The Sentient Enterprise: The Evolution of Business Decision Making by Oliver Ratzesberger & Mohan Sawhney

Author:Oliver Ratzesberger & Mohan Sawhney
Language: eng
Format: epub
Published: 2017-08-23T14:30:00+00:00

PROACTIVE DATA STANDARDS AND DESIGNING FOR THE UNKNOWN

You probably noticed that we’ve made more than one reference in this chapter to artistic realms like sculpture and live theater productions. These references are intentional: the data scientist looking for patterns amid all the noise is blending mathematical and engineering expertise with no small amount of artistry in making innovative assumptions and coming up with novel theories and insights.

That creative guesswork can be time-consuming, however, and a certain amount can be avoided if we do a better job standardizing the way we collect and use data to begin with. We need to record and manage data with an eye toward how it’s going to be analyzed in the future, even if we don’t know exactly what that future will look like.

We need, in other words, to design for the unknown. The roll-ups and optimizations we talked about earlier in this chapter may work okay for today’s known applications, but our analytic architecture and design choices today should not inhibit new and currently unknown use cases tomorrow.

For example, technologists often place clickstream data into session containers on file systems to make things like bot detection or per-visit analysis easy. But doing so sacrifices entirely different use cases— situations that may not start with individual website visits such as financial checkouts, call center interactions, or machine/sensor events.

Whenever we place data in a certain order, grouping, or structure to allow for one type of access today, chances are that we may complicate things tomorrow as patterns and questions shift. That’s bad news, because we cannot afford to redesign or rearchitect the data over and over. Yet, this is what’s most likely happening right now in your company over and over.

If we remedy such issues—standardizing wherever possible, avoiding roll-ups, and preserving the granular detail when staging data for analysis—we reduce friction, enhance governance, and make our organization better and faster at wrangling data. Few companies today do this really well, meaning anybody who is diligent about standardization can get real competitive advantage.

Dell is one such company. In Chapter 3, we heard from Vice President Jennifer Felch about Dell’s evolution toward global manufacturing processes and infrastructure. To its credit, the Dell analytics team understood clearly and early the value of standardization. “When we organized the company into some globally shared capabilities, that’s where we had this realization that there were different processes, and even different definitions, of key data pieces,” she told us.

“We had regional general managers, regional manufacturing, and regional IT systems,” she said. “When it came time to set up a global manufacturing site, we had to ask, ‘Are we going to support three or four different inputs and processes, or are we going to standardize before we go? Do we spend three times the integration costs, or do we just do it right from the start?’ And so we decided to standardize from the very beginning.”

We’ve spent the past few pages looking over the shoulders of experts who first examine our data for patterns and context.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.