Gilmore: Isn't there some element of truth in that idea though? I certainly don't hear hardly any stories of the kind of quality problem we use to see say in the 1990s that would cause real operational problems, bar codes that just wouldn't scan, etc.

Hill: No question bar code scanner and printer providers are producing much better products than they did back then. But don't think there aren't still occasional problems, and that's often again because users just take it for granted that the auto ID will work without doing enough homework on what is needed to make it work.

Just consider: is 99.5% success good enough when you are processing say hundreds of thousands of packages or cartons per day? No.

Gilmore: So your message here is that even though the technology has become much better, that taking it too much for granted is a mistake.

Hill: Right. Auto ID systems still need attention.

Gilmore: So let's stay on topic of auto ID. When we talked in 2003, the first EPC conference was just happening, Walmart had announced its mandate, and it looked like RFID was going to take over the world. Then the Walmart program imploded, and while RFID is being used in many places, in the US at least it is not much used in distribution. Is that going to change, do you think, or did RFID just miss the wave in distribution and it's just not going to happen?

Hill: The Walmart announcement was a watershed event, and cost a number of companies a lot of money in terms of preparing for those mandates. What we wound up with unfortunately were cost and performance issues that I might compare to "square peg-round hole." The technology was not ready at a price point that could be cost justified for the applications Walmart was trying to pursue.

You know I spent part of my career in RFID. In the 2000s, RFID could be made to work in those Walmart applications, but you'd never pay it back.

Gilmore: But that is a key reason why in the end the Walmart program collapsed.

Hill: Exactly. But much progress in the technology is being made. They are doing things with RFID today that are far beyond what was possible at the time of the Walmart announcement.

Gilmore: That's in part because so many venture capitalists poured hundreds of millions or billions of dollars into RFID companies after the Walmart announcement, and most of them lost their shirts over it, but it did move the technology along even if they turned out to be dumb investments. But still, not much activity in distribution.

Hill: There is obviously a lot of action in the apparel sector for item-level tagging. But they are doing it where they can find the ROI. Eventually that will drive tracking not just in store but throughout the supply chain. It will come.

Gilmore: Here's my theory John. If you look at bar code, it really started at the item level - the UPC code - and once that was in place in consumer goods to retail then eventually it led to case-level identification with the UCC128 label, and use in distribution. I think the same thing will happen with RFID. Once item-level tracking, starting in soft goods, is broadly in place, I think use of the tags will move back up into distribution.

Hill: I think you're dead right, and also what is driving it is ecommerce and efulfillment from stores. To do that requires high levels of inventory accuracy, which is what RFID can give you, in real time. In general though, when RFID is considered intelligently and dispassionately, married when it makes sense with bar coding and Voice technology and others coming along such as smart glasses, augmented reality, smart phones, even something called the HoloLens from Microsoft - I think some very powerful capabilities can be created, and that RFID will become an important part of the supply chain for many companies.

But I think it will usually be part of a team not the only solution.

Gilmore: Let's go back to WMS. My perception is that the implementations are too often too rocky and painful, and that in 2003 we would certainly have thought that would have changed dramatically by 2015, but it hadn't. Do you agree, and if Yes is there anything companies deploying WMS and WMS vendors themselves do to improve that situation?

Hill: First, I do agree with you.

Let's start with end users first. In my experience, supply chain projects fail when users didn't fully do their homework, or skipped important steps in the process, whether its failing to well define workflows and processes, performance goals, all the needed capabilities, etc.

But even when they've done all that, some companies still don't well connect all those requirement to what's in the different WMS packages they are evaluating. It's hard work, but it has to be done correctly, and if it's not in can lead to the kind of deployment pain you referenced.

We see that still today. When we sit down with a prospective WMS user, and stress the importance of understanding current workflows and desired workflows, and expected end results, a surprising number of even large companies want to shortchange some of that, and get right to selecting a system.

In fact, I recently had to meet with a company looking for a WMS to reinforce the need to go about it the right way. And again, this isn't an anomaly, and is a key factor in implementation issues.

The other issue with end users is a failure to build clear ownership and commitment of the WMS project from day 1. You simply can't spring a new system on workers and managers without their participation in the process, from requirements definition all the way through the system launch.

When companies fail to do that engagement - I saw one where this had happened earlier this year - things are bound to hit a rocky road, or even fail.

Gilmore: To me, you don't see the outright failures like you sometimes did in the old days so much anymore, but it's still just too painful too often, and I think the vendors also have to take some responsibility for that.

Hill: As the WMS market has grown and matured, I think frankly that the level of talent on the WMS vendors' front lines has diminished. And a number of them are subcontracting implementation to third parties. There are some good third parties out there, but some of them don't know a warehouse from an outhouse.

WMS vendors need to up their games in terms of who they hire and how they train them for these deployments. You need people who understand distribution, not just how to configure the WMS product. And that includes the front end, the salesforce, who I think need to have stronger knowledge of both distribution and the product they are selling.

I've received some sale pitches over the past few years where you wonder if they've ever walked a warehouse.

Gilmore: It's interesting that back in the highly customized WMS days, you needed to have very knowledgeable people up front, because you had to precisely define what code was to be written. Now here in the much more packaged WMS era, the idea sometimes is you don't need those kind of experts any more.

Hill: I think you are correct. When I look back 30 years ago, when we started turning Logisticon around is when I replaced all the good looking, well dressed sale staff who didn't know much at all about distribution with engineers. When I put them on the front lines, shortly thereafter our fortunes started to change.

Any reaction to our interview with John Hill? Let us know your thoughts at the Feedback button (email) or section (web form) below.