Forget the bar code. In a few years, it may be replaced by networks capable of following products from factory to warehouse to supermarket shelf. The mastermind behind this system -- called radio frequency identification (RFID) -- is Sanjay Sarma, a professor and research director of Massachusetts Institute of Technology's Auto-ID Center.
With the help of dozens of companies and five other universities, he has cobbled together a network that hinges on all pallets, cases, and individual items -- whether packs of razor blades or bottles of shampoo -- carrying their own RFID tag. He recently spoke with BusinessWeek Boston Correspondent Faith Arner about the struggle to make the new system work without breaking the bank.
Q: How did a roboticist get into RFID?
A: Quite simply I got into auto ID because of my colleague David Brock. One day he came to me and said, "Why don't we just put tags on everything and instead of having the robot recognize with vision? Why not just read a bar code or an RFID tag?" That's how we got into it, around 1998.
Q: Were you working on tags already?
A: I had looked at tags maybe eight years earlier [to automate machine tools and production lines]. It was very fleeting, and I gave it up because it was just so expensive. Then Dave came back and said "Why not just put it on everything, put a number on it, and just bypass this whole robotics thing?" He was thinking of it more in terms of put a number on it and put the data on the network instead of recognizing it through image-capture.
Then the challenge was to make the tag cheap. You make the chip small, then how do you handle the chip? Well, could we not invent new ways to handle chips? That's how the Auto-ID Center took shape.
Q: How did you hook up with Kevin Ashton, the executive director of Auto-ID?
A: In 1999, I bumped into Kevin. He was a brand manager for Procter & Gamble (PG) trying to figure out how to keep bottles of Oil of Olay on the shelf and if it was possible to make RFID tags cheap. He had heard it was impossible to do so from the RFID community.
So I told him, "Listen, I think we've figured out a way to make it very cheap. And I know it has to have an application in the supply chain." And he said yes, but not just the supply chain, in retail. And then we started looking at the numbers, and we said, "You know, in retail, when you're talking about billion of items, how cheap does it need to be?" He said less than 5 cents. And I said "I think we can get it there.'"
Q: When you entered this field, did you think you would be starting revolutions in supply-chain management?
A: The first dreams you have are always out there. I look back and say reality turned out to be almost as spectacular as our dreams about where the technology would end up. Kevin said you need to put these things on bottles of Oil of Olay. That hasn't happened yet, but today we're looking at it and see there's a pathway.
Q: Did the path require a lot of bushwhacking?
A: About 2 1/2 years ago, we were tying to convince someone to join the center. This guy says "You're telling me you can make a 5 cent tag?" I said yes. And then I didn't sleep for about two weeks. But at the end of two weeks, I had pretty much figured out the key innovations, inventions you would have to make, technology we would have to master, companies we would have to get to cooperate, the people who would have to buy it. We pretty much laid out a path.
So you wrestle with what it is people want, think about what's possible, and you push the limits of possibility to meet what they want. Very often, I'd agree to something and go back and say, "I don't know why I agreed to that." But we've been able to meet most of those goals.
Q: What's your biggest accomplishment? How much is a tag now?
A: It's still fairly expensive. The biggest accomplishment is the system. A consortium of 100 companies and six major universities put together a system, which on the whole stands. But take away any one piece and like a house of cards, it would fall over.
I'm absolutely confident that with volumes of 5 billion to 10 billion a year, which is not a lot, tags can be made for 5 cents. People were laughing at us when Kevin first said this. But almost everything we said has been borne out.
Q: Was Wal-Mart (WMT) pulling out of trial with Gillette (G) a big setback?
A: No, I think that has been spun the wrong way. I think it's a case of the glass being three-fourths full rather than one-fourth empty. The reasons I say that is that 40 companies have been enrolled in the field trail. And it was targeting three things: pallets, cases, and items.
In my wildest dream I didn't think items would be announced by 2005. In my less wild dreams I didn't think cases would be adopted as early as they have been. I know the items didn't happen after there was some news about privacy advocates. So people have juxtaposed the two. But I don't see it that way at all.
Q: So it's not about the privacy backlash?
A: It might be a distraction. You don't want to get involved in an argument right now, because it's moot. It's a distraction, and perhaps it played a role from that point of view. But it wasn't really a structural change. I don't think there's any fundamental change in the sponsors' interest in this technology. If anything, it has sharpened.
Q: When will we start to see tags on items?
A: I think we're still two or three years away. But I think there will be spots, like pharmaceuticals, where you might see this happen earlier. Counterfeiting is a huge issue. In pharmaceuticals, it can kill people. So RFID can detect counterfeits.
Remember the Tylenol cyanide thing? Imagine if RFID had been around then. I bet you in a year the pharmaceutical industry will adopt RFID. Because you'll know exactly which part of the supply chain is compromised.
Q: Where do you think the biggest payoff is?
A: I think the shelf is the biggest payoff. Capitalist economies are based on supply meets demand. So if a consumer goes to a store and finds the shelf empty, it's a fundamental failure of supply meets demand and probably worth 5% to 10% of total sales of the company.
The [other payoffs] are smaller but also huge. Tightening up the supply chain is another big application. Smaller than the shelf, but by itself it more than justifies RFID.
Q: What are you working on now?
A: The Auto-ID Center is transitioning into Auto ID labs, and the Uniform Code Council has set up a new nonprofit fully owned subsidiary called AutoID Inc., which is taking over all the standards-making activities. So I'm in middle of transitioning that over. We'll then return to labs and do more fundamental research, and so it's the denouement. It's a winding down on one hand, but going back into a more academic role on the other.
Q: Does this mean you have reached a plateau and it'll take a while for next big leap?
A: In a good way, that's true. The Auto-ID industry was at a certain stage, and we wanted to make this giant leap, and we made the leap. So yes, you could say we've reached the first step, and there's a plateau aspect to it. Which is, now companies are adopting it, Wal-Mart is gearing up to use it. So the innovation is still happening, but the technology transfer now will slow down for some time, and then we'll make another leap.
Q: What do you think that next leap will be?
A: Probably protocols that give you even better performance. Things you can do with tags so you can personalize them. Now you kill the tag if you want privacy. In future, you can personalize them so only you can read your tags. You can take them home, do stuff with them.
For example, you put medicines in the cabinet. If the cabinet figures out this medicine shouldn't be taken with another, it tells you. Or if your frig runs out of milk, it orders it for you. This is way beyond retail.
Q: Will chickens ever be able to tell the oven how to cook them?
A: It's a very visual example. I don't think it's the most compelling one.
Q: What's more compelling?
A: I think medicine is huge. First, we have a retirement boom coming up. Medical care will have to be largely automated. Second, we have the genome-sequencing advances. So all this makes it more likely that medicines will be tailored to particular people. The problem of that is the logistics are terrible. Not only are there manufacturing challenges but also distribution challenges. And so it becomes a supply-chain issue.
RFID tags will play a big part there. As we start inventing thousands of new medicines, we'll start discovering thousands of new interactions, and we'll need to tell people about them. We'll need much more intuitive interfaces. And the medicine cabinet is a great interface. That's going to be a very big deal.