Real questions come up on calls every week, demos, onboarding, support. Most of them don't go anywhere afterward, a good answer given once, to one person, then forgotten. Here's a handful from the last few weeks, kept on record instead of left to rot in someone's call notes. 

This one's a bigger edition than most, because it's been a busy stretch. A demo with a skincare brand's warehouse director covers three of these on its own. The rest are scattered across other calls from the same few weeks: a fragrance brand setting up their first system, a soft furnishings brand doing around 120 orders a day, a couple of supply chain managers who'd clearly had the same argument with their current software more than once before they got to us. It's as good a place as any to start, with the questions that have come up most often lately, gathered into one edition. 

Different businesses, same underlying pattern, and it's one I'll keep pointing out until vendors stop making it necessary: the fix is almost never a person trying harder. It's one missing field, one manual step nobody automated, or one assumption a vendor baked in that doesn't hold at your volume. None of these are hard problems. They're just problems nobody senior enough had been annoyed by yet. 

Can You Prioritize Picking by Courier Cutoff Time? 

A warehouse director at a skincare brand put it plainly: the courier collects at six, and by two there's still a pile of orders sorted by nothing more useful than when they arrived. Yes, a pick queue can be sequenced by time remaining to cutoff instead of order age, so a next-day order due in four hours outranks a B2B order with three days' grace, and the queue re-sorts itself as each cutoff passes, on its own, with no supervisor playing air traffic control from memory. Worth naming the bit that undoes the fix if you skip it: this brand told us straight that picking wasn't their constraint, packing was. Reordering the queue moves the bottleneck to whatever station is slowest. Measure that before you assume you know where it lives. 

Can You Trace Who Built a Kit If a Customer Complains? 

Same call, a different worry: "when we've built the kit and I find something's wrong with it, how can I go back and look at who actually built it?" Most systems can't answer that, because a kit build gets logged as one stock adjustment, components in, finished kits out, nothing recorded in between. Real traceability needs a works order, scanned component movements, an operator attached to the build, and batch numbers captured at the point of assembly. Their own rule was better than anything we'd have written ourselves: "it's critical to scan everything that moves around the warehouse." Skip that, and a supplier recall turns into quarantining every kit that might be affected, instead of a short list of the ones that are. Most warehouses learn this the expensive way, once, and never again. 

Can You Stop Pickers Taking Stock Reserved for a Retail Order? 

Same brand again, protecting a large retail order by physically carrying the stock to a corner and labeling it by hand. It works until someone doesn't read the label, or the website sells it anyway because the stock file never knew it was spoken for. The actual fix sits in the system, not the floor: exclude the location from available stock, and every sales channel recalculates immediately. One distinction the call kept circling back to, and I'll be pedantic about it because it matters: this isn't quarantine. Quarantine is stock you can't sell because something might be wrong with it. This is stock you won't sell because it's promised to someone. Same mechanism, opposite reason, and your reporting should keep them separate. 

What’s the Right Way to Name Warehouse Locations? 

A fragrance brand setting up their first WMS started naming shelves top to bottom, A, B, C, until someone on the call asked the question that mattered: "if we end up with loads of racks, we might run out of letters." The fix is hierarchy, not cleverness: rack letter, shelf letter, position number, so BA4 reads as rack B, top shelf, fourth box, and a tenth rack just becomes J instead of breaking the whole scheme. The more common mistake isn't the letters, though. It's naming a location after whatever's currently sitting on it. Locations are coordinates, not descriptions; that's the system's job to remember, not the shelf's. It's a five-minute decision that people somehow manage to get wrong for a decade. 

What’s the Best Layout for a Small Warehouse? 

A founder walked us round his warehouse on his phone camera: racking in a U shape, pickers working the inside, five steps to any shelf regardless of where they stood. Already closer to right than most warehouses ten times the size, and the physics holds regardless of scale: pick rate is a function of walking distance, so layout is just how you choose to spend it. Two changes mattered more than any new racking would have. Separate packing from picking once you move past pick-and-pack-as-you-go, because a bench full of boxes and label printers slows both jobs if it sits inside the pick route. And stop reserving shelves by product; let any empty shelf take the next delivery and let the system remember where it went, instead of shuffling stock to preserve a layout nobody needs. 

How Do You Stop Pickers Walking the Warehouse Once Per Order? 

A soft furnishings brand doing around 120 orders a day was sending someone into the warehouse once per order: pick one, pack it, walk back for the next. The fix depends on what's in the basket. Single-item orders get bulk-picked as one wave and sorted at the bench, because nothing needs matching mid-aisle. Multi-item orders get tote-picked instead: a trolley of barcoded trays, one circuit of the warehouse, and the system assigns each order to a tray the moment its first item is scanned, not before. Ten trays means ten orders picked in the time it used to take for one, and the fact that this still lands as a revelation on calls says more about how low the bar's been set than about the idea itself. 

Why Can You See Every Outbound Order Live but Not What’s Coming In? 

A supply chain manager named something obvious in hindsight: outbound gets a dashboard because late dispatch is visible pain, and inbound doesn't because its pain is internal, an overtime bill, a truck idling on the yard. What he wanted: "I know how many are going to arrive this week. I need this many guys on inbound, I need this many guys on outbound." The data already exists in any system that tracks purchase orders properly, supplier lead times, PO status, expected delivery date, it's never been put on a screen. Nobody built the inbound version first because nobody on the outbound side ever complained about it loudly enough. That's not a technical limitation, it's whoever shouted last. Inbound planning isn't a forecasting problem. It's a visibility problem wearing a forecasting costume. 

Should Shipping Labels Print Before or After Picking? 

A brand running a lean team pre-prints a batch of labels, then picks and sticks as they go, because it saves a walk to the printer. The honest cost shows up the day stock runs out mid-pick: a customer gets a dispatch email and a tracking number for a parcel that doesn't exist. In their own words, "we get to a box and it's empty." Nobody puts "we occasionally tell customers their order's on its way when it isn't" in a sales deck, but that's the honest description of what pre-printing does at scale. Print after the pick is scanned and validated instead, and the label becomes proof a correct pack happened, rather than a promise made in advance of checking. It costs a few extra steps back to a printer. Wrong-item claims and phantom dispatch emails cost considerably more than that. 

Why Do International Parcels Get Refused for the Wrong Weight? 

A soft furnishings brand weighs every international order by hand, edits the order, then prints the label. Their reason: "it gets refused if the weight's not accurate." Certain destinations, islands especially in their experience, will bounce a parcel outright when the declared weight doesn't match what's on the scales at a fine pricing band. The actual fix is duller than the workaround: declared weight should be item weight plus packaging weight, both held as system data, so the label calculates the true combined figure and scales stay on the bench for spot checks only. Their software simply had nowhere to enter a box weight. That's the real bug: a missing field, not a missing habit, and an oddly expensive one to leave unfixed for something this boring to solve. 

Can You Run Two Royal Mail Contracts on One Account? 

A Royal Mail account manager told this brand to split their volume across two contracts and save roughly three pounds a parcel, because their single blended rate was priced on an average weight dragged up by their heaviest items. Their software could only hold one account, so the saving sat unclaimed. Yes, you can hold two contracts and route between them by weight band automatically, light parcels to one, heavy to the other, provided the platform treats each contract as its own courier instance rather than assuming one carrier means one account. Software shouldn't be the reason your own account manager's advice doesn't work. The carrier offered the saving. The software decided whether they were allowed to take it. 

Is It Ever Cheaper to Ship Two Boxes Instead of One? 

Past a carrier's volumetric threshold, yes, and a supply chain manager we spoke to already knew the number, he had it trapped in a spreadsheet his IT team maintained by hand: "if the order contains these SKUs, use the cheaper carrier." A static list rots the day a product's dimensions change. The alternative is a rule rather than a list: hold real dimensions on every SKU, calculate volume and weight per order at the moment it ships, and branch automatically. His own description of the fix was cleaner than ours would have been: "multiply all my dimensions together, and say if it's over a certain value, ship it differently." One number in one rule, editable the day the rate card changes. His IT team could be spending that spreadsheet-maintenance time on anything else. 

What Happens When a Marketplace Issues Its Own Shipping Label? 

Very books its own courier and hands sellers the label. TikTok Shop runs something similar, though sellers who'd rather use their own courier there pay for the privilege. Marketplaces are generous about handing you their label requirements. They're considerably less generous about handing you the admin budget to manage them. One of these is a minor inconvenience. Three is a packer doing a different manual routine per sales channel, downloading a portal's labels each morning and matching them to parcels by eye. The fix is to treat the marketplace as a channel and a courier at once, so its label arrives at the packing bench through the same scan-and-validate flow as everything else. The label prints because the pack was checked, not because someone remembered to visit a portal that day. 

What Does It Cost When Your WMS Won’t Integrate a Carrier? 

A supply chain manager wanted Amazon Seller Fulfilled Prime. His vendor refused, called itself a WMS company rather than an IT company, and pointed him at a third-party platform instead. His reply is the whole argument on its own: "what's the difference? A WMS is software." He's right, obviously, and any vendor who needs that explained to them has told you something important about the next three years of your relationship. Refuse to see that, and the cost stacks in three places: a per-parcel toll to a middleman, an IT team quietly rebuilding the vendor's integrations themselves because support moves too slowly, and the commercial deals you can't do while you wait for permission. Ask a vendor who owns their carrier integrations before you sign, not after. 

Should a First-Time Buyer Take a Trial or Go Straight to Onboarding? 

Two founders, first warehouse system, weighing a free trial against signing properly with a break clause. A trial is a good idea for someone who already knows what they're checking for. For everyone else it's an expensive way to discover, after the fact, that they didn't know the right questions to ask. The general rule: a trial measures the software, onboarding measures the fit, and only one of those questions matters if you've never run a WMS before. An experienced operator can extract a verdict from a trial in two evenings with a checklist. A first-timer opens an empty system with no instinct for what to test first, and often concludes nothing, not because the software failed, but because the evaluation never really ran. One of the founders settled it herself: "I'd rather do it properly." That's risk management, not caution for its own sake. The real cost of a first WMS was never the subscription. It's running a warehouse on a badly configured system for years because nobody wanted to commit to finding out properly. 

Fourteen different problems, and underneath nearly all of them the same shape: a person compensating, by hand, for something the system should have been holding. A label walked to a printer. A shelf labeled in pen. A spreadsheet an IT team maintains out of habit. None of it is stupidity. It's usually the sensible response to a gap that's been there so long nobody questions it any more, which is exactly why an outsider asking an obvious question on a call keeps being the thing that moves it. 

If any of these sound like your Tuesday rather than someone else's, bring the messy version to a demo, not the tidy one, and we'll show you what changes: talk to us.