Management vs. Operational Productivity: A Costly Mix-Up

In 1992, Steve Jobs stood in front of a room full of MIT business students and admitted that the first decade of his career had been pointed in the wrong direction. Not the wrong company, not the wrong product — the wrong kind of productivity. It's a confession that has aged, in the worst possible way, into a prophecy.

That's the frame Dr. Errol Brandt builds from in a recent video that's worth taking seriously, even with a grain of salt kept handy. Brandt is the founder of something called the Kira engine, and he makes no secret of that — the product pitch arrives on schedule in the back half of the video. But the diagnosis he delivers before the sales portion earns a proper hearing. The argument draws on a framework that Jobs himself cited, and the internal logic holds up better than most tech-sector productivity discourse tends to.

The Jobs speech Brandt references is real and publicly available. In it, Jobs recounts reading economist Paul Strassmann — then serving as Chief Information Officer at the Pentagon — who had studied the IT spending patterns of a wide range of companies and found a consistent split. The underperformers concentrated their technology budgets on management productivity: email systems, document tools, presentations, reports. The outperformers directed theirs at operational productivity — the tools and systems that made the actual work faster, cheaper, or more reliable.

Here's the thing about that finding that should make any skeptical reader pause: it goes against the grain of how almost every large organization actually behaves. Most IT budgets, in practice, flow upward — toward systems that serve the people who approve those budgets. The fact that the data apparently ran the other direction is the interesting part. Jobs, to his credit, read Strassmann and concluded that the personal computer had mostly failed on operational terms. PCs turbocharged management. They didn't move the factory floor.

Brandt's argument is that we've spent thirty years proving Jobs right, and not in a flattering way.

The mechanism Brandt describes is plausible and recognizable. In 1996, Kaplan and Norton published The Balanced Scorecard, a management framework that encouraged leaders to track performance across the full breadth of an operation rather than fixating solely on sales and profit figures. Reasonable idea. The trouble is what happened to it in the wild. Brandt describes watching the concept metastasize into a "dashboard arms race" — every department demanding its own metrics, every executive wanting visibility, every IT team hired to build the visualization layer that sits on top of the actual work without touching it.

He tells a story about a dashboard project he watched up close. A CEO wanted comprehensive visibility across the organization, so the most experienced IT manager available was dispatched to gather requirements. What followed was organizational physics: every business unit wanted its number on the dashboard, nobody wanted to be left off, and the initial list of ten key performance indicators swelled to 127. "No human can act on 127 metrics," Brandt says flatly. "You can barely even read them." Months of effort produced, in his words, "a beautiful 127-metric artifact that changed absolutely nothing about how the business actually ran."

That, he argues, is management productivity in its most seductive and destructive form: "enormous activity, beautiful outputs, zero productivity. In fact, negative productivity."

The negative productivity framing is Strassmann's, not Brandt's invention, and it's the sharpest concept in this whole lineage of thought. The idea is that reporting on problems is not the same as solving them, and that the resources consumed by reporting can actively crowd out the resources needed to fix things. You end up, in Brandt's formulation, with "teams of developers who spend their day making pretty dashboards for everybody, and there's simply not enough resources left to help people on the ground fix the problems."

He sharpens the point with a real conversation. Brandt describes talking to a COO at a potential customer recently — the Kira sales context is explicit — who laid out the allocation clearly: dedicated analysts serving sales and marketing, almost no analytical resources dedicated to production and inventory. The COO's view was that a small improvement in production yield or the prevention of a single scrapped batch would have a larger financial impact than meaningful optimization of the marketing budget. The company's IT resources were aimed at the easier-to-measure problem, not the bigger one.

"Most of the companies I've ever been involved with have most of their precious IT resources focused on management productivity," Brandt says, "and they spend [nothing] on operational productivity."

The AI application is where the argument gets most pointed, and most timely. Brandt's read on why Microsoft Copilot has largely disappointed is that it was aimed squarely at management productivity from the start — helping people write emails faster, summarize meetings, generate presentation drafts. That's not nothing, but it's Strassmann's losing category. "This thing was doomed from the start," Brandt says, "because it never got close to the kind of problems that actually need to be solved first."

That framing puts a useful lens on the broader AI-in-enterprise conversation. The tools getting the most marketing attention — the ones being deployed first and most visibly — are mostly language models embedded in the applications that managers already use. The productivity gains being claimed are real but modest: faster drafting, quicker summarization, easier search. What's happening on the warehouse floor, the production line, or the inventory system is a different and slower conversation.

Brandt's pitch for Kira is built around the idea that analytics should live where the work happens — "close enough to change the outcome and not just simply describe it." Whether Kira delivers on that is not something this article can assess, and Brandt's conflict of interest in making the case is worth naming plainly. But the distinction he's drawing doesn't depend on whether his product is good. The distinction was there in Strassmann's research before Jobs cited it, and it's still there now.

The harder question the video doesn't fully answer is why the pattern is so durable. Brandt gestures at it — everybody wants to keep the CEO and CFO happy, operational problems are harder to measure, generic software is easier to deploy than custom solutions. Those are real forces. But there's also something structural worth noting: the people who control IT budgets are, almost by definition, managers. The tools that serve managers are the tools that get funded. The forklift driver doesn't sit in the budget meeting.

Jobs put it plainly in that 1992 speech, and Brandt is right that the observation has barely aged: the companies that win are the ones that spend on making the work better, not on making it easier to discuss.

Whether that lesson lands differently now that AI is available to deploy on either side of the ledger — toward the dashboard or toward the floor — is probably the most consequential technology allocation question of the next five years.

— Jin Seo