Using Exception Reporting So Managers Only See What Matters
A manager with thirty stores does not need thirty sales reports a day. They need to know which three stores need a phone call. How to design exception reporting that people trust.

A regional manager responsible for 30 stores receives a daily sales report with 30 rows and 15 columns. On a good day she scans it in ten minutes. On a busy day she does not open it, and the one store whose card terminal was down all afternoon goes unnoticed until the weekly review.
Exception reporting turns this around. Instead of showing everything and hoping the reader spots the problem, the system checks every number against an agreed rule and shows only what broke the rule, to the person who can act on it.
The principle: normal does not need attention
Most numbers in most reports are within the normal range most of the time. Reading them confirms that things are fine, which has some value, but it consumes the attention managers need for the few items that are not fine. Exception reporting is built on a simple bargain: if you hear nothing, it means everything is within limits. If you hear something, it matters.
That bargain only works if people trust it. Two things break trust quickly. The first is missing a real problem, because a threshold was set too loosely or a data feed failed silently. The second is crying wolf, sending so many alerts that people learn to ignore them. Good design is mostly about avoiding both.
Four kinds of rules
Most useful exceptions fall into one of four types. Most good rule sets use all four.
| Rule type | Example | Good for |
|---|---|---|
| Absolute threshold | Stock of an item below five days of cover | Limits that matter regardless of history |
| Relative to target or history | Store sales below 75 percent of the same weekday average for the past four weeks | Spotting unusual days without fixed numbers |
| Trend | Gross margin down three weeks in a row | Slow problems that never cross a single line |
| Missing or late data | No transactions from a branch by noon | Catching broken systems and broken processes |
The missing data rule is the one most often forgotten and often the most valuable. A branch with zero sales is not having a quiet day. Something is wrong with the till, the connection or the opening, and somebody should call.
Every exception needs an owner and a next step
An alert that goes to a group mailbox is an alert nobody owns. For each rule, define three things before switching it on:
- Who receives it. The person who can act, usually the branch, category or account owner, not their director. Directors get a summary of open exceptions instead.
- What they are expected to do. Investigate and explain, take a specific action, or escalate. Write it in the alert itself: “Check the terminal and confirm by reply” is clearer than a bare number.
- When it escalates. If the exception is still open after an agreed time, it moves to the next level. This keeps alerts from being silently ignored.
Where possible, let the owner record a short reason when closing an exception. Over a few months those reasons become a valuable record of what goes wrong and how often.
An illustrative scenario: a home goods retailer in two countries
Take a home goods retailer with around 25 stores across Jordan and Saudi Arabia, plus an online shop. Head office produces a daily sales pack and a weekly stock pack. Area managers say they spend their first hour each morning reading reports and still miss problems.
The operations and finance teams agree a starting set of rules. Store sales below 70 percent of the four-week weekday average by 6 p.m. No transactions from any store by 11 a.m. Any of the top 200 items with less than a week of stock in a store that sold it last week. Refunds in a store above twice their normal daily level. Online orders not dispatched within 48 hours.
Each rule runs on data already flowing from the point-of-sale and ERP (enterprise resource planning) systems into the reporting database. Alerts go by email or Teams message to the store manager and area manager, with the reason fields captured in a simple form. The daily sales pack is still available on demand, but most managers stop opening it. Instead, the weekly meeting starts with a list of exceptions from the past seven days and the reasons recorded against them.
After the first month, two rules are adjusted. The refund rule fires too often in stores running promotions, so promotion days are excluded. The stock rule is extended to the top 300 items because some important seasonal products were missed.
Building it with tools you probably have
Exception reporting does not require a special product. Power BI supports data alerts on dashboard tiles and can be combined with Power Automate for routing. Many ERPs have built-in alert features. A scheduled script or a workflow tool such as n8n can query a database, apply the rules and send messages. The tool matters less than three supporting practices:
- Clean, timely data. If the source data arrives late or contains duplicates, rules will fire for the wrong reasons.
- Rules written in plain language and kept in one list, with an owner who can change thresholds when the business changes.
- A monthly review of how many alerts fired, how many led to action and how many were noise. Retire or adjust the noisy ones.
Less reading, faster action
To start, take the report your managers read most often and ask which three things on it would make them pick up the phone. Those three become your first exception reporting rules, each with an owner and an escalation time. Designing those rules, building the data workflows and setting up automated alerts is part of our Intelligent Automation service, built on the principle of the KEMLAT Method that you automate only what deserves it.


