Monday, before anyone praises the UI
A supervisor comes in, unlocks a phone that already has twenty apps, and has a list in their head: which loads went, which did not, who is off, what the owner will ask at eleven. The official system is a URL they last opened at the training. The password is in an email from March. The screen wants a cost centre before it will show today’s dispatches. They open WhatsApp. The group already has photos and a yes or a no. That is the competitor. Not another vendor. The group.
We have sat in rooms where the new system was shown on a projector the Friday before. Everyone nodded. Monday, the projector was off and the group was busy. Nobody was making a point. They were trying to finish the morning. If your tool cannot survive that morning, it is not an adoption problem. It is a product that does not do the job that morning contains.
This is not an argument against building software. It is an argument against building theatre. A field story we keep seeing: a handsome operations screen, role names from a workshop, charts that look like a SaaS advert — and a printed list still on the desk because printed lists do not expire a session. The printout is the system of record. The software is a visit.
Ask who opened it today without you in the room. If the answer is “they will, once we add reports”, you do not have a tool yet. You have a promise. Monday does not run on promises. It runs on a list, a login that works, and a way to undo the mistake someone will make before tea.
Login is the first feature, not a preface
If staff share a password on a sticky note, you do not have security and you do not have usage data. You have a rumour. If the official login needs a VPN that drops on the warehouse Wi-Fi, you have a museum exhibit. If the OTP goes to a manager who is in traffic, you have trained people to avoid the front door.
Monday-morning login looks like this: the person who does the job can get in on the device they already carry, in under a minute, without calling IT. Roles are few. The warehouse role sees today’s list, not the chart of accounts. The accounts role is not asked to confirm a dispatch. Forgotten passwords are a path, not a ticket that waits until Wednesday.
We still see tools that open with a blank dashboard and twelve tiles because the design system had twelve tiles. That is a Monday tax. The first screen after login should be the job, or a short list of today’s jobs. If you need a home page, make it today’s list. Save the tiles for a product that staff already trust.
Single sign-on is fine when the company already lives in it. Forcing a new identity provider in the same month you ask people to leave WhatsApp is two projects. Do one. A working password reset on the phone they have is more “enterprise” than a SAML slide nobody completed.
One job, finished, without a tour
The job is a sentence. Create today’s dispatch against an order. Record the return. Mark the machine down. Log the visitor. If the sentence needs a tour, the screen is wrong. If the sentence needs three departments to click in the right order before anything saves, the process is still a meeting. Software cannot host a meeting that the building has not agreed.
One job means the happy path is short and the exception is visible. A missing item code should not dump the user into a master-data module they have never seen. It should say what to do: pick from a short list, hold, or call the person who owns the list. Exceptions that fall on the floor become side channels. Side channels become the real system by Tuesday.
We watched a team ignore a beautiful job card because completing it required a customer PO number that customers in that trade do not send until later. The official process was a fiction. WhatsApp held the real sequence: start the job, get the number, patch it. The tool that wins Monday is the one that allows that sequence without pretending the fiction won.
If you need a name for the smallest version that tests “they will leave the group”, that conversation lives on what MVP development is. This article is the Monday after that page: the slice is already supposed to exist. Does anyone open it? If not, do not add a module. Fix the minute that failed.
Undo is how you earn a second week
People will post the wrong quantity. They will pick the wrong customer. They will confirm a dispatch and then the vehicle will not come. If the tool treats those as moral failures — no edit, no reverse, find a superuser — staff will stop touching it. The workbook lets you change a cell. The group lets you send a second message. Your system has to be at least that kind.
Undo is not “delete from the database”. It is a reverse that leaves a trail. Who changed it, when, why. Accounts can live with a correction they can see. They cannot live with a silent rewrite. Supervisors can live with a voided line. They cannot live with a locked screen and a queue in someone else’s inbox until Thursday.
The first version can be crude. A “wrong — undo” action that asks for a reason and puts the line back on today’s list is enough. Polish later. What you cannot skip is the feeling that a mistake is recoverable before the owner walks in. Fear is why pretty systems die. Not taste.
Write undo into the first list, next to login. Teams that cut it to hit a date ship a prototype and call it live. Staff discover the missing reverse on a noisy morning and never come back. That is not resistance to change. That is a reasonable reading of risk.
Today’s list is the product
Not last month’s utilisation. Not a map of India with dots. Today’s list: what is open, what is late, what is waiting on someone. Sort it the way the supervisor already sorts the morning, even if that way offends a data model. If they work by vehicle, show vehicles. If they work by customer shouting, show the shouting in a column you can filter.
A list that needs a date range and three dropdowns before it shows Tuesday is a report. Reports are for Friday afternoon. Monday wants defaults: today, this location, this team. If the default is wrong, they will not reconfigure it. They will leave.
Print and share still matter. A lot of Kerala shop floors still need a paper in a pocket or a photo in a group for the driver who will not log in. If the tool cannot produce a simple list they can send, the group stays the distribution layer. That can be fine for a phase. Admit it. Then make the official list easy to send so the photo is of your list, not of a notebook.
When today’s list is trusted, you have earned a report. Not before. Dashboards built first are how you get a wall-mounted TV that nobody looks at except visitors. Visitors do not run the morning.
Why pretty unused systems die
| What shipped | What Monday needed | What staff did |
|---|---|---|
| A dashboard of KPIs | Today’s open work | Opened the group |
| Twelve modules | One job that saves | Used the module that prints, then left |
| A new login every week | A session that lasts the shift | Shared a password, then avoided it |
| No reverse on a confirm | Undo before eleven | Stopped confirming |
| Masters only IT can edit | A hold when the code is missing | Kept a side workbook |
Pretty is not the enemy. Pretty without those rows is a brochure that happens to need hosting. Staff are not design critics. They are people with a list. If your screens lose to a chat thread, the chat thread was clearer about the job.
The side channel is a measurement
WhatsApp, the notebook, the “just for this week” sheet — those are not moral failings. They are where the official path is too slow, too strict, or too broken. Count them. If after two weeks the group is still the dispatch list, the tool failed the only test that matters. Adding a dark theme will not fix it. Adding a customer portal will not fix it. Adding a mobile app on top of a login nobody completes will not fix it.
Sit on the floor on a Monday you did not announce. Watch the first hour. Photograph the unofficial list. That photograph is the spec for the next change. It is also how you avoid a second project that encodes the workshop instead of the morning.
Owners sometimes want a ban: “from June, only the system”. Bans without a working Monday produce secret sheets and worse numbers. Make the official path the fastest path for the one job. Then the ban is almost unnecessary. People are busy. They will use the short road.
If the short road needs a process change — a customer code before work starts, a stock check that used to be a guess — write that as a management sentence, not as a validation error in red. Software can enforce. It cannot persuade a trade that has billed without codes for twenty years. Do the persuasion in the room. Then enforce.
What we build when we mean Monday
We build on ASP.NET Core because the long cost is change: a new exception, a new role, a reverse you did not imagine in week one. The stack is not the story staff tell. They tell whether they got in, whether the list was right, whether a mistake was recoverable. Those are the acceptance tests. Colour is later.
The first slice is allowed to look unimpressive. It is not allowed to be unmeasured. “They said it looks modern” is a Friday demo. “Today’s list was completed in the system four mornings out of five” is a Monday result. Only the second funds the next screen.
Handover still includes backups, a named person who can reset a login, and a week where we sit with the first shift. Cutting those to look lean produces a pretty unused system with nobody to call. That is how the group chat wins again, quietly, without a meeting to record the loss.
If you already have the pretty unused system, do not start a rewrite on pride. Watch a Monday. Cut the tiles. Fix login. Put today’s list first. Add undo. Retire one side channel on purpose. That sequence has saved more budgets than a new design language. If after that nobody opens it, the job was wrong, or the incentive was. Those are management facts. Software cannot paper them. We can help you see them early. We cannot pretend a dashboard is a morning.
Bring us the unofficial list and the login complaint, not a mood board. We will tell you whether the next spend is a thin tool on .NET or a conversation you have not had with the people who actually start the week. The MVP page is there if you need the planning words. This page is for the Monday those words were supposed to survive.