On 1 September 2026, James Dyson launched a toothbrush (opens in a new tab) at an event in Paris. The Dyson CameraJet carries a 100k-pixel macro camera in its head, which uses machine learning to find the gaps between the teeth as the user brushes and fires a conical jet of mouth rinse (opens in a new tab) into each one, fed by a small tank built to hold its pressure at any angle. Oh, and the user can watch the inside of their mouth live in the MyDyson app.
A feat of engineering that took six years, 661 engineers, 38 patents and supported by a proxy plaque model built by Dyson with the National University of Singapore's Faculty of Dentistry to tune the jet to remove up to 72% more plaque than leading electric toothbrushes, according to Dyson. The CameraJet sells for 500 dollars in the US, and sells as a complete dental system with its own non-foaming toothpaste and its own mouth rinse.
The CameraJet has been on sale for a few months now, and its usefulness will be settled over the next year by something its engineering cannot answer, which is whether people keep reaching for it every morning and keep buying the rinse. Note that the British Dental Association answered that people do not need an expensive toothbrush and should brush twice a day as their dentist advises.
Most hardtech founders are engineers and scientists, and the problems that attract them are the hard ones, like a sensor that detects what others miss. Those problems also produce the demos that win grants, press coverage and first meetings with investors, which is why a hardtech startup can spend years making its technology remarkable before anyone checks whether a specific person needs it often enough to pay for it.
At Garage&co, usefulness is one of the first conversations we have with founders, because a hardware product can pass every test at the bench and still fail to find a place in someone's day. And in hardtech, that lesson usually arrives after the tooling, so before any production at-scale begins.
So what makes a hardware product useful, rather than impressive?
A person, a job, a place, a price
A useful product does a job that a specific person needs done, in the setting where that person works or lives, at a price they accept, often enough that it earns its place.
The person is someone a founder can name and meet : could be a line operator on a night shift or a physiotherapist between two patients. The job is a task they already perform today, slowly or badly or at a cost they feel. The setting includes the dust, the gloves and the noise. And finally, the price includes everything the person pays over the life of the product.
Repetition is what turns a purchase into a habit and it is where most hardware could see the revenue and the referrals come from.
Going back to Dyson, its first product started from a job that was easy to see. In a 2007 interview with TIME, James Dyson recalled assuming that a vacuum's suction came from its motor, then finding that the first dust to enter the bag blocks its pores, so the machine loses suction as the bag fills. Dyson dates the idea to 1978, when the company built an industrial cyclone tower to clear a filter that kept clogging with powder in his Ballbarrow factory and wondered whether the same principle would work in a vacuum cleaner. For James, households with a bagged vacuum already knew the problem, so the product had a person, a job, a setting and a measure of success (suction that holds as the bin fills) before the first prototype existed.
You could also argue that the CameraJet also starts from a real job, since James Dyson said that the company's research shows people consistently miss the gaps between teeth where plaque forms, and he called flossing an awkward and time-consuming chore. What remains open is whether the people who skip flossing will adopt a 500-dollar device, a special toothpaste and a refillable tank to do it.
Two audiences for one product
An impressive product is judged by people who see it once, such as a jury, a journalist, an investor during a demo or a visitor at a trade show. A good hardware demo can win them over in five minutes.
A useful product is judged by the person who uses it on an ordinary weekday, for the hundredth time, after the novelty has worn off, and that person judges whether it saved time, removed a nuisance or paid for itself.
Grants and early capital often follow the remarkable demo, and the risk for a hardtech startup lies in letting the first audience's applause stand in for the second audience's verdict. Product-market fit for hardware shows up in the behaviour of that second audience, in reorders, in units still switched on after three months and in customers who complain when a device is taken away, and none of it can be measured on launch day.
Usefulness tends to show up in the numbers of the second month: how many units are still in use, and how many refills were ordered.
Going back to Dyson, one example of the importance of being useful comes with the Dyson Zone, a set of noise-cancelling headphones with an air purifier and a visor that sends filtered air to the nose and mouth. Unveiled in 2022, at a 949 USD price tag. You can already guess that it had a very muted commercial reception.
The costs that come after the purchase
Usefulness carries costs that exist long after the purchase, especially the recurring cost of maintenance and support utilities.
As previously mentioned, the CameraJet needs a non-foaming toothpaste, a clear rinse for its 12.5 ml tank and brush heads carrying RFID tags that signal when to replace them. It adds up to the maintenance and the cost of charging the device. And the necessity to keep maker’s supplies, which the user has to trust that those supplies will stay available and fairly priced.
Apart from hard costs, the adoption of a hardware device needs also a behaviour change. It is usually the heaviest of these costs, because every new step in a routine competes with a habit the person has kept for years.
For a founder, a consumable that people reorder improves the unit economics of a hardware business, because each refill carries margin long after the device has been sold, and the same consumable becomes a reason to stop using the product when the job it serves feels optional. The recurring revenue a hardtech startup writes into its financial model exists only if the use repeats, which is why frequency of use belongs among the first hypotheses a team tests.
Testing usefulness early
Usefulness can be tested long before the product is finished, and most of the methods rest on a principle of customer discovery, which is to watch what people do, since what they say about a product that does not exist yet predicts little about what they will do with it.
Watch the job being done today. Spend time where the task happens and note how long it takes and what people do when it goes wrong, before showing them anything.
Look for people who already built a workaround. A spreadsheet, a modified tool or a jig cut in the shop shows that the problem costs someone enough to make them act.
Measure repeat use. Leave a prototype with a user for several weeks and compare use in the fourth week with the first.
Ask for money or for commitment. A deposit, a paid pilot or a letter of intent with a price gives a signal that compliments cannot.
Find out what they would stop using. A new device enters a space already occupied by tools, habits and budgets, and a product that replaces nothing adds a cost and a step to the person's day.
Consider all the people who used your innovation
We’ve talked a lot about consumer products, but, in B2B hardtech, usefulness is judged by several people who rarely sit in the same meeting : the buyer, the operator and all other people involved.
The buyer signs the purchase order and weighs the price, the expected return and the risk. The operator uses the device during a shift and notices whether it slows the work or adds a step to a procedure that already has too many. Then, we need to account for the person who maintains it, often a technician, since he cares about how often it fails, how long a repair takes and whether spare parts exist.
A device that pleases the buyer and frustrates the operator can end up unplugged in a corner of the plant long before the contract comes up for renewal.
Founders tend to spend most of their time with the buyer, who controls the budget, and a pilot designed around the buyer's criteria alone can end with a positive report and no second order, while a pilot that also tracks how often operators use the device and logs every service call comes much closer to predicting the renewal.
What this means for founders
Four things you need to know to consider usefulness of your innovation.
Write a usefulness statement before the next engineering milestone in one sentence. It needs to name the person, the task, what they do today and what it costs them in time, money or risk. Then, you can test every planned feature against this statement.
Start your customer discovery with the users who have built a workaround. They can tell you quickly whether your product beats what they made themselves.
Measure repeat use in the user's real setting over several weeks. Count units still in use, refill orders and devices left in a drawer, and set the threshold you need to see before the pilot begins. Put the full price in the same test, including consumables, maintenance and the time it takes to learn the product, and ask for a payment or a commitment that reflects it.
Before you freeze the design, spend a day with the person who will maintain the device. Show them how it opens, how it is cleaned, how a part is replaced and how it reports a fault, and change the design wherever they hesitate. That person will decide how long the product stays in service.
Note: We used AI to polish the text and make it flow better.




