Teradata chases hyperscaler, SI partnerships in cloud push
We're not just your grandmother's data warehouse, CEO tells El Reg
CEO Interview Cloud ambitions came with a plan to pull out of on-prem consultancy and services at Teradata, the data warehousing specialist whose userbase includes some of the world's largest banks and retailers.
In 2021, The Register revealed that support for Teradata's American hardware customers was transferred to IBM, to the concern of customers who feared losing expertise.
Now, CEO Steve McMillan has told El Reg the move was made globally. "We don't want to be a services company. We want our core competence to be our technology platform. The future at Teradata is in the cloud," he said. Customers continue to be happy with Teradata's on-prem support, which includes software support and consultancy, aside from IBM's assistance, he said.
"We still have a strong team across the world that are highly integrated into some of our customer environments to make sure that they're getting best from the platform," he said.
McMillan took over as CEO at Teradata in May 2020 following the nine-month tenure of Oliver Ratzesberger, who had overseen the launch of the company's cloud-based data warehouse and analytics platform Vantage and the transition to object storage.
A string of disappointing financials followed. But the company managed to double its public cloud revenue to $202 million for calendar 2021, although the majority of its business remained on-prem.
In its Q4 2022 results, total revenue fell to $452 million, down 5 percent on a year earlier. However, the company boasted that annual recurring revenue (ARR) for cloud was up 45 percent in a year and total ARR growth up $108 million quarter on quarter, driven by "outstanding" cloud migration activity.
It's something of a turnaround for the company that helped found the data warehouse concept by building hardware-integrated BI and analytics systems from 1979 onwards.
"We're not just your grandmother's data warehouse: we now have a data lake capability; we've got a full multi-cloud capability that spans across on-prem and cloud environments but also cloud-to-cloud environments; we've got a fantastic analytics engine," McMillan said.
It did not hurt that the market was getting a lot of attention, with investment interest in Snowflake, Databricks, and a string of startups in the cloud-based data management and analytics space.
"We're in a great market data and analytics market. All the surveys are pointing to the resilience of spend in data and analytics," McMillan said.
As well as moving the company to a new platform and business model, he has been forced to navigate a geopolitical storm. Teradata counted Sberbank among its customers – the majority Russian-state owned financial institution even presented at the company's Universe conference in 2019.
McMillan said Teradata had completely withdrawn from Russia following its invasion of Ukraine, and offered no hardware or software support to companies with on-prem systems it sold there – including Sberbank. The move saw the company take a $60 million hit in 2022.
As if these challenges were not enough to contend with, Teradata has opened up a whole new front of competition with its strategic shift to becoming a cloud-first business. As well as facing competition from so-called cloud-native vendors such as Snowflake – which stunned the software world with a one-time valuation of $120 billion – Teradata must confront the triumvirate of leading cloud vendors, each of which offer their own interpretation of the cloud-based data warehouse.
The company must strike a tricky balancing act because Teradata also partners with Google Cloud Platform, Microsoft Azure, and Amazon Web Services to win new accounts or transition existing customers to the on-trend computing model.
- Teradata CTO Stephen Brobst drowns data lakehouse concept
- Teradata takes on cloud-native rivals with data lakes, MLOps
- Teradata to take $60m hit for withdrawal from Russia
- Teradata: Public cloud sales soar from low base, majority of business still on-prem
McMillan said American Airlines was an example of how it worked with Microsoft to execute the cloud transition of its technology. The global carrier has been a Teradata customer for more than 20 years. Several years ago, it migrated its on-premises system to Teradata's VantageCloud on Azure.
"Working with the Microsoft team on American Airlines, we put together jointly the best possible solution to move that workload. I don't think the Microsoft cloud-native services could deliver [that] and certainly not at the kind of scale of operations that Teradata can deliver on Azure or for American Airlines," he said.
The company had grown a large services base of revenue. Still, it has been "declining that consistently" to help foster relationships with SI and cloud vendors to influence and gain customers, he claimed.
"That's what SIs want to be able to deliver to our customer, and that's what a CSP wants to deliver to a customer. So American Airlines are a really good example," McMillan said.
As data warehouse, data lake, and analytics vendors converge on the cloud, distinctions between their portfolios blur. Databricks, for example, built its reputation as a data lake vendor for unstructured and semi-structured data, based on Apache Spark and Hadoop. But as it moved to the cloud, it began to offer data warehouse-style features, supporting SQL queries for concurrent users of its so-called "lakehouse" environment.
Teradata has followed the path of Databricks, Snowflake, Cloudera and others in promising to apply its analytics engines to data regardless of whether it is in its data warehouse or not, seeming to abandon the idea of a "single version of the truth" data warehouse.
The capability arrived in December 2020 with an update to Teradata QueryGrid that integrated a new Presto connector so that users of Teradata's Vantage analytics platform can access and query a gamut of cloud and on-prem data sources.
"Our general principle is use the data, don't move it. So we would say put a query engine in our Teradata Vantage instance next to and close to your data. We're doing a lot of work to make sure that our customers can leverage cheap native object storage," he said.
But to access Teradata's secret sauce, users need to get data into its environment. It built a bunch of stuff to do with optimization – primarily to help support many concurrent user queries hitting the same database within a constrained hardware environment. Since its shift to the cloud, it has touted this approach to optimization as a way to mitigate unexpected costs which have dogged other cloud-based data warehouse vendors from a user perspective.
"Our workload management and query optimization essentially enables our customers to optimise their cloud spend. We enable our customers to really control the size of the environments that they're running in the cloud, and to optimise the workload that's in that fixed capacity environment in the cloud so that their costs don't run out of control," McMillan said.
For now, Teradata's reinvention seems to be paying off. It has beaten investor expectations and saw a jump in its valuation following its latest results. But it is betting it can make the leap from doing one thing – arguably very well – to taking on a whole field of competition at doing many different things in a market they established. It might have turned a corner, but the road ahead is far from certain. ®