Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Understanding and Reducing Errors in Warehouse Operations with Intelligent Video Analysis

Helm, Max LU and Malikova, Alexandra LU (2022) MTTM02 20221
Engineering Logistics
Abstract
Warehouses are critical nodes in a supply chain, providing a valuable interface between supply and demand. One aspect that warehouse operations struggle with is errors which, if not addressed, will impact the customers downstream from the warehouse. One of the ways to address errors in warehouses is through technology, such as automation of processes, though these are not always effective. A new technology introduced within the last five years is Intelligent Video Analysis (IVA). IVA is the combination of warehouse and video data used to identify and process event in warehouse operations. It can be used retro-actively to analyze and learn from errors that have occurred, but applications of real-time analysis and error prevention are... (More)
Warehouses are critical nodes in a supply chain, providing a valuable interface between supply and demand. One aspect that warehouse operations struggle with is errors which, if not addressed, will impact the customers downstream from the warehouse. One of the ways to address errors in warehouses is through technology, such as automation of processes, though these are not always effective. A new technology introduced within the last five years is Intelligent Video Analysis (IVA). IVA is the combination of warehouse and video data used to identify and process event in warehouse operations. It can be used retro-actively to analyze and learn from errors that have occurred, but applications of real-time analysis and error prevention are emerging as well.

We study how retro-active IVA can be used to reduce errors in warehouse operations through a phenomenon-driven exploratory multiple-case study of six warehouses in Europe. The case studies were carried out mainly through interview with managerial staff with experience of using IVA. In addition to interviews, we have analyzed metrics relevant for errors, and observed video recorded errors.

A key finding from our research is that there is a critical flaw in how errors are measured and addressed in warehouse operations. It was found that the companies mainly used customer claims to capture the occurrence of errors. However, by analyzing the outbound operations with IVA, several companies could confirm that claims were sometimes caused by errors that did no occur in the warehouse. These errors were incorrectly included in their metrics, which misguided their preventative actions and incurred unnecessary costs. By excluding non-warehouse errors, companies have been able to increase the accuracy of their metrics and avoid ineffective processes changes.

The study also provided valuable insights into the errors that do occur in the warehouse and how they can be reduced. The most common error type was delivering the wrong quantity, but delivering the wrong item or in a damaged condition were not uncommon error types either. Furthermore, retro-active IVA could be used to provide insight into the causes of the errors. We conclude that addressing errors through these insights is effective for technological and organizational causes of errors, but not as effective for human factors, like distraction or stress. Therefore, although reducing errors completely with current technology does not seem possible, companies that implement retro-active IVA can get much closer to having error-free operations and metrics that accurately reflect their performance.

This thesis is, to the best of our knowledge, the first study of practical use of IVA in warehouse operations. It provides valuable novel insights into a new technology in warehouse operations, as well as contributes to the existing knowledge of errors in warehouse operations. While we found that IVA has potential to drastically reduce errors, more research is needed. Our research is based largely on interviews with warehouse staff on a managerial level. Therefore, researching quantitative error reductions and perspective of the operators are two particularly relevant areas to further research. (Less)
Popular Abstract
Using video to eliminate the “guessing game” of errors in warehouse operations by Max Helm and Alexandra Malikova (May 2022)

How do you make correct decisions and effective improvements if you are acting on faulty information? Our study shows that many warehouses are forced to guess the cause of their errors, and sometimes, even try to solve errors that are not actually there. This phenomenon could be a thing of the past thanks to a new technology called Intelligent Video Analysis.

When competing for customers who ask for faster deliveries and lower prices, the company’s warehouse plays a critical role. However, speed and low costs are worth little to the customer if they do not receive what they ordered in the first place.

So,... (More)
Using video to eliminate the “guessing game” of errors in warehouse operations by Max Helm and Alexandra Malikova (May 2022)

How do you make correct decisions and effective improvements if you are acting on faulty information? Our study shows that many warehouses are forced to guess the cause of their errors, and sometimes, even try to solve errors that are not actually there. This phenomenon could be a thing of the past thanks to a new technology called Intelligent Video Analysis.

When competing for customers who ask for faster deliveries and lower prices, the company’s warehouse plays a critical role. However, speed and low costs are worth little to the customer if they do not receive what they ordered in the first place.

So, how can the warehouse ensure fast, low-cost service while delivering accurately? Companies have been striving to reduce their errors for many years and countless technologies exist to aid them in that purpose. In many ways, these technologies have been successful, but not successful enough to root out all errors. The remaining errors elude the warehouse managers, forcing them to guess the causes based on angry customer claims.

A recent technology in warehouse operations called Intelligent Video Analysis (IVA) has shown strong promise in better identifying errors. IVA works by combining video with data from different information systems in the warehouse, such as warehouse management systems and barcode scanning. With this data link it is possible to create searchable video material. For example, one can use an order ID or customer ID to quickly find and review how a product was handled.

Our interviews with early adopters of IVA found how problematic it is to rely on limited information, especially from a source outside the warehouse. When claims could be analyzed with video, the consequences of the guessing game were revealed. It was easy to assume that the employees were careless, when in fact many errors never even occurred in the warehouse at all! The customer and the transport provider were found to be just as likely to be responsible for the error. For example, it is easy for a customer to miscount products when receiving a large order. Furthermore, when there was an error in the warehouse, the employee was not always at fault. Namely, the technology and processes in place, which are meant to guide the employee, sometimes led them into making mistakes.

IVA revealed the truth behind many errors, shedding light on the questionable effectiveness of previous improvement initiatives. Previous corrective actions such as giving an operator feedback on an error they did not make or making investments in areas where errors did not actually occur seem like a waste of time and resources now.

Our study of IVA has shown that using claims alone to guide and evaluate improvement initiatives in warehouse operations is risky business. Claims can be misleading in identifying where the error occurred and does not provide insights into why it happened. By using video to objectively analyze what really happens in warehouse operations, the “guessing game” can be eliminated. When improvements are guided by facts instead of guesses, companies can get closer to eliminating errors in warehouse operations.

This popular scientific article is derived from the master thesis: Understanding and Reducing Errors in Warehouse Operations with Intelligent Video Analysis, written by Max Helm and Alexandra Malikova (May 2022). (Less)
Please use this url to cite or link to this publication:
author
Helm, Max LU and Malikova, Alexandra LU
supervisor
organization
course
MTTM02 20221
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Warehouse operations, intelligent video analysis, video analysis, warehouse errors, non-warehouse errors
report number
5966
language
English
id
9080780
date added to LUP
2022-05-31 14:53:22
date last changed
2022-05-31 14:53:22
@misc{9080780,
  abstract     = {{Warehouses are critical nodes in a supply chain, providing a valuable interface between supply and demand. One aspect that warehouse operations struggle with is errors which, if not addressed, will impact the customers downstream from the warehouse. One of the ways to address errors in warehouses is through technology, such as automation of processes, though these are not always effective. A new technology introduced within the last five years is Intelligent Video Analysis (IVA). IVA is the combination of warehouse and video data used to identify and process event in warehouse operations. It can be used retro-actively to analyze and learn from errors that have occurred, but applications of real-time analysis and error prevention are emerging as well. 

We study how retro-active IVA can be used to reduce errors in warehouse operations through a phenomenon-driven exploratory multiple-case study of six warehouses in Europe. The case studies were carried out mainly through interview with managerial staff with experience of using IVA. In addition to interviews, we have analyzed metrics relevant for errors, and observed video recorded errors.

A key finding from our research is that there is a critical flaw in how errors are measured and addressed in warehouse operations. It was found that the companies mainly used customer claims to capture the occurrence of errors. However, by analyzing the outbound operations with IVA, several companies could confirm that claims were sometimes caused by errors that did no occur in the warehouse. These errors were incorrectly included in their metrics, which misguided their preventative actions and incurred unnecessary costs. By excluding non-warehouse errors, companies have been able to increase the accuracy of their metrics and avoid ineffective processes changes.

The study also provided valuable insights into the errors that do occur in the warehouse and how they can be reduced. The most common error type was delivering the wrong quantity, but delivering the wrong item or in a damaged condition were not uncommon error types either. Furthermore, retro-active IVA could be used to provide insight into the causes of the errors. We conclude that addressing errors through these insights is effective for technological and organizational causes of errors, but not as effective for human factors, like distraction or stress. Therefore, although reducing errors completely with current technology does not seem possible, companies that implement retro-active IVA can get much closer to having error-free operations and metrics that accurately reflect their performance. 

This thesis is, to the best of our knowledge, the first study of practical use of IVA in warehouse operations. It provides valuable novel insights into a new technology in warehouse operations, as well as contributes to the existing knowledge of errors in warehouse operations. While we found that IVA has potential to drastically reduce errors, more research is needed. Our research is based largely on interviews with warehouse staff on a managerial level. Therefore, researching quantitative error reductions and perspective of the operators are two particularly relevant areas to further research.}},
  author       = {{Helm, Max and Malikova, Alexandra}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Understanding and Reducing Errors in Warehouse Operations with Intelligent Video Analysis}},
  year         = {{2022}},
}