A job tracker that got maintenance jobs out of WhatsApp groups
A property maintenance company tracked jobs across phone calls, email and WhatsApp groups, so nobody could say which jobs were late. We built one job board with technician sign-off and automatic updates to landlords.
- Client
- A property maintenance company
- Industry
- Facilities services, 30 staff
- Engagement
- 8 weeks, then a monthly retainer
- Published
Note.Client name is anonymized and figures are illustrative until publication is approved.
- 5
- Late jobs per month, from 22
- 2.5 days
- Average job completion time, from 4.6 days
- 11 hrs
- Office hours per week on status updates, from 19
Business context
A property maintenance company of thirty staff looks after rental homes for landlords and letting agents. Work ranges from a leaking tap to a full boiler replacement. Office staff take calls all day, and technicians are on the road. The owner asked for a job tracking app for maintenance companies because the current setup was costing them repeat visits and trust. We built it as a standalone web application alongside their existing systems.
The problem
Jobs arrived by phone and email. Someone typed the details into a spreadsheet, then copied them into a WhatsApp group for the right area. Technicians replied "on it" or sent a photo, and the office tried to remember what had actually been done.
Nobody could answer a simple question: which jobs are late? The spreadsheet had dates, but nobody updated them once the day got busy. Landlords called to ask for progress, and staff had to ring the technician to find out.
Urgent jobs were handled first, which is correct, but the quiet ones slipped without anyone noticing. A repair could sit for a week because no single person owned the list.
What we built
We built one place for every job. Office staff create a job from a phone call or a forwarded email, set the property, the category and the priority, and assign it to a technician. The technician sees it on their phone with the address, contact details and any photos that came with the report.
Statuses move through a short list: new, scheduled, in progress, done. Technicians upload photos before and after the work and add a sign-off note. That note and the photos are what the office sends to the landlord, so nobody has to chase for evidence.
Landlords get an automatic update when a job is scheduled and when it is closed. The weekly report lists overdue jobs by technician and by landlord, which the office manager uses to plan Monday morning. We sent the updates through Twilio, so the same text message works for landlords who never open email.
A small detail mattered more than expected: the photo upload had to work on a weak signal. It now saves the photo on the phone and sends it when a connection returns.
Rollout
- Week 1 to 2: Sat with office staff and one technician, and wrote down every status a job could be in.
- Week 3 to 5: Built the job board, the mobile pages and the landlord updates.
- Week 6: Ran the app beside the spreadsheet so staff could compare, without abandoning WhatsApp yet.
- Week 7 to 8: Moved every job into the app, kept WhatsApp for chat only, and trained technicians in small groups.
Before and after
| Step | Before | After |
|---|---|---|
| Intake | Phone and email, retyped | One form, from a call or email |
| Tracking | WhatsApp groups and a spreadsheet | One job board with live status |
| Evidence | Chasing photos | Photos and sign-off attached to the job |
| Landlord updates | Staff called or emailed by hand | Automatic on schedule and close |
| Overdue jobs | Nobody knew | Weekly report by technician and landlord |
Results
Late jobs fell from 22 a month to 5. Average completion time for a standard repair dropped from 4.6 days to 2.5 days once technicians stopped waiting on unclear details. Office hours spent on status updates came down from about 19 a week to 11, close to a half day back.
The number the owner cared about most was repeat visits. They dropped because technicians arrived with the right photos and the right parts already noted.
Lessons learned
- Technicians adopted the app only after photo upload worked on a bad signal. The first version failed in basements and behind thick walls, and a few technicians quietly went back to WhatsApp. Fixing the upload brought them back.
- We built a long status list in the first draft. Technicians ignored it and used three statuses anyway, so we cut the list to match what they did.
- Automatic landlord updates looked like the easiest part. They took the most review, because a wrong update is worse than no update. The same caution showed up in the online quote calculator we built for a cabinet workshop, where the price rules had to be agreed before anything could be sent.