Inventory

The real reason your stock count never comes out right (hint: it isn’t your warehouse staff) — and the one thing to look at on your last count that tells you which problem you’ve actually got

By Jay at Footing Tech — I put small and mid-sized businesses onto Microsoft Dynamics 365 Business Central at Clear Horizon Solutions. Which is to say I have a commercial interest in this subject, and I’ll come back to that before I ask you for anything.

The customer rang about the order that shipped short.

You didn’t send anybody out to check it. You went yourself, which should have told you something — the screen said eleven, and you have been doing this long enough to know what eleven of them looks like sitting on a shelf.

There were four.

And you stood there slightly longer than you needed to, because the arithmetic was not the problem. The problem was the sentence that arrived while you were looking at the gap.

Then you did what almost everybody does next. You went and found your warehouse manager.

It is never just the one order

That one you could live with. The customer was decent about it, which was somehow worse, and you sent the balance on Thursday and knocked the freight off.

But it is not one order, and you know it is not, because you have been keeping a private tally for a while now.

There is the pallet you bought in March that you already owned — two rows over, under a dust sheet, behind the thing nobody has moved since the extension went in. There is the item you are apparently out of on the screen, which your foreman can walk to blindfolded. There is the run you couldn’t quote on because you weren’t confident enough in your own numbers to promise a date, so you let it go and told yourself it was a thin margin anyway.

And there is the money. Not a figure you’ve ever written down, because writing it down would make it real — but you know roughly what is standing in that building doing nothing, and you know that a good chunk of it was bought twice.

Then there is the count itself. A Saturday. Four people, one of them you. Somebody brings pizza at two. And at the end of it the sheet and the system disagree by an amount that is too big to write off and too small for anybody to find, and the fix goes in as an adjustment with a note that says count.

Which brings us to the part that actually bothers you, and the reason you are still reading an article about stock counts.

You have now had the same conversation with your warehouse manager four times. He is not lying to you. You have known him for years, you have watched him work, and you would trust him with the keys and the float. He is not lying, and the numbers are still wrong, and those two things are not supposed to both be true.

You have already tried the sensible things, and that is the confusing part

Run through what you’ve done, because I suspect it’s most of this list.

• You started counting more often. Cycle counts instead of the annual bloodbath. The variance got smaller for about six weeks, which felt like proof, and then it drifted back.
• You tightened receiving. Genuinely a good move. Goods get booked in the same day now, mostly. It helped. It did not close the gap.
• You had the conversation. Then you had it again. Then a third time, more directly than you wanted to, and each time things were better for a fortnight.
• You hired somebody to own inventory. And now you employ a person whose actual job, when you describe it honestly, is reconciling two numbers that will not agree. You did not fix the disagreement. You staffed it.
• You bought the scanners. Or the app, or the add-on. The wrong number now arrives considerably faster.

Every single item on that list is the correct response to a discipline problem. Not one of them is stupid. If a business has sloppy handling, that is exactly the sequence a competent owner runs, and it works.

Which ought to tell you something, if you’re in the mood to hear it. If the correct treatment for a discipline problem doesn’t work, the odds are you haven’t got a discipline problem.

Two completely different problems that look identical from where you stand

Every inventory discrepancy in the world is one of two things, and they present exactly the same way — a number on a screen that does not match a number on a shelf.

Discipline fixes the first one. Discipline cannot touch the second one, because there is nothing there for anybody to be disciplined about. You can hold a meeting about it every Monday for a year and the number will keep coming out wrong, and the only thing you will actually achieve is making a good employee feel accused.

Here is how invisible the second kind can be.

Your system tells you what is on the shelf. It is right. It has been right all week. What it does not tell you — because there is nowhere in it to hold the thought — is how much of that shelf is already promised to orders you have taken and not yet picked. So your salesperson opens the screen, sees eleven, and sells eleven.

Nobody miscounted. The eleven were genuinely there. Four of them had already been sold to somebody else that morning, and your system had no way to mention it.

Don’t believe me. The evidence is already in your building

This is the point in an article like this where you would normally get a customer story. A distributor in Ohio, a percentage, a photograph of somebody in a hi-vis vest looking pleased.

I’m not going to give you one, for two reasons. The first is that you have read enough of them to discount them automatically, and so have I. The second is better: you can test the whole of this yourself, today, without me.

Get your last count. Not the summary sheet — the line detail. Then:

1. Sort it by absolute variance and look at the top thirty lines.
2. Ask whether those thirty are a random cross-section of what you stock, or whether the same names keep appearing.
3. Then sort the same thirty by location, and ask the same question again.
4. Then mark which of them are kits, assemblies, returns, or things that moved between sites.

If it is genuinely scattered — different items, different places, no repeats — you have a handling problem and you now know it for free.

If the same twenty names have been at the top of that list for three counts running, you have just watched a system fail the same way three times, and no amount of talking to your staff is going to change what happens on the fourth.

That will take you about ten minutes and it does not require my permission, my software, or my opinion. It is the only kind of proof worth anything at this stage, because it is yours.

What you are probably thinking right now

• “We’re not big enough for this to be a system problem.”

  Size is the wrong test. The test is whether what you do is more complicated than what your system can describe. A twenty-person distributor holding stock in two places, selling in different units to how they buy, and assembling anything at all, has already outrun most small-business accounting software. A ninety-person firm with one location and one unit of measure may be perfectly well served. It is about shape, not headcount.

• “Our records are a mess. Wouldn’t any system struggle?”

  Mess is the normal starting condition and it is not disqualifying. But it does matter for the diagnosis, which is why the checks ask you to look at the pattern rather than the accuracy. A mess scatters. If your mess is neat — and some of them are unnervingly neat — that is the signal.

• “This is a long way of telling me to buy software.”

  It would be, except that the checks have an outcome that says don’t. The first of the three scoring bands tells you in as many words that your problem is process, that you should fix it with process, and that buying a system would mostly make your existing habits more expensive to keep. I’d rather you got that answer from me than paid somebody to discover it eight weeks into an implementation.

• “We already have a system.”

  Almost everybody does. The question is not whether you have one, it is whether the one you have can hold the things that are actually happening on your floor. Several of the checks are simply asking your existing system a question and watching whether it can answer.

• “I haven’t got an afternoon spare.”

  Four of the nine you can do sitting at your desk in the next ten minutes, and the one described in this article you can run off a count you have already paid for. Start there. If nothing looks off, you have your answer and you have lost ten minutes.

What this costs, and what I am actually asking you for

This is where an article like this usually reveals a price, generally with a line through a bigger one next to it.

You are not getting a price, because I am not selling you anything today. What I am asking for is an email address, and what you get for it is the nine checks — each one with the instruction, the answer you want, and what a wrong one is actually telling you — plus a scoring page that turns your nine results into one of three conclusions.

It opens on the next screen. Nothing arrives in your inbox that you have to go and dig out, because I would rather you read it while you are still annoyed about the shelf.

And to be straight with you about the far end of this: if your answers come back structural, then eventually somebody quotes you for a system that can hold your business properly, and it might be us. Our prices for that are published on our own website and you can go and read every line of them without speaking to a human being. But that is a conversation for after you know which of the two problems you have got. You do not need me for this part, and you shouldn’t pay anybody until you do.