Productivity and Plain Vanilla Software

I’ve heard it often: “I’m installing [ERP, CRM—practically any enterprise software you want to name] plain vanilla. No modifications, no customizations.”

The argument for plain vanilla is compelling: the fewer customizations you make to an enterprise software system, the lower your support costs are and the more likely you’ll be able to upgrade to the software manufacturer’s next release without painful rewrites of custom code.

There are two problems with this logic: One is that it assumes that the business processes embedded in the software are better than the ones you use currently—which is not always the case. Secondly, it assumes that the processes dictated by the software will lead to increased productivity among the people using them. We’ve seen very little evidence to support that theory.

Productivity is the higher goal that gets lost in these software projects. A much more tactical goal, the quest for low-cost upkeep of the software, tends to take precedence. Part of the reason is that productivity—the amount of output per worker—is so hard to measure. Support and maintenance costs are black and white. But it’s the wrong way to look at these projects. Look, in ten years of covering this stuff, I’ve never heard a CIO—even the ones who made plain vanilla an iron-clad rule from the start—get away with less than 20 percent customization of the software. It’s just too difficult to match one of these packages to the typical mid-size or large company.

Economists and IT pundits are divided as to whether IT has any real productivity impact at all except in the precipitous price drops for computing equipment and bandwidth. But neither of those things have a direct productivity impact on how people do their jobs, unless the business processes attached to all those cheap computers improve somehow. The business reengineering craze of the 80s and 90s—which focused on rewriting work processes to improve information sharing and flow—foundered on the rocks of internal change management and the use of enterprise software as a battering ram to enforce business process changes.

Eric Brynjolfsson is outspoken in his belief that there is an improvement in productivity and competitiveness among companies that spend heavily on IT. He did a study of 1000 companies where he found a “loose” correlation between IT spending and productivity—meaning not all companies that spent heavily on IT were productive, nor were all who spent little not productive. But Brynjolfsson, who is management professor and director of the Center for eBusiness at the Massachusetts Institute of Technology’s Sloan School of Management, emphasizes process change, rather than software. In a recent interview with Gartner, he identified seven characteristics of the “digital organization” from his study. Here’s an excerpt from Gartner’s interview where he discusses them:

…We saw a pattern that emerged. The first practice we saw was an obvious one. In fact, when we first did it, we didn’t even list it, but now we break it out explicitly, which was that they converted a lot of analog work practices to digital work practices. They moved from paper-based systems to electronic systems.

I mentioned Cisco earlier. There’s very little paper in that organization. Brad Boston, the CIO, says he’s signed only three pieces of paper in the past year; everything is done electronically. A lot of other firms have done that. And that speeds up the organization’s metabolism quite a bit. That’s to be expected when you go digital. A lot of firms just sort of stopped there and said, “OK, now we’ve gone digital.”

But the other six practices really had much more to do with the human-resource side of things. One was the distributing of decision rights throughout the organization more broadly. Interestingly, in some ways this seems to contradict the first one, because going digital often entails a lot of centralizing of decision-making and rule-based systems and Oracle databases.

If you look more closely, you see there are certain kinds of decisions—quantitative, rule-based decisions—that can be centralized through a computer-based system. At the same time, decisions that involve human interaction, judgment, exception processing and creativity tend to get pushed out into organizations—out to the people in the field or out to the factory floor, for that matter. So, different types of decisions go in opposite directions.

On average, for the main decisions we asked about, they tend to be distributed to line workers rather than to the center and top of the organization.

A third practice we found that goes with the decision rights is strong performance-based incentives. And that helps support distributed decision rights. If you’re going to give people the authority to make decisions, you have to make sure they’re making them in the interest of the organization.

Another practice is a policy of much more open information access and communication. So, we found that these organizations had more lateral as well as more vertical sharing of information. Every dimension we looked at had more sharing of information through e-mail messages, intranets, etc. Again, you can see how this supports distributed decision making. If you’re going to give people the authority to make decisions, you need to also have them have the information available to make those decisions.

We also found that these firms were much more focused in their business lines. They tended to prune out non-profitable lines and invest heavily and explicitly in corporate culture. At Cisco, they have a director of corporate culture.

Again, I see this as a necessary complement to having people have more information and authority. You have to do what you can to keep them on the same page. Performance-linked incentives help with that. But there are some things you can’t incentivize and quantify. So, another way to keep people on the same page is through having a corporate culture and a relatively good focus in terms of what your goals are for the organization.

The sixth and seventh practices had to do with the kind of employees that firms hired and what they did once they got there. These firms were much more likely to invest heavily in hiring the best employees. Of course, everybody says they want to have the best employees. But these firms tended to spend more money and time on recruiting. They interviewed more people for a given spot. They were more likely to have top and middle executives actively involved in the recruiting process, rather than delegating it to, say, an HR department.

We asked about various standardized tests and what sort of thresholds they had. These companies tended to have higher thresholds in analytical ability, verbal skills, teamwork and people skills. In all of those dimensions, they put the thresholds higher. They tended to hire more educated workers, people with more experience.

And then the seventh practice had to do with what they did once those employees got there. These firms invested more money in training once they got there to further increase their human capital—more days per year training and more use of Internet-based training systems. So, when you put it all together, they tended to put more effort into having really good people, giving them authority to make decisions and giving them information to make the decisions—incentive systems that support that, a culture that supports that. And it all seems to fit together with the pattern of a high metabolism—a high information metabolism organization that keeps track of what its customers, suppliers and environment are doing. And it responds to that using information technology.

You can see that very few of these practices have to do with installing “plain vanilla” software. The ultimate goal of productivity has more to do with the “Monster Mash” of IT, people and process outside the software project. What do you think?