Cloud Data Insights (CDI) had the opportunity to talk with Sean Zinsmeister, SVP of Product Marketing at ThoughtSpot, at the Gartner Data & Analytics Summit. We covered new trends in data access technology making data exploration easier than ever, how the cloud has driven these trends, and how ThoughtSpot is integrating with a partner ecosystem.
Interview has been lightly revised for clarity and readability.
CDI: Just before our interview started, you mentioned that a large part of your time was spent with partners. That caught my attention because not everyone begins their story with partners. Why are partners so central to your product strategy?
Sean Zinsmeister: You can think of ThoughtSpot as an experience layer or a data consumption layer. The ecosystem is important to us because you need a cloud data platform or a database to do analytics. We rely on partners to deliver the performant database needed for a search engine on relational data to have search-level speed.
When we first started, that performance didn’t exist. Customers had to centralize their data in a proprietary database to search it. That has changed dramatically in the last 3 to 5 years with cloud data platforms like Snowflake and Redshift, to name a couple. We rely on partners for moving the data and partners to provide the platform we can connect to directly through APIs.
Sean Zinsmeister: What’s very new is the two new types of partners we’re working with. One is the query accelerators like Starburst, which gives us the performance to deploy into a data lake. This technology combination gives customers a single point of access to their data architecture. The other technology we see more and more demand for is data catalogs. These build more trust, reliability, and governance into data access. The vendor landscape has grown quickly–a customer told me they evaluated 50 data catalogs before selecting one.
CDI: What’s the relationship between ThoughtSpot and a customer’s data catalog?
Sean Zinsmeister: How do you trust the search? The ThoughtSpot search engine interacts with a logical view in our product called the worksheet. That’s what leverages a customer’s data catalog and gives the search governance and reliability. Let’s say I do a drill-down on a map of California to look at sales data. I can get very fine-grained detail. Because of the worksheet, I’ll not be able to access anything I don’t have permission to access, and I’ll also know the quality of the data the search is bringing up. Is it third-party data? Unwashed data? You can put guardrails in place to bridge the data analyst’s interest with the business user’s. Maintaining interactivity with data and doing it in a governed state is a difficult balance.
CDI: Governance has become such a hot topic because there are now so many more people with different roles in a company that access data for data exploration. How is ThoughtSpot responding to this surge and expansion of data utilization?
Sean Zinsmeister: It’s not just that there are more users–the users are trying to do more things with the data, and that’s caused us to develop two products. We have a standalone SKU for self-service analytics, that is, in internal business use cases. The surprise is the rapid growth of uptake for our embeddable offering that lets you deploy outside the firewall. That allows customers to deliver a data product to external customers.
A data product could be a standalone point solution that provides simple search capabilities. It could be something embedded into a portal. Many of our larger customers often have a unified portal, a center of excellence for analytics where a business user can go to meet their analytics needs.
CDI: So, data used by more people in more scenarios. We also see that more roles are engineering and consuming data. Are you seeing that play out?
Sean Zinsmeister: If you hear “analytics as code” as a name or a mantra, it’s a real movement. Data analysts are being forced to be more technical. They’re finding useful ways to repackage their skills, asking themselves whether they can do more with SQL. Do they want to work like a software developer? Sure, but probably using the language they know, which is very often SQL. This is driving a lot more demand for what I would call an open platform. An open platform has flexible APIs so that the data analysts and engineers can take the technology and build whatever they want.
CDI: You’ve mentioned centralization, that is, centralization in the cloud. ThoughtSpot started as an on-prem solution. Now you’re selling mostly in the cloud. You have a lot of valuable experience migrating to the cloud. Does that help you understand your customers’ cloud migration? And do you ever become entangled in their migration strategies?
Sean Zinsmeister: We have undergone a massive cloud transformation, which sounds remarkably simple but is remarkably difficult. And yes, we run across a lot of customers who are in the middle of a migration and want to put off implementing a search and analytics layer until they finish. We like to talk to customers about bridging the gap as fast as possible. They can have a single federated exploration layer that any BI tool access, so they don’t have to wait.
CDI: Waiting. Not a popular word in the data world. Your technology depends on performance since you’re querying massive amounts of data. Is real-time data exploration, or near-real time, a requirement for your customers?
Sean Zinsmeister: If you look at it through the lens of the customer, financial services customers are looking at multiple data sources and need refresh rates of their cloud data platform in seconds. A retailer like Canadian Tire, which had to pivot to adapt to the new pandemic market, probably needs a 12-hour refresh rate. Canadian Tire quickly realized that fitness equipment and bicycles were hot items because we couldn’t work out in gyms anymore. The pandemic revealed how important the need for speed was, but the timing really varies. [Read about the Canadian Tire case study here]
In any case, static dashboards are not going to be enough for business. When you build a dashboard, you have an answer in mind; the dashboard is just visualizing it. But businesses need true data exploration that’s not about answers but about the next question and the next question. That’s where ThoughtSpot’s dynamic search architecture and search engine are a massive differentiator. It’s what lets us build interactive dashboards based on data that doesn’t have to be moved before it can be searched.
CDI: Looking ahead, what’s on your product roadmap in the near term?
Sean Zinsmeister: Work on search. We feel that we’re only 2% done. There are so many ways for it to manifest in analytics and we’re nowhere near done with those opportunities. For example, bringing in automation and AI. We have algorithms that work behind the scenes to enable” auto drill down.” When they detect an anomaly, they will drill into that and auto-generate the insights that were the drivers behind the anomaly. We believe there will be a big demand for automated business monitoring. Customers are realizing that BI is not a terminal endpoint. It can become an operations tool, meaning that the BI tool can talk to a parent application or another business system or service and execute actions.
CDI: Thanks for sharing your take on these data exploration trends. You are indeed living at the intersection of several movements–the democratization of data and its reliance on governance, the scalability of the cloud, the rise of the citizen data pro, and the business need for speed–and you’ve painted a picture of how these are coming together.
Elisabeth Strenger is a Senior Technology Writer at CDInsights.ai.