In today's food industry, driven by the quest for supply chain efficiency and the demand to target customers more effectively, it would be difficult to overemphasize the importance of generating accurate scan data.frequent shopper clubs.For some chains, the goal of achieving highly accurate scanning levels already has been reached. But many retailers, especially smaller independent operators, are

In today's food industry, driven by the quest for supply chain efficiency and the demand to target customers more effectively, it would be difficult to overemphasize the importance of generating accurate scan data.

frequent shopper clubs.

For some chains, the goal of achieving highly accurate scanning levels already has been reached. But many retailers, especially smaller independent operators, are just beginning the struggle to get up to speed in this area.

For everyone, though, the issue is clear. To compete effectively in the information-driven supermarket industry today requires having precise data on what products are being purchased, when, and who is buying them.

Here is how SN's discussion with five leading MIS executives went on the subject:

SN: How is the industry doing in obtaining accurate scan data?

RAY HAMILTON: I've grown up with scanning here. It's been the bible for us. If we didn't do it [properly], we didn't keep our job. It's that important. Those companies that have not already pursued scan data accuracy for the benefit of their customers need to do so today if they plan to participate in ECR and computer-assisted ordering. PATRICK STEELE: The issue of scan data accuracy is now being re-emphasized in our industry. It's amazing we need this issue to be brought back to the forefront of everyone's attention, because having a product ring up at the correct price at the front end of the store should be the No. 1 issue for all of us. We all have to make sure our customers are getting what they paid for and are being treated fairly at the front end.

DAVID HAYES: We should be doing a lot better in the area of scan data accuracy. The data we're getting may be statistically correct for signaling trends, but it's not necessarily numerically correct. We've gone many years without using this data for reordering, for example, because it hasn't been accurate enough.

BILL MAY: Only about 47% of our stores scan right now, so we have a raft of stores we work with that don't even scan yet. But scanning is absolutely critical to the future. If we want to get into programs such as continuous replenishment and category management, we need accurate, crisp scan data. That is a huge issue right now.

PETER ROLANDELLI: We're in very good shape at the front end in terms of scanning. We're able consistently to get the information we need from our stores. I was just looking at a report that showed we had 99% data retrieval for the past 50 weeks.

SN: Why is accurate scan data especially crucial today?

MAY: In my mind, it all starts with scan data. If we don't have correct data on what is moving through the stores, then anything we try to do will be jeopardized. As we go forward, if the 50%-plus of our stores that aren't scanning don't get involved and do it right, then their costs of doing business are going to be phenomenal.

As ECR moves along, it's going to become an even bigger issue. If I'm doing business with a retailer and need to send an order book of 500 pages each week to show what products are available, and it costs me $50 in paper costs to do that vs. 3 cents to transmit it electronically, then the retailer requiring the paper version is going to have to pay $50, and the retailer equipped to receive it electronically will pay 3 cents. But I don't see how any retailer can remain competitive doing business that way.

ECR is going to drive a lot of these points home. It already is and will continue to do so. I think the number of stores doing scanning will increase dramatically over the next several years.

HAYES: The whole ECR push now requires accurate data. Scan data needs to be very accurate for ECR programs such as computer-assisted ordering. There's still a long way to go in this area, especially in terms of better discipline in capturing information at the store level. The discipline is there at the headquarters level to deal with these issues, but it's not there at the store level. The smaller the store, the more likely there are to be problems.

Customer-specific accuracy is critical when it comes to areas such as electronic marketing and niche marketing. It's necessary to know what specific customers buy to make decisions when it comes to frequent shopper programs. STEELE: Scan accuracy is going to play an increasingly important role in the future, because it's very hard to drive programs such as computer-assisted ordering and category management if accurate data about what's selling in the store and when it's selling isn't available.

SN: What is involved in generating accurate scan data?

HAMILTON: One good way to help ensure accuracy is to remove the department keys from the cash registers: If it doesn't scan, you can't sell it. To have complete accuracy requires vigilance regarding cashiers, who face front-end pressures to maintain a certain speed in terms of rings per minute or sales per hour.

A cashier faced with 24 flavors of baby food all at the same price may feel pressured to use the 24-count key and scan just one item. But that makes it difficult to do computer assisted ordering. Those situations and pressures at the front-end have to be alleviated, perhaps by removing the quantity key so that the cashier is forced to scan every item.

If an item doesn't scan, we have a policy that requires cashiers to do what we call a validation. In many ways, the specialty foods business we're in is a lot tougher in terms of scanning accuracy than most supermarkets. We import a lot of items from obscure places or from smaller vendors. We launched a campaign when I first arrived here to eliminate nonscanned product. We're now up to a 99.9% rate for accurate product scanning.

STEELE: Part of the issue here is technology. With some of the older scanning systems, it's not as easy to bring back data on a daily basis as it is to retrieve it on a weekly basis. Also, there are issues about storage and file size capabilities. So there are some technology difficulties in obtaining proper scan movement. But it's certainly much easier with the new PC-based systems out there today.

The other issue involves how products are scanned at the store. The classic use of the quantity key at the front end, for example, can cause problems. Changes may have to be made in how products are checked out at the front end so each item is recorded. Are we accurately recording the individual sales of each product at the POS and bringing reliable data back from the stores to feed all the merchandising systems being developed?

HAYES: To ensure scan data accuracy, it's critical that retailers do a better job at the front-end checkout. Within five years, every store will either be scanning or go out of business. But there is still a preponderance of one- or two-lane small stores today not scanning. To survive against the chains, these stores will have to get up to speed and reap the benefits of scanning and electronic marketing.

MAY: At the store level, we're still struggling with issues such as making sure that the shelf tag matches up with the price in the POS system and the price in the computer in the back room. This whole area needs a lot of work and discipline right now. ROLANDELLI: It used to be that when we brought back weekly POS or item-level data, we would miss 100 stores a week because hardware was down or files were missing. Now we get every file every week. It's to the point where we can bank on it. There are still other areas to work on, and we can't get complacent, but we are making progress.

The data we're getting now is highly accurate. We're at the short straws of cleaning this stuff up, and we have to be if we're going to move into computerized ordering in a big way. The fact that many retailers are already going to computerized ordering programs means that they have virtually licked the problem.