The Teradata Universe event and beyond - where does Teradata's transformation go from here?
- Summary:
- Data warehousing stalwart Teradata is in the midst of a provocative and difficult transformation. Here's my take on how they are re-positioning around analytics, big data and cloud.
Since the Teradata Universe event in October, Teradata released missed earnings with its latest third-quarter results. While revenue fell marginally short of analyst estimates at US $459m, earnings missed forecasts by a whopping 46%, coming in at just US$0.09 per share.
And then they removed President and CEO Oliver Ratzesberger, and removed him from the Board. Acting CEO Vic Lund, who had originally promoted Ratzesberger to CEO, takes over.
I know Ratzesberger. I like him and admire his skills and accomplishments, but being at the top of a public company going through a do-or-die transformation is a treacherous place to be.
Now about that transformation. Until three or four years ago, Teradata was the premier data warehousing company for the biggest data warehouses for the most prestigious organizations. Teradata was expensive, but if you had a data warehouse that was too big or too busy for alternatives, you paid the price. Revenue was a nice combination of services, software and hardware. Then something happened.
Other relational databases for data warehouses challenged, especially the MPP (Massively Parallel Processing) aspect that Teradata dominated. Commodity hardware and software based on open source opened up a gap. It seemed like Teradata would prevail, because the others could not compete in terms of functionality, and the cost to move was prohibitive. Teradata data warehouses were (and still are) strategic in many organizations, and switching was risky.
Then came Hadoop. Teradata was under pressure because Hadoop was proposed as a data warehouse replacement. Everyone thought it was cheap. Cheaper. But not for long. Organizations that implemented Hadoop for important applications, wanted premium storage, premium servers, so the cheap part of the proposition vanished. And the primary mission of a data warehouse is to support analytics, but Hadoop was suboptimal for that. It turned out Hadoop was not a replacement for a data warehouse.
Then the cloud. Teradata had to come up with a cloud strategy, but that would, of necessity, eliminate the hardware component. Plus, licensing for the cloud worked against the old perpetual licensing that was rapidly fading from the enterprise. To their credit, Teradata did deliver a robust cloud solution.
That is where the transformation comes in. Teradata Anywhere. Cloud. Data Center. Hybrid. Multi-cloud, and all subscription, not perpetual.
I came away from the Teradata Universe conference cautiously optimistic that Teradata was making good progress on their transformation. But here is what they did. No longer "the data warehouse company," Teradata is now an analytics company and their branding is all wrapped up as Vantage.
What is Vantage? Under the covers is still the Teradata database, but they’ve finally integrated and enhanced the analytical capabilities of Aster (Time Series, Graph, Path & Pattern, Machine Learning, Text & Sentiment), aptly titled Analytics, which they acquired over ten years ago. Another component is the Customer Experience, and a third is the Hadoop migration offering. Lets look at these one at a time.
Teradata Vantage – "the platform for pervasive intelligence"
In Teradata literature, Vantage is:
Unified integrated environment for analytics, build with your preferred tools and languages, Work with flexible support for multiple data types, formats, and heterogeneous data stores.
Teradata Vantage is an enterprise grade software instantiated on multi-form factors (e.g., Public Cloud, Managed Services Private Cloud, On-Premises, Hybrid) that ingests multi-structured dataat scale and from diverse sources while enabling the native implementation of multi-genre analytics and visualizationthrough a variety of programming frameworks(e.g., SQL, Python, R) via a group of client interfaces(e.g., TD Studio, AppCenter, JupyterNotebooks, R Studio)”
So that’s a lot, but the question is, how many alternatives do organizations have?
Teradata's proposition is that by combining all the features in one platform, organizations can reduce the time to insight. However, they also heavily promote an open ecosystem for analytics with dilutes the all-in-one message.
I'll give Teradata the benefit of the doubt. They have always been a superb engineering company. But, they’ve ventured into an area where they have plenty of competition. Existing Teradata customers will probably adopt Vantage and Teradata’s flexible pricing.
But best-of-breed alternatives are going to be a stiff climb for Teradata to attract new customers. There are plenty of cloud-native data wrangling, data management and analytics products. Teradata’s contention is that integrating all of the pieces (best-of-breed) yields a whole lot less than the sum of the parts.
Teradata Hadoop migration
They start the pitch by dumping on Hadoop (which is a la mode):
- Costs more than expected
- Long time to get projects into production
- Limited in analytical scale
- Relegated to storing data versus gaining insights
- Infrastructure complexity
While these criticisms are true for many Hadoop implementations, Spark and Kafka and a host of other projects can provide respectable analytics. Nevertheless, it’s interesting how Teradata proposes to do this, which is essentially cherry-picking data from Hadoop and leaving the rest:
Teradata had an Oracle migration program for many years. Oracle customers who ran out of steam in their data warehouse migrated to Teradata and a substantial number of Teradata’s existing customers went through that process.
Teradata customer experience
Teradata's third major announcement was their Customer Experience application. I didn’t have time to drill into this; my only comment is that Teradata tried selling applications over a decade ago and it wasn’t successful, so I’m not sure why they are doing this again.
My take
Teradata revenue from existing customers will sustain them, now that they (customers) don’t face costly upgrades from “balanced nodes” which was a policy of upgrading software storage and proprietary compute together. How Teradata will attract new customers is an open question. The Teradata brand is not especially known for analytics. Teradata data warehouses typically passed analytics to the ecosystem. Revenue from Aster, before it was merged into Vantage, was most likely a small part.
One other thing that I found curious was the absence of their Big Data/Hadoop/AI consulting group, Think Big, which they acquired a few years ago, and as I recall, started to ramp up. I suppose it’s not so strange that, strategically, they veered away from Hadoop. I just don’t know what happened to that group, but in the analysts call for third quarter results, they said they expected consulting revenue to be down 25 percent.