Q&A: Shifting Emphasis on Speed over Consistency Changes Data Landscape
An adaptive decision-making process can bring together business factions and help knit different processes together. Well-known author and consultant Barry Devlin explains how in this, the second part of a two-part interview.
- By Linda L. Briggs
- February 24, 2015
Businesses now emphasize speed over traditional BI's predefined reports and queries. In this second of a two-part interview, well-known BI author and consultant Dr. Barry Devlin explains further details of his "Adaptive Decision Cycle" process and how it can be used to join disparate decision-making segments of the business to accelerate data and decisions.
BI This Week: As you described in a recent TDWI Webinar, your adaptive decision-making process essentially bridges two approaches, the edge-on, spreadsheet-based decision cycle and the center-out, data warehouse-based decision cycle, correct?
Barry Devlin: Yes. ... How can we get past the idea that we needed to be able to have flexibility and the ability to adapt within the edge-on decision cycle in order to get innovation and at the same time have the level of consistency and quality control that [we had with] the data warehouse?
It really proves impossible to figure out a way that you can have those two things in a single process. You simply can't, so essentially, if you look at the adaptive decision-making cycle, what you find is that the two older, more basic decision cycles are embedded in there -- and they happen in sequence. We start with the idea that a decision-maker has a problem and explores that problem using the data, then goes through this process -- which is analogous to the spreadsheet world -- and then eventually to peer review and to promotion in a collaborative sense with IT. ... It becomes more of an adaptive process rather than a completely new architecture.
Is it correct to say that most companies practice both decision-making styles?
Absolutely, and the old debate that has gone on for many years about spreadmarts is, of course, the view you have if you live in the central BI organization and you look at what's happening in the spreadsheet world. Of course, you have a different perspective from the spreadsheet world, or what I called the edge-on decision-making cycle. From there, you look at the central BI cycle, and you put your head in your hands and say, "Why can I never get the data I need?" Those two styles exist in pretty much every organization. My belief is that the adaptive decision-making process knits those two processes together in a sensible way.
How is the tool landscape merging to combine data discovery and business intelligence features?
I don't dig into tools in deep detail, I have to say. I'm more of an architecture person. However, at a high level, what I see is that the data discovery tool vendors, the QlikViews and Tableaus, for example, have already made their names in terms of doing data discovery -- in terms of making it easy for the user to play with the data, to visualize the data, to get on with the job of what I call the edge-on decision cycle.
However, if you look at what those vendors are doing with their products now, they are definitely moving in the direction of saying, "How do we support IT in getting the data to the users?" From those companies, we're seeing a move from that data discovery view of the world to a more centralized and data-quality-directed view of the world.
Then we have the older and more traditional vendors, Business Objects and Cognos and so forth, bringing in the idea of data discovery to their tools while at the same time maintaining the data management and quality management aspects.
You see things going in both directions, depending on the starting point of the particular vendor. It's a pendulum, as we said earlier, swinging between the need for speed and innovation on one side and the need on the other side for quality and consistency. It's a situation that's going to be with us forever.
What business drivers are required for the kinds of speedy and flexible decision-making today's business environment requires?
With always-on, mobile connectivity today, customers demand instant gratification. It applies to everyday transactions, enquiries, and, in particular, when anything goes wrong. It's not just a case of timeliness in business-as-usual transactions, but also in those that demand some decision-making, either automated or manual. Needless to say, the emergence of pervasive Internet of things use cases will drive this need for speed even further.
What are your clients struggling with in terms of moving to an adaptive decision-making process?
I always find that it's more about process and organization than tools. I often say to folks that more projects have died because of organizational issues than have ever failed because of the tools, and yet we technology folks tend to invest so much time and energy in tool evaluation and tool selection. The real question is one of organizational skills -- the ability to get things moving, to get the organizational aspects working properly. You know, data quality and consistency and data governance are things that are very difficult to get the business excited about, and yet they are at the core of making sense of the business in a consistent and meaningful way. The issue often is this: How do we get the organization to understand the value of both sides of the equation -- the value of speed versus the value of quality?
For example, in an organization where you have a strong marketing department and a strong finance department, the finance department will be, in a way, focused on quality and consistency while the marketing folks will be focused on moving things forward and innovation. Getting those two functions within the business to agree on priorities is often quite a struggle.
What is a good first step in those situations?
It's a specific company-by-company decision. You have to look at how the culture of the organization can be moved forward. Who are the people who are the roadblocks? It comes down to consulting and counseling rather than architecture and design. It's the kind of thing that management consultants really should be doing, but I've found in many cases that they don't do. It's tough work. It's almost like relationship counseling. You have to work with people, as teams and individually.
What does your book, Business unintelligence -- Insight and Innovation Beyond Analytics and Big Data, bring to today's discussion?
One of the key concepts of the book is this idea that there are three conceptual spaces that we need to think about -- information, process, and people. I often play it as a sentence in the opposite order -- "People process information" and that becomes very important when you think about the decision-making cycle and about moving decision-making BI forward.
To make it work, you have to start with information. You have to make sure that the information is both consistent and delivered in a timely fashion. You have to have a process that enables you to make that information available in order to make the decisions work and go back out into the organization. You have to educate (draw out, in the real meaning of the word) people in order to get that to work. A key part of the thinking behind Business unIntelligence is this idea of balancing people, process, and information.
It's an interesting title -- Business unintelligence.
We've talked about business intelligence for many years, but this idea that we have of intelligence is very much based around logic and around rational decision-making. Yet, if you think about most organizations, so much of the decision-making is based on politics, it's based on intent, it's based on all sorts of things beyond rationality -- even intuition. I wanted to get something into the title that both echoed the idea of business intelligence but emphasized the fact that in some ways we're not doing just intelligence. Really, we're doing something other than classical intelligence. It's not an either-or, but instead, more of a both-and.