Summary
On 2026-05-12, Fred posted a "FLASHING RED LIGHT" finding in #direct-from-fred: Claude has now analyzed the entire MLS and identified listing description quality as a keystone driver of price outcomes. Fred's legacy listings ($200K+) score 95/100 on the new rubric; the MLS average is ~52/100. Sellers don't know this lever exists. This is positioned as a tool-build trigger.
The Numbers
- Fred's score (legacy listings $200K+): 95/100
- MLS average: ~52/100
- Gap: ~80% relative quality differential between Fred's listings and the median MLS listing on description alone
Why This Matters
For years Fred had hunches — he was using Zillow's published "high-price words / low-price words" lists, going to maximum character count, banking on emotional storytelling. The 95-score wasn't accidental; it was the result of deliberate craft. But until now, even Fred only had hunches about how much that craft mattered.
Claude's MLS-wide analysis converts hunches into a defensible stat. That stat unlocks two distinct plays:
- Listing appointment differentiator — show the seller the score gap, then show what Artemis does differently. The 95-vs-52 spread is the conversion lever.
- Productized scorer — Fred flagged this as "gotta make a tool for it." A description scorer slots into the IRIS / Spark / Athena tool stack.
Tool-Build Trigger
Fred's language ("gotta have it, gotta remember it, gotta fix it, gotta make a tool for it") puts the description scorer in the build queue. Candidate placement options:
- IRIS pre-listing — agent runs the address, gets a CMA + a description score + suggested edits before listing
- Spark at build time — Spark already orchestrates the showcase page; could score the MLS description as part of the same pipeline
- Standalone (Athena/Delphi-class) — a dedicated description scorer with input box + score + rewrite suggestions
Decision pending. The screenshot Fred attached (F0B36AGS78D, 247.3 KB, in #direct-from-fred 2026-05-12) likely contains the actual rubric — should be captured into Training & Playbooks before the build kicks off.
Open Questions
- What specifically is on the rubric? (Pending: download + transcribe the screenshot.)
- Does the 95 vs 52 spread hold at sub-$200K price tiers, or only $200K+?
- Where does this tool slot — IRIS, Spark, or standalone?
- Is there a per-bedroom / per-price-tier breakdown that would let agents benchmark against the right cohort instead of the MLS median?
Listing Appointment Use
Until the tool ships, the finding itself is a listing-appointment data point. Reframe the description conversation from "we'll write a great description" to "the MLS average description scores ~52/100; Fred's average is 95/100; that gap shows up in time-on-market and final sale price." Lands in the Proof beat of the listing-marketing-engine four-beat structure (Problem → Conventional Failure → Your Method → Proof).
Related
- listing-marketing-engine — Description quality is the upstream lever every listing ad amplifies; the new finding should land in the Proof beat
- iris-tool-suite — Most likely host for a productized description scorer
- spark-showcase-builder — Alternative host; already orchestrates the showcase pipeline
- seneca-adoption — Seneca can already coach on description rewrites; rubric integration TBD
- case-study-model — STAR framework benefits from the 95-vs-52 gap as a quantified "Result"