The Five Usual Suspects

There’s a very specific conversation that it would be great to see happening in more warehouses, much more often. It’s that one about whether the warehouse KPIs that are being displayed proudly on the dashboard (the ones that have been tracked for the last however many years, and that are trotted out at every monthly review) are actually telling you what you think they’re telling you.

Most warehouses pay close attention to the same five or six headline numbers: Pick rate. Scan accuracy. OTIF. Dock-to-stock. Inventory accuracy. You know… the usual suspects. They show up each month at the board of directors’ (BOD) meeting to eat all the prawn cocktail sarnies. They get reviewed every quarter.

They’re a BENCHMARK, Baby.

But in a lot of operations, they may actually be telling you something rather different to what everyone believes that they’re telling you.

This isn’t a list of metrics to chuck in the bin, by the way. They all measure something. The trouble is that quite often, they don’t measure what a lot of people THINK that they’re measuring. And because of that, you might find that more often than it should be, a perfectly healthy-looking headline number is hiding an operation that’s behaving rather differently underneath.

Here are five of the biggest offenders, in reverse order:

We consider these to range from a little bit sneaky to deeply unhelpful.

5. Scan accuracy

This is the number that every supervisor can quote without even checking their notes. 99.5%, 99.7%, and sometimes it’s even higher. It’s a nice metric that gets trotted out whenever picking quality comes up in a management meeting, and on the face of it, it all sounds lovely.

Except that scan accuracy is not really measuring what most people think it’s measuring.

What it actually tells you is that the barcode the picker scanned matched the barcode the system was expecting. That’s all. Nothing else. It tells you nothing about whether the right quantity went into the tote. It tells you nothing about substitutions that might have been made off-system. It tells you nothing about anything that happens after the scan is complete.

WERC’s 2025 best-in-class order picking accuracy benchmark is 99.68%, with the industry average sitting at 99.49%. The gap between those numbers and what your scan accuracy report is showing is, more often than not, exactly where a lot of your customer complaints are coming from.

A better measure: order accuracy per customer.

Did the right product ship, in the right quantity, at the right time? Blue Yonder’s Dispatcher WMS captures every single thing that you might need to track that information already. The trick is making it easily visible at the right level, to the people who need to keep an eye on it. User Services Portal (USP), one of the Blue Yonder approved products from Socius24 Limited, turns information like this into a role-based dashboard that supervisors can see while they’re actually on the floor, and our latest product, our natural language AI, AskUSP, lets managers ask, in plain English (or whatever language they happen to prefer), which orders had discrepancies last week.

4. Inventory accuracy percentage

This is the headline trust number. It’s the one that your WMS gives you every morning to reassure you that what’s on your screen is what’s actually on the floor.

Most operations report somewhere between 97% and 99%. And WERC’s 2025 best-in-class sits at 99.5% or better, which is a perfectly respectable target. Dispatcher WMS users, incidentally, typically report close to 100% accuracy.

The trouble with this number is that a lot of cycle counts can hammer the fast movers and then accidentally on purpose forget the bulk locations. Which means that your A-items might be spotless, but your slow movers and your pallet locations might simultaneously be doing who knows what, and you’ll only find out what that is when an order short-ships.

A better measure: inventory count accuracy by location, broken down by velocity class.

You’ll want to know what your bulk and slow-mover accuracy looks like, not just the aggregate number. Dispatcher WMS tracks everything at location level by default. And USP can present you with a breakdown without anyone having to ask IT for a custom report. If you want even more info, AskUSP can tell you which locations haven’t been counted in the last X days, as quickly as you can ask it the question.

3. Dock-to-stock time

This is often treated as specifically the receiving team’s KPI and it’s often used to judge whether inbound is keeping up.

What it’s actually measuring, however, is how fast you can get a trailer’s worth of stock into an available status. Not whether that stock eventually ended up anywhere sensible.

WERC’s 2025 best-in-class is under 3.5 hours. Slightly worryingly, that’s slower than 2024’s sub-3-hour benchmark. And APQC data shows a gap of more than 44 hours between the best and worst performers.

Plenty of warehouses (ones that don’t use Dispatcher WMS, with its system directed putaway, for example) can hit a super-duper dock-to-stock stat, just by dropping incoming pallets wherever they happen to fit. The scan will go through nicely, the WMS will record the location, the stock is technically available, and the KPI looks healthy.

The problem only tends to appear a couple of weeks afterwards, and it’ll show up in your pick-path travel time, when those pallets have to be retrieved from wherever whoever it was happened to have parked them.

A better measure: dock-to-stock that’s paired with putaway quality.

How often do your goods land in a velocity-appropriate slot, rather than just any open slot? Speed without sensible slotting is a false economy. Dispatcher WMS’s directed putaway uses velocity, dimensions and configurable putaway rules to drop those pallets in the most sensible place by default. To further confirm that this is happening, USP can make any exceptions visible to supervisors, which means that any pallets that ended up somewhere they probably shouldn’t have will already have been identified, and before they cost you outbound travel time.

2. OTIF (aggregated)

This is the big-customer metric. It’s the one that ends up on the BOD report and the one that gets quoted at industry conferences: On Time, In Full.

What everyone believes that it means: “we’re hitting all of our service commitments”.

What it actually means: if you take the average, across everything, we’re probably hitting most of our service commitments… most of the time.

A perfectly respectable 96% headline OTIF is absolutely capable of hiding a problem. And the mysterious laws of the universe can often mean that those specific problems are with the inventory of the two or three retailers who will absolutely, positively, definitely deduct money when you miss their deadlines.

The 14 largest UK grocers operate under the Groceries Supply Code of Practice, but service level charges continue to be one of the most common reasons that suppliers see unexpected deductions. The Groceries Code Adjudicator’s investigation into a major retailer recently, specifically, called out service level charges as a recurring source of unilateral supplier deductions, and the practice continues across the sector under lots of different names.

And the pattern is absolutely consistent when you start looking more closely at it: an acceptable-looking OTIF aggregate can at the same time hide a small handful of accounts that might turn out to be costing you a lot of money because they’re not achieving agreed SLAs.

And that means that, while the headline might be fine, the customers who are paying your bills might be less so.

A better measure: OTIF that has been disaggregated by customer, with any penalty-likely accounts being flagged separately.

Aggregate OTIF might be the one that you show in the BOD presentation. But Customer-level OTIF is the one that you need for managing the relationships that are actually paying your bills. Dispatcher WMS already captures the data that you need, per customer and per shipment. And again, USP can turn that data into a customer-level scorecard, if that’s what you need.

1. Pick rate / lines per hour

This is usually the most tracked, most reported, and most incentivised number that shows up in any warehouse, anywhere. The Picking process typically accounts for around 55% of total warehouse labour cost, so the pressure to make this number look good is permanent and constant. And in certain circumstances, achieving the relevant KPI can become quite creative.

Officially: it is a measure of picker productivity.

Actually: it’s a number that responds to all sorts of things that have nothing to do with how productive your pickers actually are.

Pick rate is structurally quite easy to influence, and not because anyone is being dishonest about it. People respond, absolutely rationally, to how they’re being measured. If pick rate is the headline KPI that they’ve been told they need to pay attention to, well, any easy-zone orders will obviously get prioritised, because that’s what the number that’s being measured is rewarding. Travel time, for example, which often accounts for at least 50% of total picking time, can get buried inside an average that no one ever has the time to break down properly. And none of this is anybody’s fault, but it does mean that the headline number can sometimes drift further and further away from what it’s supposed to be telling you.

The industry average for manual picking sits at approximately 71 items per hour. Best-in-class operations can achieve 250 picks or more. Which is a significant gap, and can be distracting. But the more useful question is how much of any reported number is a true, and meaningful measurement, and how much of it is a function of which orders happened to end up in the mix on any given day.

For example, a picker who is hitting 120 an hour in an easy zone and 40 an hour in bulk will be reported as 80. And that number, 80, tells you very little about what’s actually happening on your floor.

A better measure: a pick rate that’s been normalised by zone and order profile… combined with accuracy.

Or, if you want a single number that’s really worth chasing, you might want to focus on the share of picker time that’s actually spent picking versus how much time is spent walking. And that’s where Pulse from Optioryx, the AI pick-path optimisation (that runs directly on Dispatcher WMS), has the most impact. It can reorder pick sequences in real time, and it’s based on what’s in the warehouse today, rather than what was in the warehouse when the putaway algorithms were last reviewed. Because of it, improvements can often show up in the part of the operation that your headline pick rate has been accidentally hiding.

So what?

Well… none of these numbers are bad.

They all measure something. The problem is that as soon as they stop being a starting point for inquiry, they can start to become an end point for it instead. And that can mean that as soon as your specific KPI goes green, any conversation around it may stop. Which in turn might mean that the operations underneath it will just carry on doing things that you might actually want to know about.

The data to do things better is almost always already inside your WMS. And if you’re a Dispatcher WMS user, it certainly is. But there can be bottlenecks to getting exactly what you want, exactly when you want it. Who can see what? In what level of detail? And how quickly can they follow up the answer with a sensible next question? Sometimes those things can create a gap.

But it’s a gap that USP, AskUSP and Pulse (on top of Dispatcher WMS), can help to reduce, and instead, turn that same underlying data into a more useful picture of what’s actually going on.

And we’d respectfully suggest that it’s worth doing sooner rather than later. According to Logistics UK’s Q1 2025 Skills and Employment Update, the advertised salaries for UK warehouse operatives increased by 49% year-on-year. The figure will have changed again since that stat was published, but the direction is pretty clear: every misleading or incomplete labour KPI you are looking at may well be costing you considerably more than it was a couple of years ago.

And that pressure won’t be going away. Ever.

So, it’s worth asking about every metric on your dashboard: can this number move without anything in the operation actually getting better? Plenty of them can. And that’s a useful thing to be aware of. And certainly before the next time that someone presents you with a deck full of green numbers and tells you that everything is fine (because they believe that it is).

If you’d like a hand finding out what your dashboard is actually telling you, or a personalised demo of how things COULD look, we’re always happy to have a chat. Book an obligation-free discovery call now.