GS1: 10 Percent Of Mobile Barcode Info Wrong
Written by Evan SchumanWhen GS1 U.S. and GS1 Canada on Tuesday (Nov. 16) formally unwrapped a coalition of many of the world’s largest consumer goods manufacturers to address accurate product data, it mentioned that “more than 10 percent of searches for information about allergens, nutritional characteristics or other data return incorrect or incomplete results.” There are reports of even higher percentages of problems with pricing data.
The issue behind those inaccuracies is actually the meteoric rise of mobile. It didn’t make much sense for the industry to hammer out mobile pricing and data consistency issues before mobile marketshare was meaningful. Now that it’s soaring, retailers suddenly have shoppers scanning information in the aisles that comes from a huge number of independent databases. All things considered, it’s impressive that the inaccuracy figures aren’t a lot worse.
The coalition is not dealing with pricing accuracy and is fully focused on product data, especially information such as allergen interactions, ingredients, expiration dates, product dimensions, recalls, instructions, etc.
It’s that kind of information that is accurate only about 90 percent of the time, GS1 U.S. said, citing an Auto-ID Lab report, prepared for the coalition. The group is called The Business to Consumer (B2C) Alliance and includes several major consumer goods manufacturers, including Procter & Gamble, Coca-Cola, PepsiCo, J.M. Smucker, Johnson & Johnson and Kraft Foods. One retailer—Kroger—is also a member and other retailers are also involved, although they have asked that their names not be made public, GS1 said.
The B2C group was actually formed back in March and was known internally as the Mobile Effort until June, said Gay Whitney, GS1 U.S.’s Senior VP for industry engagement.
The data inaccuracies stem from the way the data is maintained. If a consumer, for example, uses an iPhone to launch a barcode-scanning application such as RedLaser or ScanBuy, that application doesn’t do a lookup against the databases of either the manufacturer or the retailer, but instead uses a wide range of third-party sources, Whitney said.
On the pricing front, Whitney said her group is not focusing there right now. “Pricing is a very challenging issue. It’s very fluid, very dynamic, almost non-standardizable,” she said. “The search engines are working very hard in that space.”
Although the group is not focusing on that pricing, it did crop up at one of the group’s meetings this summer in Boston. While Whitney was presenting, some members were in the back of the room running an experiment. They were using a smartphone and scanning various products around the room, according to two participants in the meeting and confirmed by Whitney.
A bottle of Coca-Cola was scanned by an iPhone and it was identified as a blanket, said those participants, who added that a 2-liter bottle of Pepsi was then scanned and it returned no data. Whitney did not confirm the brands involved, as she was in the front of the room presenting at the time. One participant said that the unscientific sampling of consumer products scanned in the room that day had “the vast majority coming through (with) incorrect” information, including pricing.
As indicated by the group’s original working title, mobile is overwhelmingly making these database issues critical. But now that mobile is forcing the data to be updated and made consistent, it is likely to be used on the Web and leveraged with desktop devices as often—if not more often—than through mobile devices.
-Marc
November 18th, 2010 at 8:01 am
Why would this be a surprise to anyone?
When the data sync initiative was started many years ago, the majority of companies participating from the supplier side chose to utilize web portals and Excel spreadsheets to manually enter their data – the same process that led to the data quality problems that drove the data sync need in the first place.
This shows the retail industry valued getting suppliers to participate (ie. pay annual fees to be part of the data sync movement) more than they did the actual accuracy of the data. At least one leading retailer still needed to hire consultants to recheck product data for many suppliers for years afterwards (heck, they may still be doing so!) while many (most?) of the rest did the same with their own people and additional Product Information Management software – or just lived with the bad data.
Have all the retail suppliers adopted true data governance and master data management practices and properly synchronized both inside their businesses and with all their retail customers in the last year or two? If not, we’re probably dealing with more bad data rather than less. When it comes to data quality, institutional control problems are the norm.
Into this world where we try to fix errors caused by fat fingering with more fat fingering we overlay mass-consumer mobile initiative where consumers are fed data based on randomly bad data gathered from multiple sources and we expect that all should be right? 10% is amazingly good!
With only 10% inaccuracy, perhaps the bad data is being neutralized by crowd sourcing-like models, but that assumes that the accurate data actually exists in the first place and that it is somewhere where it can be aggregated from. And that it is in the majority during aggregation. If not, I hope you are cold and not thirsty when you reach for that Coke, er…blanket.
November 18th, 2010 at 11:56 am
Thanks for the interesting insights. If more and more consumers are going to be using these barcode scanning apps from their smart phones, then this information needs to be as accurate as possible or else there will be an increased mistrust of business by consumers. That’s not good, especially if it’s not the fault of the business!
November 18th, 2010 at 10:35 pm
Won’t any casual shopper just trying a phone scan assume that something is wrong?
Possibly that something is mislabeled?
Where they will blame the retailer, probably, when the $12 wastebasket shows as a $199 Waste Basket System at the checkout?
Just asking…
R
November 20th, 2010 at 11:06 pm
Last year I scanned a small package of Oreo cookies and the best price that came up was from Ace Hardware at $22.00. Obviously a mismatch.
This problem has been on-going for quite some time. Nice to see that it might finally get addressed. Funny though, that the problems occur with the bar codes of major manufacturers, all of whom belong to GS1.
December 2nd, 2010 at 11:46 am
The problem exists because there is no vast central database of UPC/Product assignments. There never has been.
During most of the history of the UPC, which began in the 1970’s, only large corporations could afford the technology to support it. They all created their own internal databases and regarded them as highly proprietary – they don’t share or sell them to anyone.
So, if you create a smart phone scanning app, your app has to access some database. Whose?
There are some web heavy hitters working on it, notably Amazon and Google. If someone manages the daunting task, they will have enormous power over the retail sector.