Roost where small businesses come home
← all posts
← back to the Journal
October 11, 2026·5 min read·by Lynsie

The thing your numbers will never tell you

Your till knows exactly what sold. It does not know that a screen has been dark for nine days, or that your best cook is about to quit. Four questions, once a week, fixes that.

This one was not my idea. It came up in a conversation about a business with several locations, and once it was said out loud it was obviously the most important thing in the whole build.

Here is the problem. An owner with more than one location has excellent data and almost no information. The point of sale tells them revenue, covers, best sellers, hour by hour. What it cannot tell them is that the ice machine is making a noise, that a regular complained twice last week, or that the person who actually holds the kitchen together is quietly looking elsewhere.

Four questions, once a week

So each location got a section called the Report In. A manager fills it out once a week, and it takes about two minutes.

How did the week go. What is worth watching. How are your people. Is there anything you need a decision on.

That is it. It lands in one stream at the top level, alongside every other location, and the owner reads four short notes instead of chasing four phone calls.

the tell

The board shows how many locations have reported this week. Three of four, with the missing one named. Not to be punitive. Because the week somebody stops reporting is usually the week something is wrong.

Worth watching is the important one

The four questions are doing different jobs. How it went is warm-up. People is the one that saves you money, because replacing a good cook costs more than almost anything else on the list.

But worth watching is the one I would not give up. It asks for the small thing that is not a problem yet. A screen dropping out for a few seconds around six. A supplier arriving later than they used to. Nobody would ring you about those. They are exactly the things that become expensive three months later.

Why it has to be short

Every version of this that fails, fails the same way. It becomes a form. It grows fields. Someone adds a dropdown. Then it takes fifteen minutes, and by week five nobody fills it in and the owner concludes their managers do not communicate.

Four questions, free text, two minutes. If it ever takes longer than that, it has been built wrong.

It works at home too

I keep noticing this pattern is not really about restaurants. Anyone who is not physically present has the same gap. A family with a house sitter. A board with a part-time director. The numbers travel fine. The noticing does not.