The question tech parents ask schools, and owners should ask vendors
What is this tech supposed to change, and what's the plan if it doesn't?
That's what a parent in San Mateo says families should ask their school district, in a The Guardian piece this morning on Silicon Valley families keeping tech out of classrooms. It's the best vendor question I've read this year.
I've built software since 2010 and my son's kindergarten has no screens. The parents in that article aren't afraid of computers. They know most products are built and sold against usage: logins, sessions, minutes in the app. A school wants children who read better. A reading app can be great at the twenty minutes and do nothing for the reading.
Owners buy the same gap. You want fewer "where's my order" calls; the portal reports logins. You want fewer empty slots; the renewal deck counts bookings your front desk would have taken anyway.
The test has two halves. Name the number that should move before you sign. Then decide, on the same day, when you'll check it and what you'll do if it's flat.
I wrote up the rest, including what to ask before pricing "hours saved."
The Guardian ran a piece this morning on Silicon Valley parents who pay to keep the technology they build out of their children's classrooms. The engineers and AI workers in it are easy to read as hypocrites, or as a lifestyle trend. I read them differently. They know how these products get made, and they are applying a buying test most organizations skip.
One parent in San Mateo put the test better than most procurement checklists I have seen. Before a district signs, she said, ask "what specific outcomes they're looking for that this tech is supposed to help with, and knowing what the plan is if they aren't seeing those outcomes." Then: "A lot of these companies have really sexy pitches, but then on the ground, the work doesn't quite bear out."
I have built software for a living since 2010. My son is in kindergarten at a small Waldorf-inspired school in San Francisco, where the children work with their hands and there are no screens in the classroom, and I sit on its board. So I read the article twice, once as a parent and once as someone who reviews technology decisions for small business owners. The second reading is the useful one here, because nearly everything those parents are asking a school district is something an owner should be asking the person selling them software.
What the product is built to move
The parents in the article are not afraid of computers. They know what a product team gets measured on. Most software is built and sold against usage: daily active users, time in the app, seats, sessions. That is a reasonable thing for a vendor to optimise, and it is rarely the thing the buyer wanted.
A school wants children who read better. A reading app can be very good at getting children to spend twenty minutes a day in it, and the two are not the same result. One of the districts in the article is moving away from an adaptive reading and math program that has drawn criticism for repetitive lessons. Twenty minutes a day was never the problem the district was trying to solve.
A business has the same gap. You buy a customer portal because you want fewer calls asking where an order is. The vendor's dashboard will show you logins. You buy scheduling software because you want fewer empty slots on the calendar. The renewal deck will show you bookings made through the app, including the ones your front desk would have taken by phone anyway. The vendor is not lying. They are reporting what their product moves, and that is your cue to define what you need it to move before you sign.
Name the outcome, then the exit
That parent's test has two halves, and the second half is the one nobody writes down.
The first half: say what should change, in a number you already track or can start tracking this month. Calls per week about order status. No-show rate. Days from job completion to invoice. If nobody can name the number, you will have little to go on when the renewal arrives.
The second half: decide now what happens if the number does not move. By when do you look, who looks, and what do you do if it is flat? A $30,000 system at a $4M business is not a disaster if it does not work. It becomes one in year three, when the staff have built their week around it, the data lives inside it and switching costs more than the original purchase. The exit plan is cheapest to write on the day you sign, which is the day nobody wants to think about it.
When budgets get tight
The line in the article I keep coming back to is from Jon Valant at Brookings: "It's when budgets get tight that the temptation to replace people with technology is especially hard to resist." Staff are the biggest cost in a school, so the screen arrives where the adult used to be.
Small business owners are hearing a version of that pitch right now, usually priced in hours saved. Some of it is a good deal. Plenty of work in a small business is clerical, repetitive and better done by software, and nobody misses retyping an order from an email into another system.
The question is what the person was actually doing in those hours. A researcher in the article makes the point about children: they are more likely to put in effort for a teacher they love than for a computer. Customers are not children, but the mechanism is similar. In a lot of service businesses, the person who answers the phone or walks the customer through an estimate is the reason the customer comes back. When an owner swaps that person for a tool and lists the salary as the saving, they have counted the cost and left out the revenue that person was holding.
So before you price the hours saved, ask which of those hours a customer would notice. Automate the ones they would not.
Borrowing the insider's advantage
The researchers in the article worry that less-resourced districts will be left with the tools wealthier ones are abandoning, partly because the wealthier ones have parents who know the industry. One district's AI task force had volunteers from large tech companies on it, and a trustee said they could "see a longer runway of where these products are headed."
Most small businesses do not have that person on staff. What they need is narrower than a technologist: someone who understands how the product makes money and has no stake in whether you buy it. Ask the vendor three things that person would ask. Which of your customers stopped using this, and why? What does your product report on, and how does that map to the number I care about? If I leave in two years, what do I take with me, in what format?
The answers will not tell you whether the software is good. They will tell you whether anyone, including you, will be able to say so a year from now