I once timed a client’s order clerk with my phone stopwatch and didn’t tell her. It felt sneaky while I was doing it, which should have been my first clue.
She noticed within the hour. My phone was lying face up next to her keyboard, and I tapped it every time she opened a new email, which is about as subtle as a smoke alarm. She asked me, very politely, whether I was checking how slow she was. I said no. That was true, but it didn’t matter, because from her chair it looked exactly like I was checking how slow she was, and she was right to be annoyed. I apologised, explained what I was really after, and she ended up timing the rest of the week herself (with better notes than mine, and a kitchen timer she brought from home as a joke at my expense).
Her notes were the useful part.
Most of the small wholesalers I keep books for have never put a number on manual order entry. They know it takes “a while” and goes wrong “sometimes.” Both of those can be measured in a week, without buying anything, and the answer isn’t always that you should change how you work.
Time it for a week, and say so out loud
Pick a normal week, not the one before a holiday. Tell whoever runs the order desk what you’re doing and why: you’re measuring the orders, not the person, and the messy formats are the thing on trial. Then give them a simple sheet to fill in as they go (a spreadsheet works, so does a printed page and a red pen). They know where the time goes better than you do.
- When the order arrived, and when they started on it
- When it was fully entered, including any back-and-forth with the buyer
- How it came in: email text, PDF, photo, phone call, voicemail
- Whether they had to ask the buyer anything
- Anything odd, like an item code they didn’t recognise or a price that didn’t match
Time the whole thing, from opening the message to the order sitting finished in QuickBooks Online or your inventory tool. Not just the typing. The typing is often the fast bit. What eats the time is working out which “chicken thighs” a buyer means, scrolling back through last week’s order to check their usual quantity, and waiting for someone to answer a question about a substitute. If you only time keystrokes, you’ll get a lovely small number that has nothing to do with your actual morning. At that first client (a drinks wholesaler with a short product list and very chatty buyers), the full start-to-finish time came out at about 6 minutes an order, on roughly 30 orders a day.
That’s most of a morning, every morning. Nobody had said it out loud before.
Averages hide a lot, though. The repeat email from a regular took almost no time at all, while the photo of a handwritten list from a restaurant kitchen took ages, and so did anything left as a voicemail. Sort your sheet by format and you’ll see which orders actually eat the morning.
Picking, delivery or month-end: where did it go wrong?
Time is the easy half. Mistakes are harder, because they rarely show up where they happen. A wrong line typed at the order desk turns up later, somewhere else, usually on somebody else’s desk.
If it’s caught at picking, it’s cheap: the picker sees a case count that looks odd, walks over, and it’s fixed in the time it takes to ask. If it slips past picking, it gets caught at delivery, and now there’s a driver standing at a customer’s back door with the wrong product, a phone call to the office, and maybe another trip out. If it slips past delivery too, it surfaces at month-end, when the customer queries their statement and someone raises a credit memo for an order nobody remembers. Same mistake every time. The later it surfaces, the more people touch it, and the more it costs to undo.
So during that same week, add another column: every one of the order entry errors you hear about, and where it was caught.
The month-end ones are already sitting in your books. In QuickBooks Online, run the Transaction List by Customer report, customise it, filter Transaction Type to Credit Memo, and set the date range to last month or last quarter. Xero and Sage call them credit notes, and you can pull the same list there. Then read the memo or reference on each one (assuming somebody filled it in, which is a big assumption) and mark the ones that trace back to a typed order rather than damaged stock or a short delivery.
The first time I did this properly, I got carried away. A different client, a small wholesaler supplying cafés, had 9 credit memos in one month that traced back to orders typed wrong. I added them up at 11 p.m. at my kitchen table, felt my stomach drop, and drafted an email to the owner recommending they “change everything about how orders come in.” Luckily I didn’t send it. The next morning, with coffee and daylight, the picture was much smaller: most of them came from one buyer whose PDF put the product code right up against the quantity column, and from a pair of oat milks with nearly identical codes. That was a product-list fix and a friendly call to the buyer, not a rebuild of the whole order desk.
Tired brains round up.
(Lookalike codes trip up software too, by the way. If you ever hand the reading to a tool, I’d still want a person looking at the lines it isn’t sure about.)
“I never paid for the typing”
The owner of that café supplier said something on our follow-up call that I’ve since repeated to nearly every client. “I never paid for the typing. I paid for the phone calls afterwards.” He meant his own time, mostly: the driver ringing him from a customer’s loading bay, the awkward apology to a regular who’d been shorted, the evening spent digging through old emails to work out what the buyer had actually asked for. His order clerk was the lowest-paid person in that whole chain. Every unclear order that slipped through landed on someone whose hour cost more, and none of those hours showed up on any report I’d ever run for him.
So when you count mistakes, also note who fixed each one and roughly how long it took. Owner, sales rep, driver, bookkeeper (yes, me, raising the credit memo and chasing the reconciliation). That column is usually the one that settles the argument, in either direction.
When the numbers say leave it alone (and when they don’t)
Sometimes the honest answer is: keep typing. If you get a handful of orders a day from a few buyers who send them the same way every time, your product list is short, and the person on the desk knows every regular’s usual order by heart, manual order entry is probably the cheapest system you’ll ever run. Spend the effort on a clean product list where no codes look alike, and maybe a simple order form for the buyer who keeps sending blurry photos.
If your week comes out the other way, the choice gets clearer. Lots of minutes, mistakes that reach delivery or month-end, fixing that keeps landing on the owner: then it’s worth comparing options. A few buyers might use an ordering portal if you offer one, though most of the small ones won’t ever switch to EDI. Template-based scanning works when every buyer’s PDF looks the same and stays the same, which is rarer than it sounds. There are also tools that read emailed, PDF and photographed orders and match each line to your own product list and prices, ask a person when they’re unsure, and hand back a clean file for you to import into your accounting or inventory software. Whichever you look at, check it against your own sheet, not against the brochure.
And if your number was small? Then you’ve spent a week confirming that your order desk works, which is worth knowing too.
Before Monday, make the sheet: arrival time, start time, finish time, format, questions asked, plus a column for each mistake, where it was caught and who fixed it. Pull last month’s credit memos in daylight, not at 11 p.m., so you have something to compare against. Then walk over to whoever runs your order desk, tell them exactly what you’re measuring and why, and hand them the sheet.
