What makes the business case for a technology project? That’s a question that every enterprise asks many times, and the sum of how the question is best answered determines how much gets spent on IT and networking. It’s also true that the chances that a good answer exists determines which industry sectors are likely to consume the most technology, and what type they’ll consume. In short, it’s critical to understand business cases. While there’s plenty of material in how to write up something for the CFO to review, there’s far less available on what you might want to say, and even less on how to review tech project targeting to optimize benefits. Let’s try to fix that by referencing users themselves.
Let’s start with fundamentals. Almost all tech projects are justified by the impact the project will have on overall productivity. If the value of the gains exceeds the project costs by the margin set by the CFO, then the project is likely to be approved. Usually, a company has an “internal rate of return” (IRR), and if the project return on investment (ROI) exceeds the IRR, it makes financial sense. This is a kind of accounting-101 view of making a business case..
Digging deeper, how do you get this happy equation to work out? Let’s suppose we have a project that will cost a million bucks. Let’s say the company’s target is a 25% benefit gain over costs in three years. That means the three-year benefit would be $1,250,000. Where do we get that money? Let’s say that the company has ten thousand workers that are made more productive. What’s the chance the project gets approved? We don’t know, because we don’t know the unit value of labor of those workers. If we did, we could calculate how much time the project saved, and that number times the unit value of labor, equals the productivity benefit. Thus, all other things being equal, the best projects would target the productivity of workers with high unit values of labor. I target a thousand workers for my example, I’d need $1,250 per worker in benefits. It would be a lot easier to get that from workers with a unit value of labor of $75,000 per year than ones with a unit value of labor of $37,500 per year, right?
Well, not completely. Suppose our $75,000 worker is a steel worker on skyscrapers. How would we organize a tech project to make these workers more productive? We might find that our lower-compensated worker sat at a computer all day, so any number of different IT and network projects could be targeted at that worker. The information content of a job is also a very important metric. Where the information content is high, there’s a decent shot at empowerment with an IT/network project.
What kind of project? Let’s look at network versus IT/software. When is a network project likely to empower some workers? When there are workers who have to be connected to the information content of their jobs. When is that most likely to be the case? When the “knowledge workers” are distributed widely. An engineering company with a single location may offer very little opportunity for empowering network projects, while the same kind of company, with the same number of workers but distributed to 20 sites, could offer a lot of opportunity.
All of this is logical. If you ask enterprises, nearly all of them (over 95%) will agree with all these points. But if you ask how many projects are planned, justified, presented the the CFO, approved, and launched to completion, you find it happens in only 28% of cases. In fact, just a bit less than half of enterprises tell me that they even use this approach to create and approve project business cases. Almost none target projects at the places where demographics says they’re most likely to be valuable.
CIOs tell me that most IT and network projects are launched by a request from a line department. This, the CIOs say, tends to focus making a business case on the specific “making this business case” point. In other words, we launch projects in search of a justification, and the CIO is essentially charged with making the proposed projects work. We don’t look at where justifications would be the easiest to come by. This may be one big reason why the number of tech projects justified by productivity improvements has fallen steadily over the last twenty years, to the point where today they contribute an average of about ten percent to IT spending, when it was as high as just over 50% in the past.
This, it turns out, is a problem as old as computing. Back when I started being a programmer, most enterprises had an IT policy that was driven by the IT organization itself. The company I worked for had an “Electronics Research Division” that reviewed line operations overall and targeted things where “automation” could make a big difference. This approach was almost universally successful, and created a significant gain in productivity, but also created a lot of resentment in the line departments. Within five years, the company had created a parallel organization focusing on what line departments wanted, and dedicated to building what they believed were the best applications. They consigned the original one to building what would today be called “middleware” or tools. Eventually that original organization disappeared, and with it the most effective targeting of empowerment applications.
The problem here was simple; line organizations are ultimately responsible for their own people and results, and ceding this to IT created resentment, pushback, and eventually a revolution. The same thing would happen today, unless somehow the process of targeting could be inclusive of both IT and line organizations. According to enterprises, the biggest barrier to this is loss of jobs. It’s not so much that companies object to layoffs (we have plenty of current evidence that they’re willing to lay off workers to save money) but that management’s power is usually measured by the number of “reports”, meaning workers under them. Tell a line organization manager that they’re about to lose half their employees, and they see it as losing half their power.
My model says that there is enough annual productivity benefits that could be reaped by proper targeting of applications to almost double total IT spending, while meeting what enterprises say is the typical IRR/ROI relationships required by the CFO. To make that happen, the goal would be to focus empowerment on the specific industry groups where the combination of unit value of labor, information content of the job, and distribution of human assets would be the highest. This is a kind of “pull” approach to empowerment, very different from the way that projects are launched (and approved) today.
One of my goals for Andover Intel is to determine just where that is, at least identifying the industries with the most potential and noting what specific issues would have to be addressed/optimized for the best result with the least effort. That means picking the top ten to twenty industries, where the business case is strongest and where a total of at least half the benefits could be realized, and focusing on them. That’s something I’ve already started, and when I’ve completed the work I’ll report here.
