the stack and the supply chain that Tom just talked about is very much how we look at it at Snowflake. Not only is the data supply chain evolving, and certainly Simon is part of that, as well as Snowflake. The industries that we’re serving, their supply chains are also changing and evolving. And so when you have the dynamic of evolving data supply chains and evolving industry supply chains, it creates this shuffling effect that creates a lot of opportunity for the marketplace and the investment side and in companies like Snowflake.
An Example from Bill
So let me just give one example. In my 25 years of media, most of the big brands like Disney or NBC or ESPN or CNN or HBO, were what we call wholesalers. And the end customer’s relationship was with the cable company or the satellite company. So the supply chain existed to have data that came from the end customer point. But the wholesalers didn’t have a lot of data because they didn’t own a customer relationship. Now that supply chain is switching and many companies are establishing a direct-to-consumer, in this case, streaming relationship. And that is creating a new set of supply chain data customers that need to adhere to what Tom has described as the data supply chain.
What we start to see is that the data warehouse is starting to blur the line between the upstream and downstream data supply chain. So the point that I’m trying to make is that I’m not even sure convenient definitions, like data warehouse, are really even appropriate anymore because at least from a Snowflake perspective, we’re seeing data collaboration and data sharing and applications come to the data stack. And that sort of starts to change definition.
So maybe the one prediction I’ll make, is that I think what we call the data supply chain and even the names that we refer to them as are going to change and also blur so that new partners, new companies, new situations are going to emerge, and frankly, the companies, that take advantage of this. What we’re doing at Snowflake, positioning ourselves as a data cloud, not a data warehouse, will drive this evolution.
This conversation and more on operational analytics continue on the Data Unlocked podcast. Tune in here: