The client book
Table. Name, last visit, what they asked for.
Then a job. A listener emails them after a booking. A weekly cron finds anyone quiet for thirty days.
By Need
RecordsIf you think in columns more than screens, begin there. Set up the table for the business, live in the rows, then attach a small listener or cron when something should happen.
That is not a lesser way to start. It is often the honest one. You already know the nouns. You already know which columns matter this week. A screen is an opinion about how someone should work. A table is the work as you already keep it.
Records is the App Marché path for that posture. You set up the table first. You run the business from it. Then the table starts doing work — a listener when a row changes, a cron when a day passes, an email when someone did something. The person evolves. The app evolves with them.
Why this is easier
Designing an interface asks you to invent the whole app before the data is even trustworthy. Columns do not. You can be useful on day one: add a row, find a row, change a field. That is already an app someone can run a business on.
The later screens, if you need them, sit on the same records. You do not start over when you want a nicer page. You do not switch products when you want one email. The table stays. The jobs grow around it.

You do not graduate out of the table. You attach the next small capability to it. If the job is still “email them if this happened,” it belongs here. If the jobs have become the product, that is a different page.
01
Make the table
Name the thing and give it the columns you already keep. Clients. Jobs. Stock. If you would write it down this week, it belongs here.
02
Live in the rows
The table is the first interface. Open it in the morning. Find a row. Change a field. You do not need a designed screen to start running the business.
03
Attach one small job
When a row should do something, write a small listener or a cron. Email them if the status changed. Remind yourself if nobody moved. Keep it small on purpose.
Same records. One small reaction.
Table. Name, last visit, what they asked for.
Then a job. A listener emails them after a booking. A weekly cron finds anyone quiet for thirty days.
Table. Who asked, what is due, what status it is in.
Then a job. A listener pings the next person when a row moves. You stop forwarding the same update by hand.
Table. What you have, what is promised, when it is late.
Then a job. A nightly cron marks the late rows and sends one note. The count and the follow-up stay on the same table.
A listener is “when this happens, do that.” Someone books, someone pays, someone marks a row done — then the app emails them, assigns the next person, or writes the follow-up you used to remember yourself.
A cron is “when this time comes, look at the table.” Every morning, find the quiet clients. Every night, mark what is late. You are not building a pipeline. You are giving the table a clock and a reaction.
That is the point of starting here. The table is easy because it matches how you already think. The jobs stay easy because they hang off rows you already trust. Screens can wait until someone besides you needs one.
Records is the wrong first page if you already know you want a public booking flow, a customer-facing product, or a heavy approval chain. Those have their own starting shapes. Come back here when the honest first move is still a table.