When I’m working with teachers, one recurrent theme is their frustration with how school districts spend money. There’s a familiar lament that districts buy technology and resources that teachers don’t really want, while failing to procure stuff or services that teachers think would be more helpful. In fact, many teachers suspect that the administrators who decide what gets bought don’t know or particularly care what they think. I fear there’s something to that.

For instance, while school districts annually spend more than $8 billion on education technology, tech companies estimate that about two-thirds of all educational software licenses go unused each year. Meanwhile, teachers wind up home-brewing the resources they actually use, drawing on friends and old files. Indeed, a 2016 study from the RAND Corporation found that 98 percent of elementary teachers rely on materials they’ve developed or selected themselves when teaching English/language arts, with Google, Pinterest, and Teachers Pay Teachers as the three most popular sources.

Why the disconnect? In an intriguing new analysis for my AEI shop, Thomas Arnett, formerly a middle school math teacher and now a senior research fellow at the Clayton B. Christensen Institute, tackles all this through the lens of the “Jobs to be Done” business theory. Now, Jobs to Be Done sounds like one of those B.S., jargon-laden business school concepts, but it’s actually pretty straightforward. It posits that people don’t actually buy products; they “hire” a product to do a specific job for them. For instance, Harvard Business School pro Clayton Christensen, who dreamed up the concept, conducted an early Jobs to Be Done study to explore why so many people purchase milkshakes in the morning—rather than coffee, donuts, or granola bars.

What did Christensen find? As he’s explained, “They faced a long, boring commute and needed something to keep that extra hand busy and to make the commute more interesting. They weren’t yet hungry, but knew that they’d be hungry by 10 am; they wanted to consume something now that would stave off hunger until noon.” The point was that people weren’t necessarily saying, “Ah-ha, I want to buy a milkshake,” but rather, “I want something that will be filling, fun, and long-lasting.” Once you see the problem that way, it’s a lot easier to figure out what people might find appealing.

Arnett notes that this kind of approach, which is today almost wholly absent in school systems, can illuminate why so many things districts buy go unused—and why teachers wind up scrambling to assemble the resources that they actually use. From textbooks to technology, district leaders too rarely understand what jobs teachers really need to be done, which is why so many big-dollar purchases get left on shelves and servers collecting dust, while teachers flock to Pinterest.

For instance, Arnett notes that districts often embrace textbooks or instructional programs because leaders are convinced “that they work.” When these new resources are then pushed down to schools, accompanied by enthusiastic training that explains all the benefits for students, teachers are often left cold. They’re not sure that a new approach will actually help them solve the problems they see in their classroom. They wonder if major instructional changes might do more to create headaches than to help them improve. They worry that some of their most successful activities or units will be wiped out and that they’ll instead be struggling to figure out some new tool or instructional technique.

In other words, teachers often feel like district-level procurement and programming is doing more to create new jobs than to help teachers accomplish the ones they’re struggling to get done.

There are things that vendors and school districts can do to change this dynamic. For one, they should spend a lot more time watching teachers actually wrestle with challenges and talking to teachers, at length, about how they go about their work. This means spending less time on pilot data, focus groups, and surveys, and lots more time watching and listening to teachers.

Similarly, school and system leaders should spend a lot more time understanding the jobs teachers need help getting done. Instead of spending so much time trying to instruct teachers in how to use things they’re reluctant to embrace, they should spend a lot more asking what teachers actually need—and learning why they keep rejecting expensive resources that seem so compelling to the harried administrators who purchase them.

At the end of the day, what matters for kids is what happens in classrooms, with teachers, behind closed doors. Buying impressive-seeming tools or highly touted programs won’t make a lick of difference if teachers aren’t using them. That’s a recipe for wasted funds, teacher alienation, and general frustration. Rather than telling teachers over and over what administrators, researchers, advocates, or vendors think they should be using, Arnett has sketched a mind-blowing alternative: that we make it a point to ask teachers what will help them do their job. It’s a crazy notion, I know, but seems like it’s worth a shot.