How SaaS Flags works
An honest look at how we gather complaints, score risk, and decide what makes it onto a product page.
Where the warnings come from
Every warning on this site is paraphrased from a public source: a Better Business Bureau filing, a Trustpilot or G2 review, a Reddit thread, a Capterra rating, a court record, or a news article. We never reproduce reviewer text verbatim — that's both an attribution problem and, frankly, lazy. Instead, our scanner reads each source, extracts the substance, and rewrites it in our own editorial voice with the source URL attached so you can verify it.
The Risk Score
It's a 0–100 number. Calculated as the weighted sum of every documented warning:
- Severity weights: Low = 3, Medium = 8, High = 16, Critical = 28.
- Category weights: Lawsuits (×1.4) and misleading marketing (×1.2) carry the heaviest multipliers. BBB filings, billing problems, and user complaints all sit between ×0.9 and ×1.1. Churn signals weigh lighter (×0.7) since they're less indicative of harm per se.
- Diversity bonus: +4 points per distinct category present, because a product with problems across five different categories is materially worse than one with five complaints all in a single category.
The score is capped at 100 and rounded. Lower is better.
Risk bands
- 0–14 — Clean Record: No meaningful complaints have surfaced.
- 15–34 — Mixed Signals: Some issues exist, mostly minor. Worth knowing about, not worth panicking.
- 35–54 — Caution: A non-trivial pattern of complaints. Read before you swipe.
- 55–74 — High Risk: Multiple severe issues. Many users walk away frustrated.
- 75–100 — Severe: The complaint record is bad enough that you should know exactly what you're signing up for.
What we don't do
We don't accept payment to remove, downplay, or reorder warnings. We don't take affiliate kickbacks from the products we profile. We don't reproduce reviewer text verbatim. We don't claim a product is a scam unless that claim has been formally adjudicated — instead, we paraphrase what the public record shows and let you decide.
What the AI scanner does (and doesn't)
The scanner is an LLM-driven pipeline that pulls together publicly available information about a product from across BBB, Trustpilot, G2, Capterra, Reddit, news outlets, and forums — then classifies what it finds into categorized warnings. It paraphrases. It cites sources. It does not invent complaints out of thin air, and it does not copy text verbatim from any source.
Every page on this site is written to be read by a human, not to game search engines. If you ever land on a page that doesn't tell you something useful about whether to buy a product, please tell us and we'll fix it.