Most WMS requirements documents are built to be answered yes to. That's a document problem rather than a vendor problem, and it decides how the next three years go.
Dave Crysler had me on episode 36 of Everyday Business Problems in October 2024. He asked what gets missed most in vendor selection, and I took the question somewhere he wasn't expecting.
Every quote block below is me, on the episode, condensed for reading.
The RFP Everyone Clones
The typical approach, you get into the RFP. And by and large, I see more clones pulled from the internet, or pulled from a certain vendor. If you look at them, there's hundreds of questions on there. They're pretty much rinse and repeat, and most of them focus on table stakes. The features. And it's very easy for a vendor to come in, select all the rows and say yes. There's just not a lot of teeth to them.
That happens without anyone being dishonest. Ask a hundred vendors whether they support lot management and a hundred of them support lot management. The question has a true answer, and the true answer carries no information. A document made of those questions produces a spreadsheet where everybody scores well and nothing separates anybody.
A requirements list that every vendor can answer yes to has told you nothing, and you paid for it in weeks of everyone's time.
I'm not neutral on this and I said so on the show. I'm frustrated with the status quo, I don't have the replacement perfected, and I'm still adjusting it every time we run one. But the RFP as a first move is the part I'm confident is wrong. It also arrives after the real filtering has happened, since the shortlist it gets sent to was mostly built out of familiarity.
What Replaces the Checklist
I'm a big fan of turning that on its head, where it's a small assortment of questions and capabilities that we focus on. The table stakes piece, we're at a point now where most vendors have that covered. You can vet that out pretty quick.
I like asking for a short canned demo to get a feel for how their platform works. I don't need feature demonstrations for the most part. I like that because you can do it asynchronously. You can get a few of these vendors together, get those demo videos, and then people can digest those if and when they need to.
Then the questions that actually sort the field. I've been around this long enough to know most vendors won't cover the critical few, so that's where the effort goes, because that's what rules people out. What comes out the other end isn't a scorecard:
We have a comprehensive trade-off matrix. Vendor A is 80% there. Vendor B is maybe 50% there, so there's a bigger lift to get you where you need to be.
The useful output of a selection is a priced gap for each finalist. Every one of them has one, and the only question is whether you found it before you signed or after.
Lot Management, All the Way Down
The clearest example of a requirement with no teeth is the one I keep returning to. A client, multiple sites:
They had the impression that when they asked for lot management and the vendor said yes, that meant true lot management in their mind, across all their sites. They received a SKU in with a lot, and that was preserved and validated across all the different sites. If they received one somewhere else, that information was preserved, and they didn't have anything weird. Because if you get a lot with a different expiration date going up for that SKU to an ERP, bad things happen.
The WMS itself didn't have that. It really just recorded the attributes, the data that came in. There's a lot number, there's an expiration date. It didn't do any validation beyond that.
Lot management as an attribute means the system writes down whatever you type. Lot management as a constraint means the system refuses to let the same lot of the same SKU carry two different expiry dates in two different buildings, and only the second one is worth paying for.
Work orders fail in exactly the same shape. Manufacturers see the checkbox and assume it means what it means to them, and for process or batch work with yield loss it doesn't, which lands you in front of leadership asking for either another system or an enhancement you'll be the only customer of. Both of those are the sort of thing requirements written from your own floor surface before a contract rather than after one.
Nobody Is Finished at Go-Live
I think it's a misnomer that you roll out a warehouse management system and you're done. It's not. You look at your operation, it's never done. Continuous improvement is part of what you do. Same thing with the system. It's just another lever to help move the business forward.
It's fairly common for folks to roll something out and not use all the bells and whistles. Going back and looking at that and reevaluating periodically, either going to the vendor or finding someone that's independent, there's always room for improvement. That might even be a better option than just looking to an upgrade. That's what the vendor is going to point you to, because there's money in that.
The line that belongs in the original budget is the one covering the years after go-live. Customers come and go, new ones bring new requirements, regulations change. A number that covers subscription and implementation and stops is a number for getting started.
The implementation quote describes the happy path from the vendor's side of the table, and it silently assumes a maturity on your side that nobody wrote down.
It's also why the small annoyances matter so much beforehand. A workaround you tolerate today becomes something you fight daily once the system is enforcing rules around it, and that's the mechanism behind workarounds that keep multiplying.
What's Changed Since October 2024
The trade-off matrix I describe here is now a named deliverable. The System Fit Sprint produces three documents on a fixed schedule, deliverables landing two weeks after the final onsite, and the vendor analysis is the ranked version of what I was doing ad hoc on this episode.
Publishing the method as a fixed scope fixed what frustrated me most on this show. A buyer could not see what they were paying for until they had paid for it.
