Not the Upvotes or the leaderboard ranking, I think.
I've seen products hit the top spot and the comments are just empty praise. Then there are launches where the founder is actually in the replies having real conversations.
Someone points out a bug and the maker listens. Someone asks a tough technical question and gets an honest answer.
Nobody is pasting robotic sales pitches in every thread.
For the two years I celebrated each increase in our email list. I refused to delete anyone because a large email subscriber count felt like an asset.
Eventually our email open rates fell from thirty percent to twelve percent. Our newsletters began landing in the promo and spam folders of people who wanted to read them.
Mail providers look at how many people ignore your emails to decide if you are sending spam. By keeping thousands of email readers we were telling Google and Apple that our email content was not worth opening.
We finally ran a re engagement campaign. Deleted six thousand email users who had not clicked a link in six months. It felt terrible to slash our email list size in half
We spent our first three months in beta completely free. The goal was simple: get people through the door, gather feedback, and build word-of-mouth. It worked. We were averaging about 150 new signups a week, and users were exploring every feature freely.
Last Tuesday, we finally introduced our basic subscription plan ($12/mo).
I give a text generator six different voices (a dry friend, a drill sergeant, a corporate manager, and so on). Every prompt carries shared rules: word count, how many emojis, where they go, never invent a number. I tested it end to end against a fake model server: 127 checks that the exact prompt reaches the model for every voice, all passing. That proves the prompt is assembled correctly. It proves nothing about whether the model follows it.
First real run, one sample per voice:
Emojis clumped. One reply put all 12 emojis in a single block at the end. In the short replies, three voices ended with 4, 5, and 6 in a row. The rule said never more than two together.
Under target. Four of six roasts used fewer than the 8 emojis I asked for.
A copied example. Each voice had a tone example marked never reuse it . One voice reused its closing line in both runs.
An invented number. It produced a ratio that wasn t in the input, under a rule that says use only numbers the user gave.
My own rules contradicted each other. Never repeat an emoji plus use 8 to 12 emojis can t both hold when two palettes had 7 and 6 emojis. The fake model never complained. Reading real output did.
What I changed: concrete placement rules (end of a sentence, about one every second sentence, spread across the whole text), tone examples as short fragments instead of full sentences so there s nothing to copy, a rule against computing new ratios, tighter length wording, and bigger palettes. I also wrote a preview script that checks eight rules on real output and prints ok or CHECK for each, using the same checking code my unit tests load. I broke each rule on purpose 42 ways, and every break failed a test.
