When a family business outgrows the way it has always worked

A lot of the strongest businesses in this region were built on judgement rather than process. Somebody knew what to charge. Somebody knew which customer was slow paying but good for it, and which one to refuse politely. None of it was written down because it did not need to be. The person who knew was standing in the building.

That stops working at a particular size, and it stops working quietly. There is no crisis. Things simply get slower and less certain, and everyone assumes it is a bad quarter.

The signs, roughly in the order I meet them

• The founder is the only person who knows what a job should be priced at. Quotes queue behind him. He is on a flight and a customer goes elsewhere.
• Ask three people how much stock is in the warehouse and get three different answers, all of them given in good faith.
• Month end takes a week, and a good part of that week is spent reconciling what the system says against what people know.
• A customer disputes what was agreed, and establishing the truth means asking four people and scrolling back through a chat thread.
• The second generation has joined, asks which product line actually makes money, and cannot get an answer anybody will stand behind.

The old way was not primitive

This is where consultants are usually insulting, so let me be careful. The way you have worked was fast, carried almost no overhead, and beat competitors who were doing everything properly on paper. Judgement held in one head is enormously efficient right up until the head runs out of hours.

What changed is volume and distance. A founder can hold two hundred customers and forty suppliers in mind without effort. He cannot hold two thousand, and he certainly cannot do it from another emirate while three branches make decisions in his name and each of them decides slightly differently.

The aim is to write down the process without losing the judgement

There is a mistake available in each direction. One is to carry on unchanged and burn a year of the next generation's patience while good people leave for businesses that answer questions faster. The other is to buy a large system, push every decision into it, and end up slower than before with the one thing you were good at now sitting outside the process.

What works is narrower than either. Take the decisions that repeat and follow from facts, and write those down as rules the system applies: pricing bands, credit limits, reorder points, who may approve which discount. Leave the genuinely difficult calls with the people who are good at them, and make the system record what was decided and by whom.

The founder's judgement stays. What goes is the founder being the bottleneck for the ordinary ninety per cent that never needed him in the first place.

Getting the pricing out of one person's head

This is the hard one, and it is rarely a software task. What works is sitting down with fifty real quotes from last year and going through them one at a time: what did we charge, why that number, what would have changed it. Patterns start falling out within the first hour. Volume bands, customer type, how far out the delivery was, whether the client had been slow to pay before.

Most of it turns out to be a rule that nobody had ever said out loud. What is left after the rules are extracted is the actual judgement, and it is usually a much smaller share than anyone expected.

Expect resistance, and expect it to be reasonable. Writing a rule down means it can be checked, and checked means argued with by people who have not earned the right. For some founders the knowledge is also the position, and that is a real thing to work through rather than a problem to engineer around. I have watched projects stall on exactly this and no amount of software fixed it.

Where to start

Not with the whole business. Pick the one thing that is measurably slow and visibly annoying: quotes waiting on one person, or the stock count, or the week that month end eats. Get that into daily use and let people watch the new way work before you ask them to trust it with anything else.

We built a till and ERP system for the apparel trade here, for tailors and abaya and kandura ateliers, which are almost all family businesses of exactly this kind. Measurements held against a customer, deposits, work moving between branches, order updates going out over WhatsApp. None of that removed anybody's judgement about a difficult customer. It removed the part where four people had to be asked what stage an order was at.

If you recognise three or four of the signs above, the useful next step is not a software demonstration. It is a couple of weeks working out which decisions can be written down and which genuinely cannot. Send me a description of where things are getting stuck and I will tell you what that usually involves.