We Rejected Them Correctly And Explained It Wrong

Every product we turn down gets a record with the reason printed on it, gathered on The Bar. The verdicts on those records were right. Some of the reasons were not.

One rejection said a product was turned down for a confidence score below 60. We do not have a rule like that. Sixty is the score a product needs to earn the Labs Verified badge. The score it needs to appear on SleekNova at all is 35. We published that distinction as a rule of ours in a decision log earlier the same day, then found we had been printing the badge number as a rejection line on a live record. Sixty one products on this site right now score below 60 on confidence. Every one of them would have failed the reason we published.

Three records gave the floor as 40 when it is 35, and one said a product was rejected for scoring 35 when 35 is not below 35. Two more used the badge line to explain the publication line a second way, once to say a product could not reach a passing score and once on the page that exists to explain the badge itself.

None of this changed a verdict. Each of those products was correctly turned down, and the real reason was already sitting in the body of its own record: a trust score under the floor, a weighted total under 8.0, a fulfillment history we could not stand behind. What was wrong was the one line a reader is most likely to read.

One of them turns out to be more interesting than a wrong number. The bath sheet we turned down failed none of our three gates. Its confidence cleared the floor, its trust cleared the floor, and we said no anyway, because customers kept reporting orders that did not arrive. The false line about a gate of 60 had been standing in for a judgment we had not written down. Its record now says what actually happened, and The Bar now says that one rejection in the record names no gate at all, rather than promising to fix a reason that was never missing.

While correcting those, we read the reason line on all 198 rejection records. A further set were not false but were broken: a bracket that never closed, a verdict printed twice in one sentence, a less-than sign showing as raw code, and one record whose reason was simply blank, so its card gave no reason at all. Those are now written the way the rest already were.

The lesson is not the arithmetic. It is that a number can be correct in one place and become a lie when it is carried somewhere else. Sixty was always the right number for the badge. It only became false when it was copied onto a rejection.

The rule it should have matched has been published on our methodology page the whole time.