Intelligent Enterprise

Don Campbell helped develop PowerPlay Web, the first Web version of the Cognos OLAP tool, and also spearheaded development of Cognos Visualizer, the company's three-dimensional, interactive dashboard product. A 17-year BI veteran, the VP of product innovation and technology at Cognos talked to Business Intelligence Pipeline about how to get performance management buy-in, the challenge of selecting bpm metrics, and where Cognos hopes to fit in a market where customers increasingly want to consolidate on fewer BI tools.

Business Intelligence Pipeline: Performance management initiatives are obviously much broader than simple BI deployments. What strategies can IT use to get performance management buy-in from key decision-makers?

Campbell: Some of what we talk about in the performance management arena can be very culturally disruptive. To be able to see all of the information across all the departments and how they're performing against strategic targets -- some people are very sensitive about this exposition of knowledge and information. So to try to do that as a big cultural enterprise initiative is a lot to take on all at once. It's much easier when you've got a specific pain point. For example, maybe you can't do your job because you can't get the numbers you need around something like inventory. This is a pain point. So IT should deliver a project that's quick, where you can get ROI in 30 or 60 days. Then you can show that there's more and more you can do with the technology and get a little creative about it. Those projects tend to create internal champions.

It's changing. In years gone by, we would have described full performance management as a top-down initiative. And I think it certainly has to have sponsorship, and there has to be awareness from the top. But I think now it's developing more across different lines of business in an organization, where they're realizing the success they have is often driven by the information they can get from other parts of the organization. Whether they're upstream or downstream, they have to understand how they fit into the entire ecosystem of the organization. Performance management is a very big topic, a very large thing for a company to swallow. And yet, everybody wants it. I think it's becoming a little more real for organizations.

Business Intelligence Pipeline: Many companies engaged in performance management don't know what metrics they should use to gauge their success. It's the determination of metrics rather than any actual deployment of technology that presents the real challenges. Does Cognos provide guidance there?

Campbell: That's a really interesting comment, and I agree with it. I do presentations, for instance, where I demonstrate our metrics manager tool. And then companies come to me afterwards and ask, "How do I put this in place? Can I have it in place by the end of the week?" I have to slow them down a bit and tell them they need to understand their business, the metrics they want to track, the accountabilities, the network of relationships between these metrics. How are you going to actually determine metrics' values? What is a good target or a bad target? That's a big part of making this system work.

Unfortunately, we can't do that for companies by ourselves. But we certainly have built some best practices around how we've seen it done effectively. We recently bought a small consulting company in the United Kingdom, Softa Group Ltd., which allows us to do strategy formulation. Companies might ask us, "How do we describe this strategy we've got, and how can we turn it into a metric presentation?" We're starting to understand that we have to take a leadership stance there, guiding and helping companies through this. We'll do a lot of this through systems integrators. And we'll use our prepackaged applications and their ability to describe metrics throughout an organization for vertical kinds of applications. More and more, we're being asked our opinions on how to do things like determine metrics. I would say we're still in the early days of it, but we know our customers can't pull off a performance management initiative by dumping a box of software on somebody's desk and saying, "Figure it out. The manual's inside."

Business Intelligence Pipeline:You just mentioned best practices around metrics. Share some of them.

Campbell: We believe very strongly that every single metric should have one person who is accountable for it, and only one person who is accountable for it. It sounds simple, but it's key when you try to put everything in place. Another thing we believe very strongly is that metrics are not strictly hierarchical. They form a network. Metrics have relationships that are fairly complex. So one thing our tool tries to do is present the impact that changing a certain metric will have on the entire system. It's not the case that you just want to look at all the metrics that are red on your screen and simply make them better. If you could do that, it would be great. But it turns out that there are a number of interrelated levers to a complex system like that. Just pulling one lever doesn't necessarily solve a problem. It might create four other problems. People have got to understand the reality of this complexity.