← All posts

Where is the job right now? A tracker that answers for you

Automation · 18 November 2026 · 8 min read

A four-person bridal alteration studio loses two days per dress to one question: where is it, and who has it. Here is the six-step method for building a job tracker behind the scenes, so the desk can answer a client without walking across the shop.

The scenario

A bridal alteration studio in Perth. Two seamstresses, the owner doing fittings, and one casual who handles the desk twice a week. They take bookings through a booking link, notes in a shared inbox, and a spreadsheet that tracks every dress from first fitting to pickup.

The delay they cannot seem to shake is not on the machines. It is between them. A silk dress has to go upstairs to the sequin room after the second fitting, and the staff member upstairs finds out by walking over and asking. A dress that has been altered is finished, then sits on a rack for two days because the person who does quality checks only looks at the rack when somebody mentions a pickup. A client rings to ask whether the wedding guest dress is ready, and the answer is always "let me check".

None of these are big. Each one is ten minutes and a delayed reply. But a dress that sits for two days after finishing is a wedding date moving closer, and the client can feel it.

This is the automation almost nobody starts with, because it is invisible. It is a single question, asked thousands of times a year: where is this dress right now, and what has to happen next? This post is about making one place that can answer it, so nobody has to walk over and ask.

Why a status board beats a chat message

The obvious fix is a group chat. Someone finishes altering, posts a photo, tags the next person. It works for a week. Then the chat fills with other things, the tags get missed, and the dress sits again.

A chat does not tell you what is outstanding. It tells you what someone said in the last hour. You want the opposite: a list of every dress in the shop, filtered to the ones that are ready to move and have not moved.

That is a single table. One row per dress. A job number, a client name, a stage, who is holding it, and when it got there. Everything else, every message, every note, is a layer on top.

If they already use the spreadsheet, they already have this. It is just that nobody updates it at the moment the dress moves, because there is no prompt. That is what automation adds.

What good looks like after a month

After a month, the owner can walk in on a Monday and open one page. It shows every dress in progress and lists the ones that have been sitting at a stage for over two days. The casual staff member on the desk does not need to ask anyone where a dress is; she looks it up and answers a client in ten seconds. The seamstress who finishes a dress records the move in one tap from a phone, and the quality check gets a prompt the next morning. Nobody has to remember the quality check.

And the owner has stopped being the switchboard. That is the real win, because every question routed through the owner is time the owner is not working.

Step one: write the stages down, in the order a dress moves

Sit with the person who does the fittings and have them write the stages on a whiteboard, in order. Do not tidy them up, do not combine them. Get the real ones.

For the studio this comes out as: booked, measured, altered, quality checked, steamed, waiting for pickup, collected. Seven stages. There are almost always between four and eight, never two. Two means someone is hiding work inside a stage and will keep walking over to ask about it.

Next to each stage, write who handles it. Notice how many stages have nobody named. Those are where dresses get stuck, and they are what the system makes visible.

The stages are the backbone. If the stages are wrong, every automation built on top of them will send the wrong message to the wrong person, and staff will stop trusting the whole thing within a fortnight.

Step two: turn the spreadsheet into a fixed list of columns

The existing spreadsheet probably has eight columns added over two years, half of them empty. Strip it back. One row per dress, and these columns only: job number, client name, phone, event date, current stage, stage since, held by, last note.

Two columns carry all the power. Current stage is where the dress is. Stage since is when it got there, and that is what turns into "this has been sitting for two days".

If the spreadsheet is the system for the first month, that is fine. Do not buy anything yet. This step is about agreeing what a record looks like, and a spreadsheet does that for free. Buying a tool before anyone has agreed the columns just gets you a worse spreadsheet.

Step three: make the move one tap, at the moment it happens

When someone moves a dress, the record should update in one action and ask two questions: which stage, and any note for the next person. A phone form with one dropdown and one note box is enough. Anything more and the seamstress will do it at the end of the day from memory, which is the behaviour you are trying to remove.

The rule to set on day one: the dress does not get put down until the record is moved. Not at lunch, not at the end of the day. It sounds fussy, and it is the single habit that decides whether this works. A tracker an hour behind is still useful. A tracker a day behind is worse than none, because people stop looking.

Step four: let it tell the next person, instead of making them ask

When a stage changes to quality checked, a message goes to the person who checks it, with the job number and client name. Not a group chat. A direct message to one person, or an entry on a checklist that opens in the morning.

When a stage has not moved for a set time, a nudge goes out. Two days at altered, one day at quality checked. Once only, per stage.

This is the part owners get wrong. They send the nudge to everyone, and nobody acts, because everyone assumes someone else has. One named person, one message, once.

The second mistake is reminding about a stage that has no time limit. A dress can sit at waiting for pickup for a fortnight and nobody should be nudged. Put a nudge time only on the stages where a delay costs something, which here is quality checked and steamed.

Step five: give the desk the answer, so nobody asks the owner

The desk needs one page it can open when a client calls. Type a name, see the dress, its stage, and the date it was promised.

If the dress is at steamed or quality checked, the page should offer the pickup message ready to send, with the time and address filled in. The desk sends it with one click. That is the customer-facing loop opening almost free, because the stage data already exists. It also means the seamstress is not surprised by a pickup, because the rack gets checked before the client arrives.

The failure modes worth knowing about

The first is a stage list everyone agreed on in a meeting and nobody uses. The fix is a card next to the machines for a fortnight, then correcting it once. The list that survives two weeks of real work is the right list.

The second is a record that gets moved twice. A dress goes to altered, then back to quality checked, because someone skipped a stage. The tracker should allow that without fuss. Step-backwards moves are normal; forbid them and staff will fudge the stage to avoid a warning.

The third is a nudge going out to a dress genuinely waiting on the bride to bring in the shoes. Every reminder needs an off switch per dress, and staff need to know it exists. An automation that cannot be paused gets switched off wholesale.

The fourth is the one that kills the project: reports. Within a month someone will ask for a weekly summary with average days per stage. Say no until staff have stopped needing the tracker to be perfect. A report showing quality check averages four days will get blamed on the staff who are using the system honestly, and they will stop using it honestly.

What the first month actually looks like

Week one is messy. Half the dresses got moved, half did not, and the list is wrong. That is expected, and it is why you run the tracker alongside the old spreadsheet rather than replacing it.

By the end of week two the list is mostly right, and the nudges start catching dresses that were genuinely stuck. Expect a few in the first fortnight. Those were already stuck; you are seeing it for the first time.

By week four the question "where is this dress" gets asked less than half as often. The desk answers without walking over, and pickup dates get kept because the delay is visible early enough to fix.

Effort is modest: an afternoon to agree the columns, an hour to build the form and the message, then one person checking the list daily for a fortnight. That is the whole project. It is not a platform decision. It is a decision about who is allowed to be left not knowing where a dress is.

Start with the thing that is always late

If you are not a bridal studio, run the same six steps on whatever moves between people in your business. A job card, a repair, a quote waiting for sign-off, a stock transfer. The structure holds: stages, a record, one-tap moves, one named person per prompt, and a page that answers the question your clients keep asking.

Start with the item that causes the most phone calls, not the biggest volume. The phone call is the signal that a human is doing something a list could do, and the phone call is where the hour went.

Do one of these properly and you will still be maintaining it a year later. That is the only test that matters.

Common questions

How do I stop staff forgetting to update a job tracker?

Make the update a single tap on a phone form, and make it the last thing they do before the item leaves their hands. A tracker filled in at the end of the day from memory is the usual reason these systems get abandoned.

Do I need new software to work out where a job is?

Not at first. A spreadsheet with columns for the current stage and the date it reached that stage covers the first month. Move to a proper tool once the stages and columns have stopped changing.

How do I keep clients updated on a job without ringing each one?

Send an update only when the job changes stage, and make it plain and short. Clients mostly want to know it is moving and when they can collect, so one message at the ready stage does more than four progress updates.

Want this done for you?

I build one small automation at a time for North Shore businesses, usually starting with whatever is costing you the most hours.

Start a project ↗