Draft. Not legal advice. Counsel review pending.

How Our Early Ratings Were Seeded

Beta draft -- Last updated September 2026

Tried launched with a catalog problem: a consumer product database is only useful if it has coverage. Nobody wants to rate products on an empty platform. But an empty platform stays empty.

This page explains what we did about that, and why we're being transparent about it.

The cold-start approach

At platform launch, Tried populated the catalog with synthetic ratings generated by generative reviewers -- AI-driven personas trained to simulate credible consumer behavior across a range of product categories. We call this the "sim layer."

These synthetic ratings bootstrapped the catalog so that the first real users would arrive to a useful, browsable product database -- not a blank slate.

How sim ratings are handled

Sim-generated ratings are:

• Stored with a sim flag in the database, distinguishing them from ratings made by real, verified users.
• Included in aggregate rating displays during the bootstrap period. We made a deliberate choice to show them alongside real ratings rather than hide the catalog entirely -- with the understanding that we'd disclose this publicly (which is what you're reading now).
• Not attributed to real users. Sim ratings appear under synthetic handles, not real accounts. No real user is misrepresented.

The phase-out plan

As real users join and rate products, sim signal is progressively weighted out of aggregate scores. Products with sufficient real-user coverage will effectively have their sim baseline replaced by genuine community opinion.

The long-run goal is a platform where sim ratings represent zero percent of displayed scores. We expect that for popular categories, sim influence drops to negligible within the first few months of real user activity.

Nuclear option

If at any point the sim layer creates a trust problem -- whether perceived or real -- we have a "nuclear purge" available: a migration that zeroes out all sim-flagged ratings across the database and recomputes all aggregate scores from real-user data only. This would result in many products showing no rating until real users covered them, which we consider an acceptable tradeoff for trust.

We will run the nuclear purge if community concern warrants it, if a regulatory environment requires it, or if we determine that the sim layer is doing more harm than good.

Why we're telling you this

Because we think it's the right call.

Consumer trust is the whole product. If Tried's ratings turn out to be gamed -- even by us, even with good intentions -- the platform is worthless. The alternative to this disclosure is hoping nobody notices, which is not a foundation we want to build on.

Sim-seeded content at launch is a known pattern in marketplace and review platforms. It is rarely disclosed. We're disclosing it because users deserve to know, and because we believe a platform that's honest about its own mechanics is more trustworthy than one that pretends its early data came from nowhere.

Questions

If you have questions about the sim layer or want to know more about how a specific rating was generated, email hello@sugarcap.com.