CIO: Say you've made an overseas acquisition, and there is a system there that may or may not be better than what you've got already. What are the steps you go through in evaluating that?

Turner: When we acquire a business, we start by defining success with the customer. What does success with the customer look like? We haven't always been good at determining that, but we've put a lot of emphasis on it in the last 18 to 24 months. This is critical so that everyone has the same definition of success.

The next step is to eliminate before we automate. Eliminate steps, processes, reports, keystrokes; eliminate any activity that you possibly can for two reasons: one, you'll end up building a whole lot better system that's easier to support, and two, invariably you will have a better solution that's more [user] friendly.

Next is to do a business process overview, where we physically lay out what the system looks like from a business process standpoint. We validate that with the best business process experts in our company before we start the design. Once we get the system built, then we pilot it with our best and toughest customers, stores, distribution centres and clubs. If you pick the lowest volume store or easiest customers to test with, you don't find all of the things that you need to find.

CIO: That approach can work only if your people truly understand the business. How do you make that happen?

Turner: In any development effort, our [IS] people are expected to get out and do the function before they do the system specification, design or change analysis. The key there is to do the function, not just observe it. So we actually insert them into the business roles. As a result, they come back with a lot more empathy and a whole lot better understanding and vision of where we need to go and how we need to proceed.

We've learned some hard lessons in our stores, clubs and distribution centres when we developed something and people didn't use it, and they chose to find other ways to get the job done. We are working hard to develop systems that are easy to use. That puts an awesome responsibility on that developer to get out and understand the business. That's one of our key things: We're merchants first, technologists second.

ON EXECUTION

CIO: Getting beyond the up-front process, how do your people actually execute? Do you have principles of execution?

Turner: That's something we talk a lot about because, as a company, we're big on execution. You can't name a great team that wasn't good at discipline. Execution and discipline are not the enemies of enthusiasm. We have a very clear set of critical success factors that every associate in our division has to live by, and they are generally conditions of employment. Some of them are complicated and some of them are very simple: excellent customer service, testing and validation before you roll it out, balance and controls, payback and ROI. All new [IS] associates are indoctrinated into that set of disciplines, and all associates review it at least on an annual basis. The disciplines are the same for everybody regardless of what team you're on - infrastructure or development or help desk.

CIO: How do you recognise failure of execution?

Turner: Well, there are a lot of ways to recognise failure, such as if a project is missing end dates. We don't call them completion or target dates; we call them end dates because everybody should have the same definition of end date. We track projects against the end date and surface those things that are behind or in trouble. The second indicator comes from on-the-job reviewing and listening. I and others who are not on the project team go out where these new systems are deployed and talk to the people doing the work. We ask them how it's going, how things are working, what they'd like to see different. That's where you really surface feedback that can help project teams get even better.

Sometimes we have to make management changes to bring a project across the finish line. It's hard because no one wants to give up on a person or a team that's struggling. But there are times when the right thing to do for everyone involved is to change the quarterback and put somebody else on the team.