Retail, food service, and hospitality operators generate more data today than at any point in the industry's history — point-of-sale transactions, loyalty programs, video feeds, labor schedules, supply chain logs, and e-commerce activity all produce a constant stream of signals. The challenge isn't a lack of data. It's turning that data into decisions fast enough to matter, especially when it comes to protecting margin from theft, fraud, and operational loss.
That's where third-party analytics comes in. Below, we break down what third-party analytics actually means, the four core types of data analytics every loss prevention (LP) leader should understand, and how partnering with an outside analytics firm can sharpen fraud detection and unlock insights an internal team may not have the bandwidth to find on its own.
What Is Third-Party Analytics?
Third-party analytics refers to data collection and analysis performed by an outside firm on a business's behalf. Rather than relying solely on an internal team to build reports, monitor exceptions, and investigate anomalies, a company brings in a specialized partner to manage some or all that work.
In a loss prevention context, the "first party" and "second party" are typically the retailer and its customers, or the retailer and its employees. The analytics firm is the third party — an outside expert reviewing transactional, video, and operational data to flag risk and surface opportunity.
Four Kinds of Data Analytics
Nearly every data analytics program, whether built in-house or run by a third-party firm, draws on four core methods. Each answers a different question, and together they form a progression from understanding the past to deciding what to do next.
Descriptive analytics provides the historical baseline — usually delivered through dashboards, charts, and reports. A restaurant group might use descriptive analytics to see exactly how much free or discounted food a given manager authorized over the past quarter. It tells you what occurred, but not why.
Diagnostic analytics goes a layer deeper, using correlation and root-cause analysis to explain the "why" behind a pattern. If one manager's shift shows an unusually high rate of comped meals, diagnostic analysis might reveal that shift is chronically understaffed, so employees work longer hours and receive more meal compensation — a staffing issue, not necessarily a fraud issue.
Predictive analytics uses historical patterns, statistical modeling, and machine learning to anticipate what's coming. A retailer might know that returns spike every January after the holidays. Predictive models help distinguish a normal seasonal pattern from an emerging fraud trend, such as employees exploiting the high-return period to process fake refunds or fraudulent gift cards.
Prescriptive analytics is the most advanced layer — it doesn't just forecast an outcome, it recommends (or triggers) a specific response. Consider an outdoor retailer with a generous one-year return policy. If data shows a surge of returns on an older product just before a new version launches, prescriptive analytics could recommend adjusting the release timing or return window to prevent customers from effectively "renting" gear for free.
How The Four Types Work Together
These four types aren't independent tools — they build on one another in sequence, and the strongest loss prevention programs move fluidly between all four.
Take employee-driven shrink as an example. Descriptive analytics might show that a particular store's inventory shrink jumped 15% quarter over quarter. Diagnostic analytics examines POS exception data, access logs, and shift schedules to determine that the increase correlates with a new overnight crew and a lapse in dual verification for voided transactions. Predictive analytics then flags other locations with similar staffing and control gaps as at risk of the same pattern. Finally, prescriptive analytics recommends a fix — reinstating manager approval for voids over a certain dollar amount or adjusting scheduling — and estimates the shrink reduction that action should produce.
Skipping a step in that chain is a common reason LP programs plateau: teams that only build descriptive dashboards can see that shrink is rising but never learn why, while teams that jump straight to predictive modeling without solid diagnostic groundwork risk chasing the wrong root cause.
Reducing Fraud & Strengthening Loss Prevention
Shrink remains one of the most expensive line items retailers manage — but the more useful story in the data right now isn't the total dollar figure; it's where that loss is shifting.
According to the National Retail Federation's "Impact of Theft & Violence" research, produced with the Loss Prevention Research Council, retailers reported an 18% year-over-year rise in shoplifting incidents in 2024. The following year told a different story: NRF's follow-up 2026 report found shoplifting incidents declined 12.4% and merchandise theft dropped 8.1% in 2025—even as phone scams, digital and e-commerce fraud, gift card fraud, and loyalty fraud continued climbing. In other words, physical shoplifting is trending down, but the fraud that's harder to see on a sales floor is trending up. That shift is precisely why analytics-driven detection matters more than ever: a program built only to catch a shoplifter walking out the door misses the fraud increasingly happening through a screen, a return desk, or a loyalty account. A capable third-party analytics firm can identify a range of fraud patterns that are difficult to catch through spreadsheet-based reviews alone, including:
- Unauthorized "sweetheart" discounts given to non-eligible customers
- Full-price refunds processed on items originally purchased at a discount
- Employees applying their own loyalty account to customer purchases
- Customers creating multiple accounts to stack discounts or promotions
- Chargeback abuse following false claims of transaction errors
- Gift cards purchased with stolen credit cards and resold
- Refunds routed to a different card than the original purchase
Modern LP technology has also changed how quickly it detects these patterns. Rather than generating more alerts, the strongest AI-driven programs in 2026 focus on exception-based investigation that ties video directly to transactional data. Staffed, licensed security personnel, combined with that technology, have been associated with meaningfully lower organized retail crime incident rates compared to locations without staffed security. A third-party analytics partner brings both: the technical infrastructure to run exception-based reporting at scale and trained analysts who know how to interpret what the data shows.
Choosing the Right Third-Party Analytics Firm
Not every analytics vendor is built for loss prevention specifically, and the difference matters. A few things worth evaluating in a potential partner:
- Context, not just raw data. Some vendors will hand over a data feed or administer your POS reporting tool. A more valuable partner contextualizes and analyzes that data for you, connecting it to root cause and recommended action—not just a dashboard.
- Focus on what matters most. The right firm helps you prioritize the metrics tied to your actual risk profile, rather than selling a broader analytics suite you don't need.
- Demonstrated expertise, not a bolt-on service. As data analytics has become table stakes across the security industry, some firms have added it to their offering without deep in-house expertise. Ask about their team's background, request real case studies (not hypothetical models), and push for specifics on how their methodology works.
- A right-sized start. A strong partner will let you start with a focused pilot tied to a clear, measurable outcome before expanding the relationship — rather than pushing a large, all-at-once proposal.
- Transparency as a shared value. The firm you choose should hold the same standards of transparency and data ethics with your customers' and employees' data that you'd want to uphold yourself.
This is where the Zellman Group has built its reputation. As a dedicated loss prevention and asset protection firm serving retail, food and beverage, and hospitality, we provide data analytics and Exception-Based Reporting (EBR) services purpose-built to protect your bottom line. We don't retrofit generic business intelligence tools for loss prevention — our analysts work within your existing EBR platform (or help you implement one) to build a long-term strategy grounded in your operational reality, not industry averages.
Beyond analytics, Zellman's integrated approach spanning OSINT, organized retail crime investigations, civil recovery, and consulting ensures the insights our analysts surface don't stop at a dashboard. They connect directly to case resolution and recovery. With decades of experience across retail, food, and hospitality, we help clients move away from manual, spreadsheet-driven reviews and toward a data-informed loss-prevention program built on trust and measurable results.
Ready to see what third-party analytics can uncover in your business? Contact us today to learn more.
