A WMS is good at a specific job. It knows where things are, it tells people what to do next in an order that makes sense, and it records what happened in a way you can audit later.
That covers a lot of warehouse pain. It does not cover all of it, and the difference matters because a WMS costs six figures and eight months of your team's attention. Spending that on a problem the software cannot reach is how projects end with everyone quietly going back to the old spreadsheet.
Here is my working list of problems that follow you into the new system.
A layout that fights your order profile
If your fastest fifty SKUs are scattered across four aisles, your pickers walk. A WMS will route them through that walk more intelligently, and the honest gain is maybe ten or fifteen percent on travel.
Moving those SKUs to a single fast-pick zone near the pack stations gets you far more, and you can do it in a weekend with a forklift and a marker.
I have watched a company sign a WMS contract to fix picking productivity when their real issue was that the product had been slotted in receipt order since 2011. Slot first. Then buy software to keep the slotting current, which is a genuinely good reason to buy one.
Not enough people, or the wrong ones on the wrong shift
A WMS makes each person more productive. It does not make more people.
If you are missing the cut because you have five pickers doing a six picker job, directed picking will buy back some of that gap and then you will be back where you started at the next volume step. Look at labor by hour of day before you look at software. A lot of warehouses are not short staffed, they are staffed flat against a demand curve that spikes between 1pm and 4pm when the day's orders release.
Staggering start times is free. It also tends to reveal whether you had a labor problem at all.
Demand you cannot see coming
Stockouts and expedites usually trace back to planning, not to the warehouse. A WMS tells you what you have and where it sits. It does not tell you what you will need in six weeks, and it will not reorder for you.
That is your ERP's job, or your planner's, and if the reorder points have not been touched since the system went in, the warehouse is being asked to solve a math problem it does not have the inputs for.
There is a version of this that goes the other way too. Some companies discover during a WMS project that their on-hand numbers were wrong enough to break planning, and fixing accuracy fixes the stockouts. That is a real win. It is also a win you can get from cycle counting for a fraction of the price.
Data that is wrong in your ERP
This one is the most common and the most expensive to learn late.
If your item master has units of measure that do not match how product actually arrives, or dimensions and weights that were guessed at, or three part numbers for the same physical item, a WMS will faithfully act on all of it. Directed putaway will send a pallet to a bin that cannot hold it because the cube data says it fits. Cartonization will pick a box that will not close.
The software does not clean the data. The project does, and that cleanup happens on your side of the line. Budget for it as its own workstream rather than a task inside week three.
Processes nobody follows now
If receiving skips the count when the truck is late, or if pickers grab from the reserve rack when the pick face is empty and tell nobody, those habits exist for a reason. The reason is usually that the official process is slower than the pressure people are under.
A WMS makes those shortcuts harder, which sounds like a fix. What actually happens is that the shortcut moves. People will scan the wrong item to close a task, or leave the handheld on a desk and do the work from memory, and now you have a system that is confidently wrong instead of a spreadsheet that is obviously wrong.
Fix the reason first. If receiving cannot count under the current dock schedule, more software does not create dock time.
The honest counterpoint
I do not want to talk anyone out of a system they need. There is a real threshold, and past it manual methods stop scaling no matter how good your process is.
The signals I trust: you are running multiple pickers in the same zone and they collide, you cannot answer where a lot or serial went without a phone call, your accuracy holds only because one long-tenured person knows where everything really is, or you have added a second building. Those are structural. Process work makes them less bad and does not make them go away.
A cheap way to find out which one you have
Pick your three worst recurring problems from the last quarter. For each one, write down the specific moment it goes wrong, who is standing there, and what they are looking at.
Then ask whether a screen telling that person what to do next would have changed the outcome.
Late shipments because picking finished at 4:30 and the truck left at 4:00? A screen helps. Late shipments because the item was not in the building? A screen does not help.
Do that for a week's worth of problems and the split becomes obvious. In most warehouses I walk, somewhere between a third and half of the pain is reachable by software, and the rest is layout, staffing, planning, and data.
Knowing the ratio before you sign is worth more than any demo you will sit through. It tells you what to fix now for free, and it keeps you from judging a WMS by whether it solved problems it was never able to touch.