Voice Assisted Technology Providing Immediate Feedback To Reduce Employee Errors

1
Voice Assisted Technology Providing Immediate
Feedback To Reduce Employee Errors

• Samuel M Berger, Timothy D. Ludwig, Drew Link,
  Donald Kaade

2
Errors

• In large distribution warehouses errors add up
  quickly and become quite costly.
• Selectors perform a key function in the
  distribution process.
• 3 Types of errors Shorts, Mispicks, Damages
• Selectors pick hundreds of cases each day and
  receive little or no feedback regarding the
  accuracy of their selections.

3
Feedback

• Feedback can used as a tool to increase
  performance.
• When paired with tiered goals and incentives
  feedback has been shown to improve selector error
  rates (Bateman Ludwig, 2003).
• Alvero, Bucklin and Austins (2001) review of
  performance feedback.
• Immediate Feedback gt Delayed Feedback

4
The Three-Term Contingency
A - Antecedent
B - Behavior
C - Consequence
The closer in time these contingencies occur the
greater the behavior change
5
The Problem

• Managing performance in the real world has been
  hampered by
• Antecedents (e.g., response information) that
  are not proximal to the behavior they prompt or
  easily accessible in the work process.
• Behaviors that cannot always be reliably
  measured.
• Consequences that are delayed from the
  occurrence of the behavior.

6
The Problem
A - Goal to reduce errors to 2.5/1000 cases B -
picking the right product in the right amount C -
If you achieve goal you do not lose the 1.50 for
each picking error
Goal set in September in a memo
Feedback with consequence delivered with
paycheck up to 13 days later!
7
Both Scanning and Voice make contingencies
proximal to behavior
8
The Technology Solution

• Technology allows us to make the contingencies
• proximal to the behavior

A - Goal to reduce errors to 2.5/1000 cases B -
picking the right product in the right amount C -
If you achieve goal you do not lose the 1.50 for
each picking error
Goal set in a memo
Goal updated continuously and delivered at the
time of behavior
Feedback with consequence delivered with paycheck
up to 13 days later
Feedback delivered immediately after behavior
incentives earned updated continuously
9
Voice

• Battery-powered waist unit headset.
• Each unit is sent information regarding orders
  via radio-frequency.
• - Some providers claim
  
• 99.99 accuracy

10
Methods

• AB design with a non-equivalent control group and
  a control behavior.
• 3 Departments Grocery, Produce, Mezzanine
• Subjects 132 Selectors
• Audits were done on 3 of the outgoing orders
• DVs Shorts, Mispicks and Damages
• Damages were used as a behavioral control

11
Grocery
12
Produce
13
Mezzanine
14
Low vs. High PerformersGrocery
15
Low vs. High PerformersProduce
16
Discussion

• There is a clear impact on error rates after the
  implementation of voice
• 99.99 accuracy was not realized but the decrease
  in error rates was statistically significant.
• Low performers vs. High performers.

17
Cost Benefit Analysis

• Error estimate 15.75
• Total cost of voice implementation 1.3 million
  dollars
• Estimated yearly savings 1.48 million dollars

18
Limitations Future Research

• Experimental design used a pre-existing control
  group (mezzanine).
• Lack of control
• Future research should attempt to use a more
  appropriate control group and/or a laboratory
  setting.