A first-order coupon can appear as customer acquisition even when someone claims it after receiving the same offer through another account. The discount is real; the acquisition may exist only in the report. For SHIELD, a provider of device intelligence, that difference supports a more specific commercial argument than a broad promise to reduce fraud: protecting a budget intended to buy a new customer relationship.
The company presents its promotion and referral abuse solution through connections between devices, accounts and codes. The proposition helps identify relationships that a different email address does not reveal. Its value depends on turning that signal into a proportionate decision, without automatically treating a shared device as a person breaking the conditions.
A new account does not establish that another customer has arrived
In a HappyFresh case study published by SHIELD , the grocery-delivery company uses device identification to connect several accounts with the same phone. This is a vendor’s commercial example, rather than an independent assessment of savings. It illustrates the change in the mechanism: a promotion is no longer examined solely within each account.
A merchant can recognise that several attempts to claim a benefit share technical signals. That can help review an invitation chain returning to the same environment or a sequence of registrations. The signal alone does not establish who is behind them, whether terms have been breached or which benefit should be withdrawn. Those conclusions depend on campaign conditions and other evidence.
The distinction matters to marketing. If the condition grants one incentive per customer while measurement counts accounts, spending can increase without expanding the buyer base in the same proportion. The review should start by defining the entitlement a coupon provides, how long it lasts and which exclusions the user can understand.
Incentives already appear among reported forms of abuse
The series compiled by Statista from the Global eCommerce Payments & Fraud Report 2026 shows that 25% of the professionals answering fraud-management questions identified coupon or discount abuse at their company. Loyalty fraud was reported by 24% and affiliate fraud by 22%. These are three ways a commercial benefit can fund behaviour other than the action it was intended to encourage.
Abuse also reaches commercial incentives
Professionals reporting that their company experienced these fraud types in the 12 months before the November–December 2025 survey. Three selected categories.
Tap or hover for details. All values are also available in the data table.
Collapse notes
Base of 634 fraud-management professionals, within an international survey of 1,278 participants across 37 countries. Multiple responses allowed. Self-reported experiences, not the share of fraudulent orders, losses as a percentage of sales or SHIELD outcomes. The report was published in 2026; fieldwork took place in November and December 2025.
Explore the data table
| Category | Respondents identifying each type |
|---|---|
| Coupon or discount abuse | 25% |
| Loyalty fraud | 24% |
| Affiliate fraud | 22% |
These figures describe how widely each problem was reported among respondents. They cannot estimate how much of an individual store’s promotional budget is lost or be transferred to Spain. To decide whether to buy technology, a merchant needs to calculate its exposure from its own redemptions and conditions, separating the planned discount from uses that breach the rules.
The three categories should also remain distinct. Repeated coupon claims, improperly earned points and an artificial affiliate commission are different events. They may share links between accounts and devices, but money leaves different budget lines and the person responsible for validating it may change. A shared prevention agreement needs reporting that can attribute each effect.
Commercial proof starts with what remains after the discount
SHIELD has a route into conversations with growth teams: demonstrate how a promotional decision changes margin as well as the number of alerts. Records should distinguish benefits granted, rejected uses, reviewed cases and legitimate orders ultimately completed. These are proposed criteria for evaluating the service, not a description of the company’s internal accounting.
A block may avoid the cost of an improper incentive. It may also prevent a valid purchase. The economic balance adds legitimate margin retained and improper spending avoided, then subtracts service and review costs and margin lost through mistakes. Counting only stopped attempts leaves part of that calculation out.
A before-and-after comparison is weak if it coincides with a change in offer, channel or season. Where feasible, comparable groups under the same conditions make it possible to observe changes in repeat redemptions and legitimate purchases. The purpose is to establish the additional effect of the control, rather than attributing every campaign change to it.
A device adds context without replacing the entire decision
Another Statista series, drawn from a Ravelin and Qualtrics survey , places two-factor authentication among the tools that 59% of participants considered most effective. Machine learning reached 48%, device identification 40% and IP analysis 37%. The survey covered large companies selling online in January 2025.
Device identification sits within a defence of several layers
Tools respondents considered most effective against fraud, January 2025. Four selected options; each figure represents a reported opinion.
Tap or hover for details. All values are also available in the data table.
Collapse notes
Qualtrics online survey commissioned by Ravelin among 1,466 fraud and payment professionals in ten countries, including Spain, in January 2025. Companies had more than USD 50 million in annual revenue or more than 500 employees and sold online. Multiple responses allowed. This is not a comparative test, a fraud-prevention success rate or a Spain-specific result. Ravelin is a commercial fraud-prevention provider.
Explore the data table
| Category | Respondents selecting the tool |
|---|---|
| Two-factor authentication | 59% |
| Machine learning | 48% |
| Device identification | 40% |
| IP analysis | 37% |
The comparison records professional perceptions, rather than a measured competition between technologies. For SHIELD, its commercial use is to show where device information fits alongside other signals. Authenticating access can help confirm control of an account; connecting several accounts asks a different question about entitlement to an incentive. Buying both capabilities is justified only when each addresses an identifiable problem.
A family may share a phone, computer or connection. A common technical pattern deserves context before becoming an automatic prohibition. Additional review, an eligibility check or a route to challenge a decision may be more appropriate responses than stopping the entire transaction. The cost of that friction also belongs in the assessment.
The criterion for expanding SHIELD in promotions is therefore specific: grant the benefit to buyers who meet the conditions, limit repeated claims that breach them and explain the difference. A system that reduces alerts without retaining valid customers has yet to demonstrate better acquisition. The proof is the legitimate business remaining after the incentive and its controls have been paid for.
In this article
Marian Vasile
Marian Vasile is a MarTech and growth professional. His work spans product marketing, SEO, paid advertising and deep online tracking, including behavioral analytics and attribution.
