The CovetIQ standard
Methodology
Scoring Engine v1.0 · Editorial review August 31, 2026
CovetIQ combines reviewed editorial judgment with documented product and buying-path evidence. Rankings and explanations are deterministic: an AI model does not invent product scores or recommendations.
The eight dimensions
| Dimension | What it evaluates |
|---|---|
| Recognition | How readily the design and fashion house are identified in the US luxury market. |
| Visual luxury presence | Materials, silhouette, finish and recognizable house codes. |
| Resale strength | Directional secondary-market retention and demand, considered with evidence confidence. Never a return guarantee. |
| Timelessness | Design continuity and longevity beyond a particular trend. |
| Versatility | Use across wardrobes, occasions and settings. |
| Practicality | Capacity, access, closure, carrying options and everyday utility. |
| Value | The product proposition relative to its purchase-price tier. Separate from resale strength. |
| Trend momentum | Current interest, with lower confidence when fresh evidence is limited. |
Your budget is a maximum
In the evidence-led finder, purchase paths over your stated maximum are excluded before scoring. A less expensive eligible bag can legitimately rank first. The $5,000+ choice asks for your actual maximum; it does not assume an arbitrary ceiling.
New and pre-owned are separate buying paths
New eligibility requires a reviewed official US price and actionable availability. Preorder is labeled. Contact-to-order is conditional, not a claim of immediate stock or online checkout. Out-of-stock, unknown and pending availability do not qualify for New.
Pre-owned eligibility requires at least three usable authenticated-seller comparable observations after condition review and obvious-outlier exclusion. The budget check uses a derived typical-entry price, not the median or one unusually cheap damaged listing. Selecting Either ranks eligible New and Pre-owned paths separately; it does not automatically choose whichever is cheaper.
Current resale evidence is family-level and based on listed asking prices, not completed sales. It does not establish current-SKU availability, realized resale returns or an exact-item offer. Confirm authentication, condition, final price and availability directly with the seller. CovetIQ does not authenticate products.
Different goals, different priorities
Looks Most Expensive primarily considers visual luxury, then recognition, with smaller contributions from timelessness and trend. Practicality, value and resale do not affect that goal's score. Best First Luxury emphasizes lasting style, versatility and daily practicality. Best Value uses Spend Smart principles; Best Resale requires at least Medium resale-score confidence.
Quiet Luxury additionally considers reviewed logo prominence and insider recognition. Conspicuously branded options are excluded, and lower-logo designs can benefit. Eligibility and available evidence still limit the selection: a relative Quiet Luxury fit is not necessarily a logo-free design.
The strongest match stays first. Modest brand diversity can reorder sufficiently close alternatives. Remaining ties use evidence confidence, supported intent match, lower cost for Spend Smart and data freshness. Match percentages describe editorial fit, not purchase probabilities.
Confidence and editorial perception
Lower-confidence ratings are gently pulled toward neutral before ranking. Missing required evidence excludes a buying path or goal; it is never replaced with a fabricated product fact. Missing nonessential supporting dimensions remain labeled Pending and use only a neutral internal baseline where relevant.
Recognition, visual luxury, value, logo prominence and insider recognition currently use Medium-confidence editorial judgments. These are not measured consumer-survey facts. Trend confidence remains Low without fresh supporting market evidence. Resale confidence varies by the available evidence; listed prices alone are not proof of retention.
Reviewed logo prominence runs from 0 (effectively logo-free) to 100 (branding dominates); insider recognition runs from 0 (hard for informed buyers to identify) to 100 (immediately identifiable). Those ratings are candidates for future first-party Face-Off calibration, not already calibrated findings.
Focused coverage, honest empty results
The evidence-led selection starts with 15 reviewed handbag families. Not all have a verified eligible path. A budget or buying condition may yield one model or none; we do not insert legacy or unverified products to fill the screen. Product observations retain source URLs and checked dates. They are not a live inventory feed.
Brand links are reference points, not merchant-offer feeds or pre-owned offers. We do not imply that a brand page sells the observed resale item.
Original finder fallback
CovetIQ retains its original finder for operational rollback. That mode uses a broader legacy catalog with directional price estimates and a limited budget tolerance, rather than verified v1 purchase paths. Its results explicitly identify prices as directional. The evidence-led controls and maximum-budget wording identify the v1 experience.
Face-Off data
Face-Off currently uses brand and model identity, without fabricated product photography. Votes are stored separately from editorial ratings. Future perception calibration must be reviewed and versioned; votes do not silently overwrite the current scores.