Reducing rework and errors in admin: where to start
By Keith Hamilton · · 4 min read
Reducing admin errors starts with the hidden second process behind every admin task: finding and fixing the mistakes from the first one. A wrong price on an order, a customer set up twice, a delivery address typed with one digit out: each error takes minutes to make and can take hours to put right. This guide explains how to find where rework comes from, how to put a value on it, and how automation reduces errors without introducing new ones.
Why does rework cost more than it looks?
Rework costs more than it looks because a task done wrong is handled four or five times instead of once. Someone notices the problem, someone investigates it, someone corrects it, someone tells the customer or supplier, and sometimes someone corrects the knock-on effects in another system.
Rework also lands on your more experienced people, because they are the ones who know how to unpick a mess. And some errors reach customers, in the form of a wrong delivery or an incorrect bill, which costs goodwill as well as time.
Where do admin errors come from?
Most admin errors come from the way the work is set up, not from carelessness. The main sources are:
- Re-keying. Every time data is typed from one place to another, there is a chance of a slip. Our guide to automating data entry between systems covers data entry errors in detail.
- No checks at the point of entry. If a system accepts any value, a typo goes straight through.
- Several versions of the truth. When customer or product details are held in more than one place, they drift apart.
- Unclear process. If two people do the same task in different ways, the results differ.
- Interruptions and rush. Repetitive work done between phone calls, or at month end under pressure, is where slips happen.
How do you measure the rework you have?
Ask the team to log every error they find or fix for two to four weeks, before changing anything. For each one, note:
- What the error was.
- Where it started.
- Roughly how long it took to put right.
A shared spreadsheet is enough. The aim is to find the few sources that cause most of the trouble, so you fix those first.
How do you calculate the cost of rework?
Multiply the number of errors by the time each one takes to put right, then by a loaded hourly cost. Here is a worked example for a team that finds 15 order errors a week, each taking 30 minutes to investigate, correct and explain to the customer:
- Time a week: 15 errors at 30 minutes is 7.5 hours.
- At a loaded hourly cost of £22, that is £165 a week, or £8,580 a year.
That is the time cost alone. It leaves out credit notes, wasted deliveries and lost customers, which you add if you know them. Preventing half of those errors is worth £4,290 a year in released capacity.
As with any saving of time, it becomes cash only where real spending stops. Our guide to the real cost of manual work explains the method.
How does automation reduce admin errors?
Automation reduces admin errors by removing the steps where they start, rather than by catching them later:
- Copy, do not re-type. Pass data between systems directly, so it is entered once and used everywhere.
- Validate at the point of entry. Check details as they come in: does this postcode exist, is this product code real, is this customer already on the system, is this quantity in a sensible range?
- Use one source for key details. Keep customer, supplier and product details in one place and have other systems read from it.
- Flag the unusual. Have the system highlight orders or entries that look different from normal, such as a price much higher than usual, so a person checks them before they go out.
AI adds the ability to read unstructured inputs, such as emails and documents, and turn them into structured data that is then checked against these rules.
Can automation make errors too?
Yes. Automated processes can misread a document, match the wrong record or apply a rule in a case it was not designed for. The difference is that an automated error repeats, quickly and at volume, until someone notices.
That is why a good set-up:
- Sends uncertain cases to a person rather than guessing.
- Records what it did and why, so errors can be traced.
- Is checked by sampling, especially in the first weeks and after any change.
- Has a named owner who is told when something fails.
Measure the error rate before and after. If it has not gone down, the automation is not doing its job.
Where do you start with admin process improvement?
Start by logging errors, then fix the source that costs the most time:
- Log errors and rework for a few weeks to find the main sources.
- Pick the source that causes the most time lost.
- Fix the process first, such as agreeing one way of doing the task, then automate the repetitive part.
- Measure the error count and the rework time against your baseline.
Where Forwardcycle fits
Our cost reduction service helps you find where admin time and errors actually come from, automate the repetitive steps and measure the results against a baseline. To see where it fits your business, take the free AI Readiness Assessment.