Q&A: Teradata Sees Opportunity in New Technologies

Social media and unstructured data offer some exciting opportunities for your enterprise.

• By Linda L. Briggs
• 11/10/2010

Advanced analytics, the advent of social media, huge amounts of unstructured data -- all are having a dramatic impact on BI and data warehousing. In this interview, we talk with Randy Lea, vice president of products and services at Teradata -- a company that specializes in both data warehousing and enterprise-level analytics -- about what he sees ahead in trends and technologies.

Among his comments: The recession is driving up the use of analytics. “Demand by upper management to know what’s happening to the business probably goes up tenfold during bad times,” Lea says. “The analytics market in general is a pretty attractive one [right now].”

BI This Week: Is there an emerging technology that you think will particularly impact data warehousing and BI?

Randy Lea: One area that will be interesting is unstructured data and analysis. Google’s MapReduce comes to mind. Some people are calling that Big Data -- it depends on what your definition of big data is. It’s really about social media; it’s about sentiment analysis and what’s going on with social media.

Lots of our customers are coming to us for help in integrating the social aspect of what’s happening with social media, with the transactional reality of business that’s being captured in the data warehouse.

First, in many cases with data from social media, you can’t find out about the individual. You can perhaps find out about a group, and that can give you some insights.

However, what happens when you apply that [information] to your business, to the transactional reality of what’s going on? If someone is out there Tweeting or blogging about a product of yours, is it because they’re a very dissatisfied customer with a loud social media voice, or is it a trend that’s truly going to be impacting your sales? By capturing this data from social media and then integrating it into your data warehouse, where you’re also capturing the transactional day-to-day business, you can validate the gut feel of the social media with the reality of what’s going on in your business. That can help you determine the right course of action.

In some cases, it’s a product problem that you need to take care of it. In other cases, it’s somebody with a loud social presence who is not really impacting sales. In that case, it’s not really an indicator of what’s happening with your products or customers.

Are we still early into the curve in dealing with unstructured data, especially as it relates to social media?

I think we are. There’s plenty of interesting things happening in terms of trends and what customers are doing. With Facebook, Twitter, and others, one of the questions becomes, who owns the data? Who owns the analysis? Are [these companies] going to try to monetize it? Are customers themselves going to try to do this analysis, or will data aggregators -- who are using more traditional means today -- step in, then sell the trending data to companies to integrate into their data warehouses?

Are social media platforms, in other words, going to become another set of channels where data is collected, in much the same way that Nielsen collects data on television watching? Time will tell. I can say this: The level of effort, and the knowledge of systems, that is required by companies to analyze data collected from social media is not trivial, nor is it cheap. You’ll have to decide whether or not you want to invest in being an expert in that business -- or make plans to buy the data from a data aggregator and integrate it into your warehouse environment.

Speaking of the future, we’ve just had some interesting news -- that IBM is planning to acquire Netezza. IBM can be called a Teradata competitor for larger enterprise accounts, I suppose, as is Netezza in the smaller-company space. Can you comment on the news from Teradata’s perspective?

It is interesting news. My comments pertain to a couple of areas. First, it obviously validates the importance of analytics, and the investments that companies are making in this space -- one that we’ve been in for the past 25 years. It solidifies Teradata’s positioning as a leader in the analytics space, that we’re in a market that’s growing, and that there’s a great opportunity for us.

Being acquired by IBM is interesting. Who knows what the future will be for Netezza? Their technology is more hardware-centric … so the database would be a little bit difficult to integrate, I think, into existing [IBM] DB2 technology. If there were some integration, it would be rather costly and expensive.

Given that, are there plans to have two different code bases and two go-to-market strategies? That’s been a challenge for IBM in the past. Even with the DB2 product line on UNIX and Linux, they’re Intel-based, they’re P-series-based, and they’re mainframe-based. Now, Netezza joins the company of Red Brick and Informix in having [been acquired] by IBM. If you know the history of Informix and Red Brick, [IBM’s strategy was basically] to sunset the technologies.

It will be interesting to see what IBM’s plans are for the Netezza technology. They could go to a high-end/low-end strategy, but that gets a little confusing for the sales organization and the R&D organization. … It’ll be interesting to see how it’s all executed by IBM.

It puts us in a good position in the market, especially with our platform family, because much of the low-end and high-end is actually either enterprise or departmental. Because we have a standardized database with our platform family, we can sell anything from a low-end SMP -- what we call our data mart warehouse -- to our data warehouse appliance to our active enterprise data warehouse to our new SSD-only [solid-state drive] technology to our extreme data appliance. It’s all based on the same Teradata technology. The same skill sets, the same ETL code, the same data modeling, and the same BI and application development are all consistent from the departmental level all the way up to the enterprise data warehouse.

We think that’s a huge value proposition for our customers, [because] the acquisition of technology and software is typically only 25 to 30 percent of your overall investment in business intelligence. It’s all of the people costs, training, support, and development costs that add up. That’s one reason that having a standardized Teradata platform is a big value to our customers.

That’s an interesting point, that just 25 to 30 percent of a BI project’s cost is the acquisition of the product itself, and the rest is support and other items.

Yes, for some customers it’s even less than that. When you take a look at an average project, [you need to] include all of the people down to the business liaisons, whether they’re on the BI team or in the IT organization in general, include all of your ETL sourcing resources, [include] all of your BI reporting, and all of your training -- those are just some of the costs.

When we release new programs and applications, that’s where the majority of our projected costs [for customers] are. Ask the average company, “Do you have your overall project cost [locked down], and how much of that is the acquisition of the data warehouse environment?” You’ll find that it’s a relatively small part of it.

From your perspective, how heavily is the recession impacting what companies are spending on BI and data warehousing?

Much like in the years 2000 and 2001, in many cases they’re spending as much if not more. When things are going well, you can overcome a lot of your mistakes. During bad times, you’ll go out of business. The use of analytics [and] the demand by upper management to know what’s happening to the business -- that probably goes up tenfold during bad times.

[For that reason,] the analytics space is still top-of-mind for every CEO and business out there, so the analytics market in general is a pretty attractive one.

Where do you see the data warehouse appliance market headed?

It’s interesting. Our definition of an appliance is “purpose-built for the single vendor, with an integrated stack, and somewhat easy to use,” so when people ask me where I see the appliance market headed, I see it going huge. Teradata was the original appliance; all of our products are an appliance. They are just built for a different purpose; that’s what makes Teradata unique. ..

We’ve evolved over the last 25 years to an open system, all the way from our data mart appliance (which is a single SMP designed to handle smaller departmental entry-level projects) to our data warehouse appliance to our extreme data appliance. They are all appliances in that fit that single-vendor, integrated stack, built-for-a-purpose definition.

Today’s market validates that Teradata has headed in the right direction in taking the appliance approach for the past 25 years. We continue to go down that road. It’s been a good run.