Academic studies offer objective look at voice-directed picking technology

Studies show voice technology provides quick ROI and significant productivity increases but may not always increase picking accuracy.

By Staff -- Modern Materials Handling, 7/17/2007

Looking for an honest appraisal of voice-directed picking systems? Visit your local university library and ask for the Journal of Organizational Behavior Management.

The Spring 2007 issue of the academic journal includes two articles that discuss the merits of voice-directed order picking technology:

• The first article reports on a study by Timothy Ludwig and Samuel Berger, both of Appalachian State University, comparing the accuracy of voice-directed picking to the accuracy of picking directed by printed labels.
• The second reports on a study by Ludwig and David Goomas of Tarrant County College comparing the accuracy and productivity of voice-directed picking to that of picking directed by handheld scanners.

Ludwig and Berger’s article, based on a study conducted in a grocery distribution center, includes a cost/benefit analysis that shows the voice-picking system paid for itself in about one year through accuracy improvements alone.

The study found that voice technology significantly improved accuracy rates, but it didn’t boost rates to the 99.99% level often advertised by suppliers.

Voice technology didn’t reduce errors much for employees who were already performing well, says Ludwig, but the voice system made a tremendous difference for employees with a history of errors. This finding, he says, suggests voice technology might be a good training tool to improve accuracy rates for low-performing employees.

Ludwig and Goomas’s article compares RF- and voice-picking systems in an automobile parts DC. According to the article, the conversion to voice technology resulted in a 17% increase in productivity and to a minor decrease in picking accuracy (from 99.80% to 99.55%).

With the voice system, explains Goomas, picking accuracy was more dependent on upstream events. Picking errors occurred when picking locations were incorrectly labeled or were replenished with the wrong product.

The two articles may be the most reliable data available on voice technology, say Goomas and Ludwig, thanks to the rigorous experimental controls applied to the studies and to the researchers’ objectivity.

“We set out to conduct a fair test,” says Ludwig. “We weren’t funded by any company. We had nothing to prove either way.”