There is a version of this in nearly every small business. Somebody spends Monday morning, or Friday afternoon, exporting from the accounting system, the job system and a spreadsheet somebody maintains, pasting it all into a workbook and emailing it round.

It takes two to four hours a week, it depends entirely on one person and roughly half of what it produces is never read.

Start by deleting reports

Before automating anything, work out what should stop existing. For each report currently produced, ask three questions:

  • Who reads it? Ask them, do not assume. The answer is frequently nobody, and has been for two years.
  • What decision does it inform? If the honest answer is it is good to know, it is not a report, it is a habit.
  • What happens if a number moves? If nobody would do anything differently, the number does not need reporting weekly.

In most businesses this conversation removes a third to a half of what is produced, and that is free time recovered before any automation is built. Automating a report nobody reads is a very efficient waste.

The four reports most small businesses need

1. Cash

Not the profit and loss. Cash in the bank, what is owed to you and when it is due, what you owe and when and the projection for the next twelve weeks.

Weekly. This is the report that prevents the specific kind of unpleasant surprise that closes otherwise healthy businesses.

2. Pipeline

Quotes out, their value, their age and win rate by work type. Weekly.

The age column is the one that earns its place. A quote sitting untouched for three weeks is either lost or being neglected, and both are worth knowing about. This works properly only if quotes are tracked, which is covered in automating quotes and proposals.

3. Delivery

What is in progress, what is late, what is at risk. Whatever your equivalent of a job list is.

Daily if your work moves daily, weekly otherwise. Match the frequency to how quickly you would act.

4. One operational measure

Specific to your business. Utilisation for a professional services firm. First-time fix rate for a service business. Stock turn for a distributor. Response time for anybody who takes enquiries.

One. The temptation is always to track twelve, and twelve measures produce no focus at all.

A report is a prompt to act. If a number would not change anybody's behaviour this week, it belongs in a dashboard somebody visits, not in a report that arrives.

How to automate it

The principle is that data moves from the source system to the report without passing through a person.

Connect rather than export

Most business systems — Xero, QuickBooks, Sage, HubSpot, most job management platforms — offer either a direct connector or an API. Power BI and Looker Studio both read from these on a schedule.

Once connected, the report refreshes itself. Nobody assembles anything, and it is current at the moment somebody opens it.

Push a digest, host the detail

The pattern that actually gets used. A short message each Monday with the four or five numbers that matter and anything outside its normal range. A link to the dashboard for anybody who wants to investigate.

Dashboards alone are rarely opened. Long email reports are rarely read past the first section. The digest plus dashboard combination is what works.

Alert on the exceptions

The highest-value part, and the part usually left out. Rather than expecting somebody to notice a number in a table, alert when it crosses a threshold: cash projected below a floor, a quote unchased past fifteen days, a job late, an operational measure outside its normal range.

Reports tell you what happened. Alerts tell you when to do something.

What makes automated reporting fail

Nobody owns the numbers. Each measure needs a named person who is accountable for it and who is expected to explain movements. Otherwise the report is observed rather than used.

The source data is wrong. Automation makes bad data arrive faster and look more official. If your job system is only updated when somebody remembers, fix that first — the fix is usually connecting systems so records maintain themselves.

Too many measures. Twenty numbers means no focus. Five means a conversation.

No comparison. A number on its own is meaningless. Every figure needs either last period, the same period last year or a target next to it.

Where to start

Cash, weekly, automated. It is the report with the highest consequence and usually the easiest to connect, because accounting systems have the best integrations.

Then pipeline. Then delivery. Add the exception alerts once the reports themselves are stable and trusted.

Expect two to three days of work for a small business, and the recovery of two to four hours a week of somebody's time permanently.

If you would like reporting connected to your actual systems rather than rebuilt every week, that is what our system integration and automation teams do. Tell us which spreadsheet somebody rebuilds every Monday.