Anti-theft response randomizer

Abstract

Systems and methods for maximizing the deterrence effect on theft. Specifically, systems and methods for selecting and randomizing at least one response to potential theft events while minimizing impact on store personnel productivity in a retail setting. A plurality of defined event triggers detected by a monitored source results in the randomization of response to detected event.

Claims

1. A system for maximizing theft deterrence in a retail setting comprising: (a) providing at least one monitored source programmed to identify one or more suspicious events related to an action of an individual; (b) evaluating the risk associated with the one or more suspicious events; (c) selecting among one or more response types based on (b), wherein the one or more response types is selected from the group consisting of local deterrent alarm, store personnel notification, notification of adjacent stores, remote notifications and no notification of store personnel; and (d) randomizing the one or more response types based on results in (c), wherein the randomized response results in an inability of the individual being monitored to determine any relationship between the one or more suspicious events related to an identical action of the individual and the response from the system.

2. The system of claim 1, wherein the at least one monitored source is selected from the group consisting of merchandise activity sensors monitoring vibration or product removal, RFID detection, weight detection, cameras, infrared sensors, alarmed display devices, light and motion sensors and perimeter sensors.

3. The system of claim 1, wherein the at least one monitored source is configured to detect merchandise removal from fixtures, removal of packaging from merchandise, concealment of merchandise, removal of price or security tags from merchandise, or any other detection of theft related activity.

4. The system of claim 1, further comprising one or more environmental factors, the one of more environmental factors is selected from the group consisting of store traffic, staffing levels, facial recognition, mobile device recognition, regional activity, event correlation, response compliance, time of day and manual adjustment of settings.

5. A method of selecting and randomizing at least one response to potential theft events while minimizing impact on store personnel productivity in a retail setting, the method comprising: (a) providing a security system configured to identify one or more suspicious event triggers from at least one sensor or monitoring system that is monitoring an individual; (b) considering one or more environmental factors once the one or more suspicious event triggers is identified; (c) selecting one or more response types from the security system based on the one or more suspicious event triggers after considering the one or more environmental factors, wherein the one or more response types is selected from the group consisting of local deterrent alarm, store personnel notification, notification of adjacent stores, remote notifications and no notification of store personnel; (d) allowing the security system to execute the one or more response types based on results from (c), wherein the response is randomized, resulting in an inability of the individual being monitored to determine any relationship between the one or more suspicious event triggers related to an identical action of the individual and the response from the security system.

6. The method of claim 5, wherein the at least one sensor or monitoring system is selected from the group consisting of merchandise activity sensors monitoring vibration or product removal, RFID detection, weight detection, cameras, infrared sensors, alarmed display devices, light and motion sensors and perimeter sensors.

7. The method of claim 5, wherein the one or more environmental factors is selected from the group consisting of store traffic, staffing levels, facial recognition, mobile device recognition, regional activity, event correlation, response compliance, time of day and manual adjustment of settings.

8. A method of reducing merchandise shrink while minimizing impact on store personnel and shopper experience in a retail environment, the method comprising: (a) providing a system configured to detect and identify one or more suspicious event triggers from at least one sensor or monitoring system that is monitoring an individual; (b) considering one or more environmental factors once the one or more suspicious event triggers is identified; and (c) determining one or more response types from the system based on the results in (b), wherein the one or more response types is selected from the group consisting of local deterrent alarm, store personnel notification, notification of adjacent stores, remote notifications and no notification of store personnel; and (d) executing the one or more response types based on the results from (c), wherein the response is randomized, resulting in an inability of the individual being monitored to determine any relationship between the one or more suspicious event triggers related to an identical action of the individual and the response from the system.

9. The method of claim 8, wherein the at least one sensor or monitoring system is selected from the group consisting of merchandise activity sensors monitoring vibration or product removal, RFID detection, weight detection, cameras, infrared sensors, alarmed display devices, light and motion sensors and perimeter sensors.

10. The method of claim 8, wherein the one or more environmental factors is selected from the group consisting of store traffic, staffing levels, facial recognition, mobile device recognition, regional activity, event correlation, response compliance, time of day and manual adjustment of settings.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The novel features of this invention, as well as the invention itself, both as to its structure and its operation, will be best understood from the accompanying drawings, taken in conjunction with the accompanying description, in which similar reference characters refer to similar parts, and in which:

(2) FIG. 1 illustrates an overview of the response randomizer of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

(3) With reference to FIG. 1, the basic operating sequence of the invention is to: 1. Accept suspicious activity event triggers from various monitored sources (sensors and/or systems and human input); 2. Evaluate the nature of the suspicious activity; 3. Correlate that event with any others that may be relevant; 4. Identify, evaluate, and apply the applicable environmental factors and variables; 5. Select among a range or response types; 6. Apply an appropriate level of randomization; 7. Initiate the optimal response (local alarm and/or notification action(s), if any).

(4) The invention can accept suspicious event triggers from virtually any type of device or system capable of detecting events of interest; examples include but are not limited to: Merchandise Activity Sensors monitoring vibration induced into store fixtures when merchandise is removed; Merchandise Activity Sensors of any kind that detect the removal of merchandise from store fixtures. Cameras & Video Management Systems capable of detecting suspicious behavior (such as unusual loitering, rapid product removal or any unusual shopping behaviors); Infrared Sensors detecting presence dwell (loitering) at high risk locations and reaches into merchandise displays (i.e., an infrared “curtain” detecting merchandise interaction); 3D Camera Systems monitoring removal of merchandise from a store fixture; Alarmed Display Devices, often connected to cameras and other high end items, that permit shoppers to pick up the item but that detect if the attached restraint is removed; Anti-Sweep Devices & Fixtures that limit and/or mechanically monitor merchandise removal (includes instrumented locked dispensing fixtures, twist knob dispensing devices, flip doors, merchandise pushers, and peg hooks); RFID detecting tag movement from a shelf or a defined area; Light & Motion and similar devices outfitted with a transmitter that are mounted to merchandise to detect suspicious handling or, by virtue of numerous such devices subjected to near-simultaneous movement (which may suggest an in-process theft sweep); Fitting room occupancy sensors; Shopping cart sensor systems that detect a path to the store exit without a passing through a cashier station; Unauthorized presence sensors behind jewelry or other service counters; and Perimeter door switches.

(5) Once a trigger is received from one or more of the above sensing techniques, the invention evaluates recent alarm activity from the originating source and the system overall along with various environmental factors to determine what, if any, alarm or staff notification will be issued.

(6) These environmental factors may include one or more of the following: Store Traffic/Occupancy: The store's traffic monitoring system provides real time information on the quantity of persons entering and exiting the store, providing a means of determining the approximate quantity of people in the store at a given time; Staffing Level by Skillset: The store's time clock system provides information on the quantity of employees by skillset available in the store at a given time; Facial Recognition: The store's facial recognition system (typically of persons entering the store) can provide notification of the presence of known or suspected high risk individuals; Mobile Device Recognition: Mobile devices previously detected and associated with suspicious activity in this or other stores indicate the presence of suspected high risk individuals; License Plate Recognition: Vehicle plates associated with known or suspected high risk individuals or groups entering the store's parking lot; Regional Activity: Real time sharing of detected or known theft activity among stores in a geographical area (this may include human reporting of actual theft events, facial and/or license plate recognition information or may simply be limited to activity related to events of interest, such as likely ORC sweep events); Event Correlation: Receipt of triggers of other relevant events within a reasonable time proximity; for example, separate merchandise movements (on nearby display locations or even throughout the store) that might collectively represent theft sweep activity; Response Compliance: Some systems incorporate a means of confirming response by store personnel to a detected event; for example, a notification of a sweep event may be sent to store personnel who, upon responding to the area, press a button in that area or are otherwise confirmed to have responded within a reasonable time; this compliance rate may influence the probability of notifications to store personnel to subsequent detections; Randomizer: In addition to considering any combination of the above factors, the invention can be configured to randomly process event triggers within prescribed algorithm limits; Time of Day/Week/Year: Each of these three timing factors may be taken into account by the invention in determining trigger processing; Manual Adjustment: Based on observation or other factors, a manager or other authorized person can direct the invention to increase or decrease the level of aggressiveness of notifications either temporarily, or optionally, as a general setting.

(7) While systems configurations and their capabilities vary considerably, once the invention evaluates the event trigger and relevant environmental factors, it determines what response action(s) a given event will then trigger. Available response actions may include, but are not limited to, one or more of the following: Local Deterrent Alarm: Sound and/or light in proximity to the suspicious event gains the attention of nearby persons (especially thieves, who are typically hyper alert); these local alarms may manifest in a variety of form factors including: Local Annunciator: Typically a basic device with a speaker and lights; Camera: Could be a real or imitation camera outfitted or associated with a speaker or other audio device and/or lights; Public View Monitor (PVM): These video display with integral camera units are often mounted in the vicinity of high theft activity and may be activated to take increasingly aggressive sound, light, and video display actions depending on the situation; Increased illumination of merchandise: Simply turning on additional lighting in the area of interest. Locked Merchandise: Initiating an automated locking mechanism that prevents removal of merchandise from a fixture Any other theft deterrent action: The Randomizer can activate any theft deterrent device Store Associate Notifications: All or select store personnel may be notified using various communication channels including Public Address systems, two-way radios, pagers, wireless phones, and mobile smart devices; Notification of Adjacent Stores: Notification of stores within a limited distance from the store which experienced a large theft event is helpful as ORC rings hit multiple stores in a market in the same day. Remote Notifications and VMS Integrations: Especially situations in which store video cameras are monitored/analyzed at a remote monitoring station, the invention uses network and other communication channels to notify remote monitoring personnel and/or automated Video Management Systems (VMS).

(8) While the actual evaluation algorithms are a highly configurable trade secret, the following information provided below at Table 1 discloses how various environmental factors may be considered.

(9) TABLE-US-00001 TABLE 1 Algorithm Variables Factor Description Typical Impact Store Traffic Using store entry traffic As the ratio of shoppers count sensors and exit to available store sensors and/or average associates (sometimes shopping duration metrics, based on event location approximate number of and associate skillsets), shoppers in the store the threshold to is determined. triggering in-store Staffing Level Using time clock and POS response notifications login activity and data, increases (i.e., the quantity of available notifications will be store associates by less likely to trigger skillset is determined. in limited resource situations). Facial Persons entering the The identification Recognition store and/or at locations itself may trigger a within the store are notification event; compared with a database additionally, any other to identify individuals events (especially if or groups of individuals associated with a past known or suspected to be modus operandi, such as involved with theft. the theft of razor Mobile Device Similar to facial blades) will be handled Recognition recognition but with higher identification is made aggressiveness. Two or through identifiable more persons of interest signatures of mobile in close time proximity phones carried by who were previously persons of interest. detected as a group also License Plate A camera at the parking increases aggressiveness. Recognition lot entrance or other location(s) detects license plate numbers to determine if past events of interest correlate with that plate. Regional Nearby stores within a As ORC teams often Activity chain or cooperating target a series of stores of different nearby stores - chains provide real- typically sweeping the time notification of same items - awareness select events of of a team operating interest (especially nearby increases alarm theft sweeps likely and notification performed by ORC frequency and teams). aggressiveness. Event All event triggers If individual events Correlation received within a are determined to reasonable time likely correlate to frame are evaluated a suspicious activity, for possible action aggressiveness correlation with and response frequency each other. increases. Response Confirmation of Poor compliance will Compliance store personnel typically increase responding to an notification event of interest aggressiveness (e.g., in a timely fashion.* reducing the threshold justifying a notification and speeding escalations to management). Randomizer After all The randomizer action environmental factors reduces the percentage are considered, the of notifications and selection and the actual response frequency or responses and associate is then randomized to, notifications occur 1) create uncertainty, in a random fashion. 2) limit resource utilization and 3) increase compliance Time of Identifying time The algorithm uses Day/Week/Year frames during which the specified action specific system level as a final actions are desired consideration as to (e.g., high/medium/ what, if any, action(s) low event action will be taken in aggressiveness); response to a given these are often event). related to anticipated shopper traffic, staffing levels, and known theft vulnerability (perhaps related to specified store zones). Manual Authorized personnel Aggressiveness adjusts Adjustment (such as store for a specified duration management) temporarily of time. adjust notification aggressiveness based on conditions. *A variety of methods can be used to confirm response to an event of interest. Proactive methods include pressing a button or scanning a bar code located in that area, among other similar methods. Automated methods include video or beacon detection of the presence of a responding employee.

EXAMPLES

(10) A great example of the invention in use can be shown through protection of the “cosmetics wall” at a national drug store chain. In any drug store chain, one of highest revenue and profitability categories, besides prescription drugs, is cosmetics. Unfortunately, it is also the highest theft area in the store. The cosmetics category has many characteristics which make it particularly vulnerable and attractive to thieves: 1) Items tend be relatively high priced ($10 or more); 2) Thousands of SKUs (many unique products); 3) Small size makes them easily concealed; 4) High total value of products can be stolen with little physical volume of goods; 5) High product demand (especially hot new lines of cosmetics); 6) Easily resold through alternate channels (eBay, swap meets, resold to other retails, moved internationally etc.) 7) Drug stores deploy very few personnel; most are unable to leave the cash register area; 8) Stores are often open 24 hours with very limited personnel during late night hours.

(11) These characteristic make this category highly attractive to all three theft categories: opportunistic, internal, and ORC. However, due to the large quantities of merchandise stolen in each theft event, ORC theft typically represents more than half of total losses. In their highest shrink stores, this chain experiences more losses from theft than is earned in sales, resulting in a net loss for the category. In addition, following an ORC theft event, the shelves of targeted brands are literally stripped clean of merchandise. This severely erodes sales as subsequent shoppers can no longer purchase the product. In this chain's case, despite numerous efforts and approaches to reduce cosmetics theft, shrink continued to increase year over year. Given these failures, the chain elected to install a new product protection system incorporating many elements of this invention.

(12) In this application, two types of devices were installed in the cosmetics category.

(13) 1) Merchandise movement detection devices which count items being removed from shelves. These sensors were affixed to shelves with the most theft-prone products to detect when excessive items were removed within a short time frame. For example, removal of five or more units in less than 10 seconds strongly suggests an in-progress ORC theft event (a sweep event).

(14) 2) A simulated Dome Camera was installed over the cosmetics sales area. This highly visible device, with the outward appearance of a security camera, detects people dwelling in front of cosmetics merchandise. The device can annunciate voice messages and illuminate integral lights which, when flashing, simulate the initiation of active security surveillance.

(15) A range of responses initiated when a suspicious event was detected. These responses fall into two broad categories: a) local deterrents, such as attention-getting tones or voice announcements, flashing lights, activation of Public View Monitors etc. and b) notification of store personnel via walkie talkies, the store's Public Address system, or other channels. These two categories of responses were individually randomized by the invention.

(16) Given the staffing constraints of this drug store environment, store personnel notifications had to be severely limited even though, as noted previously, store personnel response is the optimal action to stop an ORC event in progress. Still, given the sophistication of the professional ORC thief, the local response also had to be unpredictable. All the while, these same devices had to deter opportunistic theft as well as internal theft. Under these considerations, the invention was deployed to randomize the response with algorithm variances influenced by time of day, day of week, store staffing characteristics, the type of thief being impacted, and inherent store shrink profile. The deployment of the invention had these behavioral impacts: 1) The random nature of the responses made ORC thieves particularly uncomfortable; 2) ORC thieves could no longer devise strategies to thwart predictable responses; 3) Opportunistic thieves received immediate local deterrents, driving a heightened sense of physical security in the area; 4) Store personnel were notified to respond to the area a small fraction of the time driving their compliance with such requests to very high levels.
CASE RESULT: After years of increasing cosmetic category shrink, this chain experienced an immediate and sustained 52% reduction in shrink directly resulting from the deployment of the invention. It was simply wholly impractical for the sensors to be deployed absent the randomization of the response. The invention alone enabled the functioning of the sensors to be not only effective against thieves but, perhaps more importantly, compatible with the constraints and realities of this challenging retail environment.