Rob Schmidt, chief data officer at Dell, knows when IT needs to get out of the way. Over the years, he has worked in application development, application support, IT operations and BI. In each role, he's seen that as technology allows business to move faster and faster, IT runs the risk of becoming an obstacle rather than an enabler.
In his mission to build a data-driven culture at Dell, Schmidt has played an essential part in transforming the relationship between IT and the business. His initiatives have transferred many of IT's capabilities to business users, empowering the business side to perform data analysis and develop its own applications.
"I think the next shift is going to make analytics available to the masses," Schmidt said, pointing to retail website features such as "People Like You Buy," which gathers and analyzes data to make personalized shopping recommendations. "In the very near future, analytics is not going to be a differentiator. It's going to be a requirement."
He emphasized that this shift creates a different function for IT. Information management will move away from application development and operations and into a consulting role, with a focus on the challenges of storing, integrating and securing data.
Schmidt sat down with SearchContentManagement to discuss Dell's approach to data analytics and the transformation of enterprise IT.
Why is it such an imperative for business executives to have better access to data now and to be able to do analysis on their own?
Rob Schmidt: The amount of data and the ability for us to change that data into action is at a phenomenal level right now. With all the things that Dell is doing around social, if I can't tie all the data we have within Dell to the social media structure, we're missing the new way to sell and to service our customers.
You could lay that on other perspectives. If I can't predict hardware failure, then I'm missing a play to be able to service my customers. If I can't predict when a customer wants to purchase something -- when they've downloaded a white paper, that they've gone on Dell.com, that they've made some social media posts around, 'Does anybody know the best storage solution for …?' -- if I can't put all those together, I'm missing that play.
So is the idea that the business side can operate as data scientists, looking for patterns or outliers, without having all that expertise?
Rob Schmidtchief data officer, Dell
Schmidt: Yes, but what we've really done is take out a significant amount of work, and we've given them a significant amount more data. Historically, they would pull out billions of rows of data from the data warehouse and put it into shadow IT databases. They would spend an inordinate amount of time converting that data and putting it back in. But they only had the data that they pulled. Let's say they pulled all laptop sales for North America for one year. That kind of limits your discovery.
We've said, 'Move to our platform. I will give you access to do whatever you want on my platform and you get instant access to all of the data within Dell so you don't have to move data around. And you get the power behind Teradata, the power behind Aster, the power behind Hadoop, as opposed to the SQL Server database you were running on.'
We're seeing cycle times for some of our queries or events going down from 16 days to five minutes because they're on this platform.
So tell us about how you use Teradata, Aster and Hadoop as data analytics tools.
Schmidt: Teradata allows us to put big chunks of data on Hadoop, structured or unstructured, at a very low cost. Clickstream is a good example, where I can remove a significant amount of cost from Teradata because Teradata is a higher price point and I can put [that data] on Hadoop.
So Hadoop allows me to store different types of data at a much cheaper rate, and then Aster gives me the ability to do that data discovery, that statistical analysis, that model creation.
How has this changed the cultural dynamics at Dell in terms of the business owning analytics more?
Schmidt: It was a huge change for us. If you think about me going to my CIO and telling her a year and a half ago, 'I'm thinking about implementing this process that removes IT from the IT development process,' most IT people start to cringe a bit. How do I make sure I have consistency, how do I make sure I have the testing and automation correct? We had to set up that framework to make sure that all still occurred, while giving the business the ability to react quickly.
From a business perspective, it's also a big shift. They had to start hiring database developers, people who had historically been in IT, but were now in the business because they own this analytics space themselves.
The second part is, how do we change the Dell culture to be a data-driven culture? About a year and half ago, we brought together an executive steering committee for data and analytics that consisted of four VPs from the business and myself. The five of us really drove out how we're going to use this data.
The sales leaders no longer get to define their own opportunities -- we do that once for the company and everybody has to adopt it. And once we get to that rationalized set of metrics, we can now start making data-driven decisions based on valid, true, governed data as opposed to a sales leader creating their own metric and showing success.
For other companies that are trying to wrap their brains around shifting their analytics model, do you have some advice?
Schmidt: Number one, it has to start at the top of the organization. This is not something you can do from a grassroots effort.
Our CIO and CMO came together -- Andi Karaboutis and Karen Quintos -- 18 months ago and defined me as the person who stands between them. It took somebody to stand in that space and drive a cultural change, a spending change. And it took the two of them agreeing that I had the leeway to drive this stuff.
It also took an executive steering committee that just didn't take no for an answer.
And how did you get adequate resources for this initiative?
Schmidt: I got to -- and within Dell, this is unheard of -- I got to deliver an IT capability before there was a documented business requirement driving it.
I put the ROI model in place. Once the business started to show significant cost reduction and a huge ROI --- close to 1% of revenue -- I had the ability to change how we grow that architecture.
Does intuition still count in a data-driven world?
How a chief marketing technologist could align business and IT
Content analytics for better business intelligence