Warehouse picking: where the time goes and what to change first

I once spent a morning on the floor of a prospect's warehouse. No racking anywhere: a lot of one kind of stock, stacked where it had ended up. I asked four questions — how many steps do your people walk in a day, how are products located around the building, do you work in zones, and do you store by ABC. Nobody had counted the steps. Products sat where there had been room. No zones, no ABC.

From those four answers one fact followed: every order picker was walking the whole warehouse for every single order. They were doing twenty thousand steps a day. A workout at work — except that the company was paying for it by the hour and getting nothing back for it. By the time we had finished talking, the owner had worked out that he had been running the building that way for fifteen years.

That building did not need racking. It needed to know where everything was, and to be able to direct a picker to the next location. Zones, ABC storage, picking several orders in one round, a consolidation area where they get sorted: every one of those is a decision rather than a purchase — and not one of them survives on paper. Something has to hold the picture and direct the work.

Most warehouse improvement starts somewhere else entirely. It starts with whatever is annoying people this week — a jammed conveyor, a printer that keeps running out, a supplier who delivers badly stacked pallets. All of it is worth fixing. None of it is where your money is going.

If you want a warehouse to get meaningfully better, you have to put your effort where the hours actually are. In almost every warehouse I have walked through, that is order picking. And inside picking, it is walking.

That single idea is the whole of this guide. Everything below is a way of acting on it, and the four questions I asked that morning are in section 8, for you to ask in your own building.

Take all the labour in a warehouse — receiving, putting away, replenishing, picking, packing, dispatch, counting stock — and roughly half of it goes into order picking. The other half is shared between everything else.

Then look inside picking. Roughly half of a picker's time is spent walking. Not reaching, not scanning, not reading a label, not putting something in a box. Walking from one location to the next.

Put those two together and you get the number that should drive your decisions: about a quarter of all the labour in your warehouse is someone walking. That is usually the single largest activity in the building, and it produces nothing. Nobody pays you for metres walked.

Both of those halves are rules of thumb out of my own years on warehouse floors rather than market figures. Treat them as the order of magnitude to expect, and then measure your own building — which is what the rest of this section is for.

This is why the arithmetic of improvement is so unforgiving. A thirty per cent improvement in something that accounts for five per cent of your hours gives you one and a half per cent. A ten per cent cut in walking gives you two and a half per cent. Two-thirds more, for less effort. Work on the big number first.

Before you change anything, measure. Walk a full shift with a picker, with a stopwatch or just a pedometer. Count the metres, the minutes standing still, and the times they double back for something they passed five minutes earlier. You will learn more in one shift than in a week of reports, and you will have a baseline to judge your changes against.

Before any of the rest, the basics have to be in place. You cannot route a picker sensibly through a building where the stock figure is a guess and half the locations have no label. Every improvement in this guide — layout, slotting, how you steer your orders, which picking method you choose — quietly assumes the opposite, which is why this section comes second rather than last.

It is also the least interesting part of the job, and that is exactly why it gets skipped.

Get this right and everything after it compounds. Skip it and every clever thing you try afterwards will underperform, for reasons that will look like bad luck.

Layout is the cheapest lever you have, because it is mostly a decision rather than a purchase. Three things matter more than the rest.

Get rid of dead ends. If an aisle is closed at one end, a picker who needs one article halfway down walks in, picks, turns round and walks back out. You have paid for that distance twice and got one line for it. Open the ends, or cut a cross aisle through the middle of a long block, so every aisle can be walked through rather than entered and abandoned. A cross aisle costs you a handful of locations. It buys you the ability to leave an aisle in the direction you were already heading.

Make aisles wide enough to pass. When two pickers meet in an aisle that only fits one trolley, both of them stop. One reverses. Neither is picking. It feels thrifty to squeeze the aisles and win another row of racking, and in a slow warehouse it is. In a busy one you have bought storage and sold throughput. Decide which of the two you are short of before you set the width.

Run the racking along the length of the building, not across it. This one is counter-intuitive until you draw it. Racking placed along the long axis gives you fewer, longer aisles. Racking placed across the short axis gives you many short aisles — and every extra aisle is two more aisle ends, two more turns, and another decision for the picker about whether this aisle is worth entering. Longer aisles also give you more usable locations for the same floor area, because you lose less space to aisle ends and cross traffic. Straighter routes, more locations, same building.

The drawing on this page shows the same five picks in the same hall, racked both ways: about thirty per cent less walking and five fewer turns when the racking runs along the length. That is an illustration rather than a measurement of your building, but it is the right order of magnitude to expect.

One more thing anchors all of it: where packing and dispatch sit. Every route in the warehouse starts or ends there. If dispatch is in a corner for historical reasons, and your fast movers are at the far end because that is where there was space, you have built a long walk into every single order. Fix that before you fix anything clever.

This section is about the warehouse you already have. If you are fitting out a new building, or deciding on height, column grid and dock count, those are a different set of decisions and part three of this series covers them.

Once the shape of the building is settled, the next question is which product goes where. This is the part most warehouses have never done deliberately. Products sit where there was room on the day they first arrived.

Fast movers closest to dispatch. Pull your pick frequency per article over the last three months and sort it. You will find a small group of articles carrying a large share of the picks, a middle group, and a long tail that barely moves. Put the first group nearest to packing and dispatch, the middle group behind them, and the tail at the far end or up high. Sorting products into those three groups by how often they are picked is called ABC analysis.

Things ordered together, stored together. Look at which articles keep appearing on the same order. If two products are almost always bought as a pair, storing them at opposite ends of the building means you walk the length of the warehouse on every one of those orders. Put them side by side and that walk disappears.

Heavy first, light last. Order the locations along the picking route so the heavy, sturdy items are collected at the start and the light, crushable ones at the end. The picker builds a stable pallet or carton as they walk, instead of repacking at the bench or writing off damaged stock. This costs nothing to arrange and removes a whole category of problem.

Spread the fast movers out. This is where a naive reading of ABC goes wrong. If you put every single fast mover in one bay next to dispatch, you have built a traffic jam. Four pickers now want the same two metres of aisle at the same time, and they queue. A few extra metres of walking costs far less than two pickers standing still. Spread the fast movers across several faces of the same near zone.

Give the exceptional articles exceptional treatment. Most assortments have a handful of articles that appear on nearly every order. Those deserve either several pick locations in different parts of the warehouse, so a picker can always grab one without a detour, or the single easiest position in the building — between waist and shoulder height, right at the front, no bending and no reaching. That easy-reach band is sometimes called the golden zone.

And then do it again. What sells changes with the season and with your range. A slotting plan that was right in March is wrong by October. Put a review in the calendar, at least twice a year. Deciding where each product lives, and revisiting it as demand shifts, is the job called slotting.

One caution before you start. Working out the ABC groups is a morning with a spreadsheet. Holding them — keeping the classification current as demand moves, sending a picker to the right one of several locations, telling anyone where an article now lives — is not, and section 7 is about what that takes.

The cheapest pick is the one you never have to do. That makes your order intake a warehouse lever, and almost nobody treats it as one.

A customer ordering eight loose pieces creates eight picks, eight confirmations and a packing problem. The same customer ordering one full carton of twelve creates one movement. If the difference in price is small, many customers will take the carton — they simply do not know the option exists, or that it is cheaper for you.

So tell them. Put a discount on the full-carton quantity. Show the prompt in the webshop or the order portal: order four more and you have a full box, at a better unit price. Set a minimum order value if your small orders are losing money. Consolidate everything one customer orders in a day into one pick and one shipment instead of three.

What you get back is not only less picking. It is less handling, fewer confirmations, fewer chances to pick the wrong thing, less packing material, less damage in transit, and fewer parcels to pay for. One commercial decision quietly removes work from every step downstream of it — and it needs no racking, no system and no floor space. It needs your sales team and your price list.

There is no best picking method. There is a method that fits your order profile, and the profile is what you should be looking at: how many lines on a typical order, how much the same articles repeat across different orders, how big the building is, and what deadlines you are picking against.

These are not exclusive: plenty of warehouses batch-pick inside zones and release the whole thing in waves. Part two of this series sets out every method side by side, including two that sit outside this guide, with the order profile each one suits.

And every one of them rests on two things. The foundation in section 2, because a clever method running on inaccurate stock and unlabelled locations performs worse than one order at a time, the errors now being batched too. And something to group the orders, split the zones and consolidate the pieces afterwards, which is section 7.

Layout and slotting give you structure. They make the average walk shorter, and they do it permanently, which is exactly why they are worth the effort.

But they only get you a good average. The best possible route for the order in front of you right now depends on which lines are on it, where those articles happen to be stocked today, which other orders are open at this moment, and who else is already in that aisle. That answer changes every hour. No supervisor with a clipboard can recompute it, and no printed pick list can hold it.

This is also the point that the warehouse at the top of this guide had reached without knowing it. Zones, ABC storage, batch picking and a consolidation area were all available to that owner, and none of them was a purchase. But a zone nobody is directed into is just a line on a floor plan, and an ABC classification nobody maintains is last spring's spreadsheet. Three decisions in particular are beyond doing by hand at any volume: which orders to combine into one round, in what sequence to visit the locations, and which of several locations for the same article to send this particular picker to. Getting those right needs software that holds a live picture of where every article is and what every picker is doing — a warehouse management system, a WMS.

A zone nobody is directed into is just a line on a floor plan, and an ABC classification nobody maintains is last spring's spreadsheet.

System-directed work also takes the knowledge out of the longest-serving picker's head and puts it on a screen.

The gain is not only metres. A warehouse that routes its work properly is a more orderly place to work in. Work arrives in a steady flow instead of a scramble before the carrier comes. Pickers stop hunting for things, stop queueing for the same bay, and stop being blamed for errors that were built into the route. System-directed work also takes the knowledge out of the longest-serving picker's head and puts it on a screen: in our experience a temporary worker on directed work reaches eighty to ninety per cent of the standard pace inside a week. That is not a market figure either, but it is the difference between seasonal staff who work out and seasonal staff who do not.

That last point matters more than the productivity figure. In a labour market where warehouse staff are hard to find and harder to keep, a building that is not frustrating to work in is a commercial asset.

These are the four I asked on that floor, in the order I asked them. You can ask them in your own building this morning, and none of them needs a budget or a project.

1. 1
   How many steps does a picker walk in a day?Walk a full shift with one of them and time the walking separately from the picking. If nobody has ever counted, that is itself the answer.
2. 2
   How are products located around the warehouse?Ask who decided, and when. If things sit where there was room on the day they arrived, there is a slotting job waiting for you in section 4.
3. 3
   Do you work in zones?If one picker covers the whole building for one order, section 6 applies to you.
4. 4
   Do you store by ABC?Pull pick frequency per article for three months, take the top twenty, and go and find them. Note how far each one is from dispatch, and whether they are all crowded into one bay.

If the answers run: nobody has counted, products sit where there was room, no zones and no ABC — then every picker is walking the whole building for every order, and you have found your project.

Three more worth doing in the same week, because the four questions above do not reach them:

1. Follow one pallet from the dock to the shelf.

   Watch whether anyone scans it to a location on the way, or whether it simply goes where there is space. This is the step in section 2 that most often turns out to be missing.

2. Walk your own aisles, looking for dead ends.

   Count how many aisles can only be entered and left from the same end. Then stand in the narrowest one and ask whether two people with trolleys could pass each other.

3. Plot your order lines as a distribution, not an average.

   Count how many orders have one or two lines, and how many have ten or more. If both ends are heavy you have a two-peaked profile, and section 6 applies to you twice over.