Retail Shrink Reduction Case Studies: What Actually Worked?
In comparison-store tests, these measures came in below control stores: protective fixtures and special handling procedures on razor blade packs (OJP); in-aisle public view monitors (57 percent below control) and domes (27 percent) (ECR Retail Loss); and safer cases, which cut shrink 53 percent against control stores in a 10-store test (ECR Retail Loss). Named results, such as Stage Stores’ 48 percent drop across fifty high-risk stores (LP Magazine), are the retailer’s own before-and-after numbers with no comparison stores. As of 2024, FY2022 remained NRF’s latest shrink data (Retail Dive).
What the strongest controlled studies have in common
The best-controlled shrink studies share one design choice: comparable stores are randomly assigned to get the change or not before anything is installed. Hayes and Downs ran a 47-store randomized trial of three treatments: in-aisle public view monitors, in-aisle dome cameras, and protective keeper boxes (OJP). For the camera comparison, ECR Retail Loss reports monitors in 15 stores, domes in 9, and 14 random control stores, with six weeks of data before and after; shrink on the audited razor blades was 27 percent lower than control in the dome stores and 57 percent lower in the monitor stores (ECR Retail Loss). ECR also notes the study set positive shrink counts to zero, which can inflate measured loss (ECR Retail Loss). ECR Retail Loss also reviews a 10-store safer-case test, in which five stores were randomly assigned to the cases and five to control, with nine weeks of data before and 13 weeks after; shrink was 53 percent lower and sales 69 percent higher in the test stores than in control (ECR Retail Loss). A randomized trial on premium shaving blade packs found protective fixtures cut losses 56 percent and special handling procedures 58 percent against control locations (OJP). A systematic review of security tags offers only a hedged parallel: it found suggestive evidence that “more visible tags are associated with greater reductions in theft than less visible tags” (UCL). When two camera options compete for the same budget, the monitor-versus-dome result is the controlled evidence to weigh.
How CCTV’s early gains faded in one study
Beck and Willis studied CCTV at a UK fashion retailer using a before-and-after design across 15 test stores (Beck and Willis). Within three months of installation, loss to sales across the stores fell from 2.45 percent to 1.97 percent (Beck and Willis). At six months, using adjusted figures because only 10 stores completed the study, the authors report loss to sales overall remained unchanged at 2.25 percent, with high-level stores rising and medium- and low-level stores falling (Beck and Willis). The authors suggest the fade can come partly from a relaxation in staff vigilance once cameras go in (Beck and Willis). For more on why cameras and tags lose their deterrent effect, see why some retail security cameras and tags fail to deter theft. Recheck shrink at the six-month mark after any camera installation before reporting that it paid off.
A case table: ranking shrink studies by evidence strength
The table puts randomized trials first and the named retailers’ own accounts last. A shrink drop during a bundled program with no comparison stores can also reflect that year’s hiring, staffing, or store traffic, which is why those accounts sit at the bottom.
| Case | Intervention | What was measured | Window | Comparison group | Result | Evidence strength |
|---|---|---|---|---|---|---|
| Hayes, Downs and Blackwood | Protective fixtures; special handling procedures | Loss on premium shaving blade packs | Pre and post test | Randomized control locations | Fixtures down 56%, procedures down 58% vs. control | Randomized controlled trial |
| Hayes and Downs (camera arms of a 47-store trial) | Dome cameras; public view monitors | Loss on audited razor blades | Six weeks before and after | 14 random control stores (ECR’s count) | Domes down 27%, monitors down 57% vs. control | Randomized controlled trial |
| Hayes et al. (10-store safer-case study) | Safer cases on razor blades | Shrink and sales | Nine weeks before, 13 weeks after | 5 random control stores | Shrink down 53%, sales up 69% vs. control | Randomized, small sample |
| Hayes et al. (2019 wraps trial) | Anti-theft wraps | Loss by product category | June to September 2010 | Random control stores (60 stores selected) | Significant drop on cordless drills; no significant difference on diet supplements or skincare | Randomized controlled trial |
| Beck and Palmer | Hard tags replaced with sewn-in soft tags, then reverted | Shrink vs. start of trial and vs. control | Trial period, then 6 months after reverting | 540 stores kept on hard tags | Soft tags: up 251% vs. start, 53% above control. After reverting: 14% below control | Large comparison group, no statistical test |
| Beck and Willis | CCTV in 15 test stores | Loss to sales | 3 months, then 6 months | None (before-and-after only) | Down to 1.97% at 3 months; unchanged overall at 6 months, with tiers moving in opposite directions | Before-and-after, no comparison stores |
| Stage Stores | EAS, CCTV and public view monitors, audits, and other measures bundled | Shrink percentage | 2002 vs. prior year | None reported | Shrink down about 48%, $3.7 million saved | Retailer’s own account, bundled program |
| Lowe’s | Not specified | Shrink dollars | One year vs. prior year | None reported | Lost fewer shrink dollars than the prior year while opening 125 new stores | Retailer’s own account |
The wraps trial shows why the product column matters. CrimeSolutions reports the trial ran from June to September 2010 with 60 stores selected (CrimeSolutions). Stores that got the wraps saw a statistically significant drop in the loss of cordless electric drills compared with control stores (CrimeSolutions), but there were “no statistically significant differences in the loss of diet supplements or skincare products” (CrimeSolutions). The researchers conclude that the effect of anti-theft wraps on in-store loss varies based on product category (OJP). Match a case study’s product category to your own before assuming its number applies to your store.
What the hard-tag reversal shows
In the Beck and Palmer study, one US apparel retailer replaced in-store hard tags with sewn-in label tags across 355 stores of one brand, while 540 stores kept hard tags as a comparison group (ECR Retail Loss). After the switch, shrink in the test stores was 251 percent higher than when the trial started and 53 percent higher than in the control group (ECR Retail Loss). The retailer went back to hard tags, and six months later shrink in those stores was 14 percent below the control group (ECR Retail Loss). ECR notes the study ran no statistical testing, so the pattern is an observation from one retailer, not proof of cause. Watch shrink in the affected stores for the months after any visible deterrent comes off.
What the named retailer accounts actually show
Both named accounts come from the retailer itself, with no comparison stores reported. Stage Stores reported about a 48 percent drop in shrinkage across its fifty highest-risk stores in 2002, a savings of $3.7 million (LP Magazine). That program combined physical security equipment, operational compliance audits, and other measures (LP Magazine), the kind of stacking the swiss cheese model of layered retail defenses describes, which makes it hard to credit any single layer for the result. Stage also found that stores with higher audit scores had lower shrinkage percentages (LP Magazine). Lowe’s said it “lost fewer dollars than we lost the previous year, while building 125 new stores” (LP Magazine). Ask what else changed that year before crediting either number to one tactic.
Is FY2022 still NRF’s latest shrink data?
A within-chain result also lacks an industry yardstick. After more than three decades of annual shrink research, NRF did not release a shrink report in 2024, and its spokesperson said the latest available shrink statistics were in the prior year’s National Retail Security Survey (Retail Dive). That survey put the FY2022 average shrink rate at 1.6 percent, or $112.1 billion in losses, up from 1.4 percent in FY2021 (NRF). Brand Elverston, a retail asset protection practitioner of more than 20 years, told Retail Dive, “It’s impossible for retailers to have sales data-like accuracy in the shrink arena” (Retail Dive). Compare this year’s total retail loss against your own stores’ prior years with a shrink calculator rather than treating the FY2022 average as current.
How to judge a shrink case study before you copy it
- Did it use comparison stores, or only a before-and-after measurement at the same locations?
- How long did the window run? ECR Retail Loss warns that many published shrink studies use data windows as short as one or two weeks (ECR Retail Loss).
- How was shrink counted: dollars, units, or a percentage of sales, and against what baseline?
- Does the product category match yours? A result on razor blades or drills does not automatically transfer to apparel or supplements.
- Who ran the test: an independent researcher, or the retailer reporting on its own program?
What to do this week
- Pull same-store shrink for your last two years so any new program has a baseline to beat.
- For any camera installed in the last year, put a six-month shrink check on the calendar.
- When a vendor shares a case study, ask for the control stores’ numbers alongside the test stores’.
Frequently asked questions
What results did Stage Stores report from its high-risk store program?
Stage Stores said shrinkage in its fifty high-risk stores dropped about 48 percent in 2002, a year-over-year savings of $3.7 million (LP Magazine). The company first designated one hundred stores but found that too many to manage effectively. The figure is the company's own account, with no comparison stores.
Do security tags actually reduce shoplifting?
A 2017 systematic review found fifty eligible studies on retail security tags, and only eight of them reported quantitative data on whether tags worked; five of those eight were positive (UCL). The authors called the overall effect difficult to determine.
How can you tell whether a shrink case study is reliable?
Look for stores that did not get the change, measured over the same period. ECR Retail Loss reviewed the published impact studies and warns that "manufacturers and providers of anti-theft interventions often make grandiose claims" (ECR Retail Loss). Give more weight to random assignment, a window of months, and an outside researcher.
Does CCTV stop retail theft long-term?
In one six-month study, it did not. Beck and Willis found loss to sales at a UK fashion retailer was unchanged at 2.25 percent after six months, using adjusted figures for the 10 stores that completed the study (Beck and Willis). It used a before-and-after design, so treat it as a reason to keep measuring.