You shouldn't wince when you log in

Watch a kid sign up for a new app sometime. Tap, tap, in. The app is on. They’re using it. No tutorial, no orientation, no document called “Getting Started With Spotify.”

Now watch a seasoned professional, somebody who has done their job for twenty years, open the software that’s supposed to help them do that job. They brace. The kind of muscular preparation you do before you lift something heavy.

That’s the gap I think about.

I’ve spent the last decade building software for regulated industries. Healthcare, telecom, now construction. The pattern is the same everywhere. People have made peace with the idea that consumer software is going to be a pleasure and work software is going to be a chore. Instagram is intuitive. The system that processes your expense report is a not.

How we got here

The first reason work software is bad is that the person who buys it isn’t the person who uses it. A VP picks the vendor based on a slide deck full of feature checkboxes. The administrative assistant who has to click those checkboxes forty times a day was never in the room. So nobody designed the product for her. The buyer optimized for “can this handle our edge case” and “what’s the discount on five-year terms.” Nobody optimized for whether it’s pleasant.

The second reason is switching cost. Once a company has trained its team on a system and migrated years of data into it, leaving is a multi-quarter project. Vendors know this. A product can be miserable and keep its customers, because the alternative is months of pain. There’s not much pressure to fix the misery.

The third reason is the distance between the writer and the reader. The engineers building this software, often, have never used it. It’s built in offices for people in warehouses, basements, hospitals, classrooms, parking lots. You end up with menus designed by someone who has never had gloves on.

What’s changing

The people who grew up with consumer software are now the people picking work software. Their baseline expectation, formed by Stripe and Notion and Linear and Figma and Spotify, is much higher than the previous generation’s. When they log into their company’s safety reporting system and it asks them to navigate a 1997 portal to file a near-miss, they don’t think “well, this is what work software is like.” They think “this is bad.”

Founder-led B2B is the other shift. The product person at a SaaS startup is often the founder. The founder is often using the product daily. They have skin in the game on the field UX in a way that an enterprise PM never does. The feedback loop is tighter, the willingness to throw out a flow that isn’t working is higher, and the buyer is closer to the user.

The bar is rising. Not all at once, not in every category, but it’s rising.

What we obsess about

I’ll tell you what this means at Nexus.

Smart import. A new shop comes to us with a binder of scanned worker certifications. The old way: somebody at the office spends three days hand-typing names, certificate numbers, issue dates, and expiration dates into a spreadsheet, then re-typing them into the system, hoping nothing got transposed. The new way: drag the binder in. Nexus reads each document, extracts the fields, and presents a confirmation pass for a human to approve. Days become minutes. That’s not magic. It’s the obvious application of a technology that’s been around for a couple of years now. The reason this hasn’t been built before in our category isn’t technical. It’s that nobody decided typing was unacceptable.

High contrast for bright sunlight. A lot of the people using Nexus are standing in a parking lot at 11 a.m. in July, looking at a phone through safety glasses. A design that’s elegant in your dim office is unreadable in their reality. So we use heavy weights for labels, real contrast between text and background, no grey-on-grey. WCAG AAA where we can, AA across the board. This isn’t a fancy feature. It’s making the screen visible to the person looking at it.

Native translations. A Spanish-speaking worker opens Nexus and the entire interface is in Spanish. Not browser-level Google Translate that mangles a job hazard analysis into nonsense. Not an afterthought language pack that covers the nav bar and quits. The locale files are first-class artifacts, written and reviewed alongside the English ones, covering the JHA hazard rows and the time-punch screens and the error states. We have Spanish and Russian today, and will add more as the trade asks for them.

Field-first design. Our tap targets are all large enough that someone can hit them on an iPad while wearing gloves. Primary actions stick to the bottom of the screen on mobile because you shouldn’t have to scroll to submit. Forms render as a single column on phones because thumbs are not mice. Where dictation makes sense we wire it up, because a foreman shouldn’t have to type a paragraph with one ungloved finger. None of this is innovative. All of it is just paying attention.

The next wave

Most people building work software don’t hate users. They’re working inside incentive structures that don’t reward field UX. Long sales cycles, enterprise procurement, customer-success teams that paper over the pain so it never reaches the design review. The system makes bad outcomes likely.

What I do think is that a new generation of B2B companies is going to keep raising the bar until the old structures break. Linear did it for engineering tools. Stripe did it for payments developer experience. Figma did it for design. The same thing is going to happen in construction, in healthcare, in field services, in the trades. Not because anybody decided to, but because the people who write the checks now expect what they expect everywhere else.

We’re trying to be part of that. Not by being clever. Just by paying attention to the person on the other side of the screen and refusing to ship software that makes them brace.

You shouldn’t wince when you log in. We think that’s a reasonable thing to expect at work, the same way you expect it at home.