2,133 products trackedOwner-verified regret scoresServing the US and CanadaNew: Amazon Associate affiliate integration liveDay 30, 60, 90 satisfaction data
2,133 products trackedOwner-verified regret scoresServing the US and CanadaNew: Amazon Associate affiliate integration liveDay 30, 60, 90 satisfaction data
How it works

The Regret Score, explained.

A single number, 0–100, that tells you how many owners regret their purchase. It's calibrated on day-30, day-60, and day-90 satisfaction — long after the honeymoon.

12
Regret
Low regret
Low regret

Owners are still happy three months in. Safe buy — sleep easy.

47
Regret
Some regret
Some regret

Mixed long-term reviews. Read the specific complaints before you commit.

82
Regret
High regret
High regret

Most owners wish they hadn't bought it. We'll show you what to buy instead.

Why day 90 matters most

Most product reviews on Amazon.com, Amazon.ca, Best Buy, and Walmart are written in the first 48 hours — when owners are still riding the dopamine hit of unboxing. That's not the moment you want to hear from them. The truth about a product surfaces later: after the return window closes, after the subscription cost stacks up, after the battery starts to degrade. Pregret's Regret Score weights day-90 satisfaction three times as heavily as day-30 because that's where reality lives.

The formula (no black boxes)

For every product, we take verified owner check-ins at day 30, 60, and 90 and compute:

regret_score = (weighted_low_ratings ÷ weighted_total_ratings) × 100
where day-30 weight = 1×, day-60 = 2×, day-90 = 3×

A "low rating" is 1 or 2 out of 5. A product with mostly 5-star day-30 ratings but mostly 1-star day-90 ratings will score high on regret. A product that stays 4+ across all three checkpoints will score low. The math is public and reproducible.

AI-estimated vs owner-verified

Products marked AI est. don't have enough real owner data yet — their Regret Scores are calibrated from publicly available review sentiment (Amazon 1-3 star themes, Reddit threads, expert reviews) using Claude, weighted by the product's Amazon star rating and review volume. As real owners contribute check-ins, the AI estimate is gradually replaced by verified data.

Why some products recommend alternatives

When a product has a high Regret Score, we surface 2–3 lower-regret alternatives in the same category. These are curated based on owner data — the products people who returned the regretted item actually switched to.

A word on affiliate links

When you click through to an alternative and buy it on Amazon, we may earn a small commission at no cost to you. That's how Pregret stays free. Commissions never influence Regret Scores — we don't get paid more for surfacing higher-priced products, and there's no path to bump a product's ranking through payment.

Serving US and Canadian shoppers

Regret data is aggregated across shoppers in both the United States and Canada. Amazon.com and Amazon.ca often ship the same product with slightly different pricing, warranty, and return terms — Regret Scores are consistent because owner satisfaction is what changes slowly, not the retailer.