A small furniture business had a showroom, a carpentry workshop, two surveyors, an installation crew, and one office manager who held everything in her head. The company handled thirty to forty orders per month, each moving through measurement, drafting, customer approval, production, painting, and installation.
Starting point: the process worked as long as nobody forgot anything
Every stage moved between people through WhatsApp messages and groups. The surveyor photographed a measurement sheet and sent it to the group. The office manager copied it into a spreadsheet. The carpenter worked from a screenshot. The installation date was agreed by phone and entered in the owner’s Google Calendar.
The consequences were consistent and familiar:
- On average, twice a month, an order went into production with an outdated measurement because the customer’s correction remained in WhatsApp and never reached the workshop.
- Average delivery time was 24 working days, while the company had promised customers a standard of 18.
- When the office manager took a week off while her partner was on reserve duty, three installations were mistakenly booked for the same morning with one crew.
What we built
We did not change the carpentry work or Hashavshevet, which the accountant used for finance. We built the layer between them: one order workflow that advances itself.
- One order card on a monday.com production board, created automatically as soon as a quote is approved. Every document, measurement, and message is gathered there.
- A digital measurement form on the surveyor’s tablet instead of a photographed sheet. Measurements go straight into the card, and every customer update is recorded as a new version. If the order is already in production, the workshop receives an automatic alert.
- An automatic task chain. Completing one stage opens the next task for the right person: drawing approval opens production, completed production opens installation scheduling, and the scheduled installation goes directly into the crew’s calendar and sends the customer a WhatsApp confirmation with a reminder one day before.
- A stalled-process alert. An order that remains in the same stage longer than defined triggers an alert for the office manager. No order waits silently.
- A connection to Hashavshevet. Order approval automatically creates a deposit request, and completed installation opens a final billing task. Payment no longer depends on memory.
Results after one quarter
- Average delivery time fell from 24 working days to 15, an improvement of nearly 40%, without adding staff.
- Production errors caused by outdated information fell from two per month to one per quarter.
- The office manager estimates that nearly 15 hours per month were freed from copying, reminders, and “Where is my order?” phone calls.
- For the first time, the owner can see every active order on one screen, including what is stalled and why.
What this story teaches
In most operations-heavy businesses, the problem is not a specific stage but the transitions between stages. Every manual handoff creates another opportunity to forget something. When each stage automatically opens the next one, the process no longer depends on one person’s memory.
Questions we are asked about this kind of process
Our team is not technical. Will this work for us? The carpenter in this story is 58 and had never worked with a management system. His interface is one card containing the measurements and drawing. If a solution requires lengthy training, it is probably not the right solution.
How much did a project like this cost? It depends on the scope, which is why we do not publish a fixed price list. Our principle is to start with the process where improvement will repay the investment fastest. Here, that was the order workflow, and the investment paid back within the first two months.
What happens if monday.com goes down or the internet connection fails? Production continues. The system records and coordinates work, but it does not block the physical process. Anything that happens during an outage is updated once the connection returns.
The client’s identifying details were removed at their request. The data was measured with the client before and after the project.
