Why 73% of Your Declined Orders Are Actually Good Customers
Evaluation and Summary
eCommerce Fastlane hosts Nicole Jass, SVP of Enterprise Strategy at Signifyd, to discuss a counterintuitive problem: approximately 73% of declined orders in Shopify stores are actually legitimate customers, meaning aggressive fraud prevention rules are costing merchants revenue from customers they've already paid to acquire. The episode positions this as a critical issue for Shopify sellers operating at various scales. Jass breaks down multiple fraud vectors beyond simple card declines, including card testing (where fraudsters probe stolen cards with small transactions), account takeover attacks, loyalty point theft, and return abuse—sophisticated threats that require more nuanced detection than binary approval/decline decisions. A key signal the episode highlights is recognizing when a store has outgrown Shopify Protect's default protections, particularly when handling cross-border orders or reaching $5M+ in annual revenue, at which point a dedicated fraud strategy becomes necessary. The episode uses Carbon38 as a concrete case study, demonstrating that by refining their fraud rules to decline fewer legitimate customers, they achieved a 99% approval rate and captured an additional 3% revenue lift—a tangible outcome that illustrates the financial impact of false declines. The framing suggests the episode is designed for Shopify merchants across different revenue tiers, from those doing $3M annually on Shopify Protect to larger operations running dedicated fraud teams. The promise is that listeners will gain actionable strategies they can test immediately, regardless of their current fraud prevention setup. The episode appears to position Signifyd's expertise as the lens through which these problems are examined, though the show notes don't specify whether Signifyd is positioned as a solution or simply as the guest's background context.
Verdict: A well-credentialed guest tackles a specific, high-impact problem for Shopify merchants with concrete examples and actionable signals, though the reliance on show notes alone prevents full assessment of delivery depth and pacing.
Business rubric
-
Actionable takeaways 4/5 20%
Shows notes promise 'strategies you can test this week' and mention specific signals (cross-border orders, $5M threshold) and a framework (spotting fraud types beyond checkout), though without hearing the episode, the depth of implementatio
-
Evidence over hype 4/5 20%
The 73% statistic anchors the episode, and Carbon38's 99% approval rate with 3% revenue lift provides a concrete case study; however, show notes don't explain the methodology behind the 73% figure or how broadly it applies.
-
Guest and host credibility 5/5 15%
Nicole Jass is SVP of Enterprise Strategy at Signifyd, a fraud prevention platform, giving her direct operational experience with this exact problem across multiple merchant accounts.
-
Substance 3/5 15%
Show notes outline multiple fraud vectors and a case study, suggesting solid content, but the 40-minute length and reliance on show notes alone make it difficult to assess whether the episode has filler or stays dense throughout.
-
Delivers on its title 5/5 15%
Title promises explanation of why 73% of declined orders are good customers; show notes directly address this claim with fraud detection nuance and a revenue-impact case study.
-
Value for time 3/5 15%
40 minutes is reasonable for a focused interview, but without hearing the episode, it's unclear whether the pacing is tight or if there are tangents; show notes suggest enough substance to justify the length.
Worth it if you run a Shopify store doing $3M+ annually and suspect your fraud rules are rejecting legitimate customers, or if you're evaluating when to move beyond Shopify Protect's default settings.
Content flags
- Sponsored Nicole Jass is SVP at Signifyd, a fraud prevention vendor; while the episode may offer genuine insights, her company has a commercial interest in merchants adopting more sophisticated fraud tools.
Based on the episode description. AI-generated; it can be wrong.