Robo Retail - a new take on theft reporting
The mobile application "Robo Retail" is dedicated to the use case of shoplifting and rethinks the way in which shoplifting reports can be processed by the retail company. With the help of AI and automation in the scanning of IDs and products, detectives and employees are enabled to report shoplifting in a legally compliant and efficient manner. As a result, the solution not only saves time and communication loops, but also builds a data set that supports the organization of detectives and the prevention of shoplifting.
Project duration: 3.5 months - 2021
Type of project: University project within one semester
Focus: Public sector , Data Driven Experience
Tools: Figma
Methods: Interviews (online and on-site), usability tests, user interviews, user journey, business model, paper prototyping, wireframes, interactive mockups
Team: Miriam Steckl, Felix Tesche
Problem
4.4 billion euros lost annually in Germany due to shoplifting
The question of how to develop a sustainable and profitable business model led the team in the direction of B2B concepts. Particularly striking during the research is the "Crime Barometer Berlin / Brandenburg", which showed that shoplifting is a serious threat to businesses and the most common crime. A study by the EHI Retail Institute also revealed the extreme figure of 4.4 billion euros lost annually in Germany due to inventory discrepancies and shoplifting. The crime barometer explicitly stated that the crime is often not reported by retailers, is not taken seriously and that the number of criminal charges must increase in order to reduce the overall crime of shoplifting.
Research
Status quo - too much time for too many paper forms
We found a standardized paper form on the Internet for small retail businesses to file criminal complaints. The three-page paper form again contained problems that we already knew well from the digital police report form. These included a lack of logic and the need to enter the same personal information about the business etc. over and over again. We came to the realization that the time taken to complete the form could be reduced by using digital means of automation and pre-built components. Our analysis of the different components can be seen below.
Data that can already be predefined
Data that can be scanned
Manual input data
Calculation
Authentification
User Research - Second Round
At the beginning of the research, we conducted interviews with retail employees and detectives. These were intended to help us gain a better understanding of how employees and detectives work with the paper forms on a daily basis and where potential problems arise when filling out the forms.
The interviews were semi-structured, partly digital and partly on-site. Our aim was to find answers to the following questions:
-
Information about shoplifting
(How often? Which products? How to recognize them?)
-
Criminal report forms
(Who fills it out? What is the process behind it?)
-
Possible future business models
(How high are the calculated costs for a theft report? How can a digital solution be implemented in existing processes? What legal issues need to be considered?
The interviewees included police officers, employees, store managers and criminal investigators from the following companies: Edeka, Rewe, H&M, Alnatura, Nanu Nana.
The three most important findings after the evaluation:
-
Theft of goods worth less than 50 euros is often not reported to the police
Since shoplifting under 50 euros was often dropped by the lawyers, the employees did not see the urgent need to press charges, as the time involved was too high to make it worthwhile.
-
Trespassing is of greater legal importance
This is the case if a former shoplifter who has been banned from entering the store violates the ban again. However, as all data is recorded manually, shopkeepers often do not know whether a person has already committed shoplifting and could be arrested for violating the ban.
-
Criminal complaints must always be signed by the managing director, which lengthens the process
Criminal record extracts are in most cases completed manually by an employee or detective, but must be signed by the manager to be legally correct.
What our experts say
We conducted interviews with the following groups of people:
5 Store manager
h&m, Edeka, Rewe,
Alnatura, Nanu Nana
2 Prosecutors
from Berlin & Brandenburg
5 Police officers
Federal Police &
State Police
5 Penalty notifier
"
"I usually fill out a criminal complaint as store manager or a deputy together with the store detective. It usually takes at least half an hour for us to take personal details and write the crime and product on the form. But it can also take one hour or two if we have to call the police."
"
"Thefts happen here almost every day, across all age groups. Detectives detect them twice a week. Liquor, salmon, coffee and sweets are stolen particularly frequently. These are then often sold on."
Frequency of theft, product type
Who files the criminal complaint?
Store manager, Edeka
Store manager, Edeka
"
"If we recognize an offender for the second time, we fill out a separate form for the house ban under the dual control principle, which is required. We send this to the main branch in Stuttgart. From there it is sent to Hamburg. If repeat offenders are detected again, criminal charges can be filed for trespassing. However, I only know who is often there from my shift. I have no idea who has been banned from other shifts or branches."
Thefts and trespassing are not known across all stores
"
"Depending on the material value of the products, it is often simply not worth reporting them. We know that the public prosecutor's office will drop the charges anyway if the value is as low as 50 euros, so it's simply not worth the effort in terms of time and personnel costs."
When is it worth reporting a crime?
Store manager, Edeka
Head of department women, h&m
Final service design concept
Based on the results of our desktop research and other findings from user research, the team developed Robo Retail. This solution reinvents shoplifting reporting by digitizing the process in an efficient way.
Before we created our interactive prototype with Figma, we first made some paper sketches as a team, then looked for commonalities and combined the ideas into a final solution.
.png)
.png)
How it works
Total duration 5 minutes
Step 1:
Scan identity card
Scan ID cards and thus speed up the process of recording personal data. Additional ML technologies can be used to check the personal data on the ID card.
Step 2:
Scan goods
The stolen goods are automatically recognized and recorded via barcode. The final amount of the shoplifting can then be calculated automatically.
Step 3:
Describe the facts
Finally, describe the facts and complete the advertisement. The process is less time-consuming thanks to digital input and the option to make suggestions. There is also no need to send it by post.
What advantages does Robo offer?
Speed
Our solution drastically speeds up the process
From a minimum of 30 minutes to fill out the paper version, we promise that a criminal report will be completed within 5 minutes, reducing the time by 5/6 of that.
Multi-detection
Multiple detection of repeat offenders
By building up a data set, retail stores are able to identify people who have already committed an offense in their store and who may be subject to a current ban, and can therefore automatically file a trespassing charge.
Better planning and deployment
Improved use of store detectives
The data set also makes it possible to draw conclusions from statistics, e.g. which products were stolen most frequently or when would be the best time to hire detectives. This will improve preventive measures.
Total duration: 5 minutes
Total duration: 45 minutes
Form working time with Robo
Form working time with paper, printer, mail
Screencast of the Prototype
User tests and evaluation
We recruited retail employees and detectives for the first user tests.
In several rounds of interactive usability tests with the Figma prototype, we gained further insights into user behavior and current problems with existing processes in the stores.
The prototype itself was very well received and users were able to quickly understand the logic of our solution.
An important insight from the testing was not to reduce the number of "pages" in the prototype too drastically. At the beginning, experts had wanted to reduce the number of pages in order to minimize the effort and time required to enter the data. During testing, however, we found that users understood the flow of the prototype better when each step of the process was on a single page, just like on paper.
Excerpt from the usability test