From a discipline tool to a pre-trade briefing

An early idea behind RiskFloor was friction on purpose: capture what a trader actually thinks before the system shows its analysis. The description of the product kept moving as I tried to say what it was.

RiskFloor started as discipline software, built on a bet from my 18 years of trading: traders lose money through broken process, not missing information. Traders saw the key levels, but had to write down their own view before the full analysis (the bull and bear cases and their sources) would unlock. By July 2026, early accounts had run 14 briefings and recorded 0 views. In my own notes, it had become "a form with research attached" instead of "research with an optional form attached."

• Keeping the gate. I set a lower bar instead: open the evidence, offer a choice, show a sample. The gate was also only on screen. The data was already in the page.
• "Risk manager" as the name. I dropped it, because it implies exposure monitoring and intervention the product does not perform.
• Treating the record as a footnote. I ruled this out. Leading with the briefing only holds while a trader's written view never feeds the analysis and recording it keeps equal weight.
• Two equal buttons on the homepage. Right inside the product, wrong for a stranger: a cold visitor can't yet tell the two apart.

I removed the gate and replaced it with two equal choices, "Record my read first" or "Show me the analysis" (July 2026). Then I repositioned the product around the briefing itself. I rewrote the homepage to lead with it ("Bull case. Bear case. Key levels. The invalidation.") and pointed the main button at a free live briefing instead of a sign-up form.

These were two decisions with different reasons, and I pressure-tested both with several AI models as critique partners. When the gate became optional, the record was still positioned as the differentiated product. By August 6 I had ruled it "secondary machinery". The homepage headline moved too: six earlier versions between July 21 and the end of August before this one.

Since August, 9 people went straight to the analysis and 4 opened the record step. The numbers are small, so I'm not claiming a lift. To learn why people stopped, I'm running a feedback study with every signed-up user this month.

A homepage's job changes when the product changes. Before free briefing pages, it had to explain the product well enough that a stranger would sign up blind. After them, its job was to send a visitor into a briefing.

I built the friction on purpose, then took it out and said why. Nothing I have shows it changed how traders behave, across about 10 recorded views in production. A principle can justify a test, but it can't stand in for one.

Test the gate with a clickable prototype and a handful of traders before building it. I checked the idea against my own experience and research from other fields, not against the people who would use it.