How to read an Adoptability score honestly.

First, the headline: the Adoptability score is a COMPARATIVE measure, not an ABSOLUTE predictor. An 80 is not "80 percent likely to succeed." A 65 is not "below the line." The number's job is to let you compare products against each other on dimensions that matter for adoption, given an operator with normal distribution access.

The 10-axis page at /factory/adoptability/ walks through what each axis measures. This essay goes one level deeper: how to read the same score very differently if you are the right operator for the product versus the wrong one.

The four reading modes

Same score, four different reader contexts, four different correct conclusions.

Your context	What a 70 means	What a 80 means
You have the named ICP in your existing audience	Strong candidate. The catalog-average score is below this. The few axes pulling it down are probably tractable for you specifically.	Near-mandatory read. The catalog flagged this as top-tier; with your distribution it is plausibly the best ROI in our index.
You have a different vertical but operator skill	Worth a Phase A conversation. The structural research is solid; you need 5 customer-validation calls to confirm the ICP transfers.	Same. The score's ICP specificity may not transfer. The methodology will.
You have no audience yet but skill exists	Read the risk memo first. A 70 with weak distribution-difficulty axis takes 18-24 months to build the audience. That is fine if you have the runway.	Same caveat. An 80 with weak distribution is still 12+ months of audience-building before the score's "easy distribution" math applies.
You are evaluating ideas competitively (VC, accelerator, FTE)	The score is most useful here. Filter to 70+, sort by year-1 take-home, look for products where the buyer-clarity axis is 9-10.	Probably already on your radar. Read the honest disclosures: 0.4 percent of catalog graduates to its own domain in six weeks.

Same number. Four reader contexts. The "right" answer is contextual, not absolute.

The three axes most operators misread

"Distribution difficulty (easy)" - read it as your access, not the channel

This axis rates how findable the buyer is. A high score (9-10) means "the buyer congregates somewhere reachable without paid ads." But that is the abstract version. The version that matters for YOU: do YOU have access to the place where they congregate? A 9 in "demand-gen managers at B2B SaaS" is not a 9 if your network is all DTC e-commerce. A 7 in "trucking dispatchers" is a 9 for someone running a Pi-based fleet hardware company that already has 50 of them on a Slack channel. The score reflects the universal accessibility; your operator-specific score may be very different.

"Credibility of claims" - it scores OUR research, not your authority

A 9-10 on credibility means the dossier's market claims are anchored in public benchmarks or interview data. It does NOT mean the buyer will trust YOU when you sell to them. Authority transfer is a separate problem the score can't solve. If you have a vertical-specific reputation (e.g., 10 years in roofing, a podcast for accountants, a Substack for compliance officers), you'll convert better than the score implies. If you are entering a vertical cold, you'll convert worse.

"Financial upside" - read with the take-home, not the ARR

This is the axis where most buyers misread the catalog. A product with $300k Year-1 ARR mid-case can have a $-15k Year-1 take-home after build cost and operating expense. The financial-upside score reflects the take-home, not the ARR. A 9-10 here is RARE on purpose. Most catalog products are 1-4 on this axis. That is not a failure; that is what bootstrapping a SaaS in Year 1 actually looks like. The score tells you "Year 1 will pay the bills" or "Year 1 is a loss leader for Year 2-3 scale." Both can be valid; just know which one you signed up for.

What a tier really means in practice

The /factory/adoptability/ page summarizes the tiers (80+, 65-79, sub-65). Here is the operator-honest version with what to actually expect at each tier:

Tier	What the score reflects	What you should actually plan for
80+	At least 4 axes at 9-10, no axis below 5. Catalog top-tier.	6 to 9 months to first 50 paying customers if you have the named distribution. 12 to 18 months if you have to build the audience. Bookkeeper-ai is the current 80; one product out of 238.
75-79	Strong axes across the board, usually one specific risk named in the Fermi summary (e.g., a free incumbent, a long sales cycle).	9 to 12 months. Phase A engagement is well-suited to this tier; the customer-validation calls in week 1 confirm the named risk is or is not a dealbreaker for your specific situation.
70-74	Solid hypothesis. Usually 2 to 3 axes flagging tractable concerns.	12 to 18 months. This is the meat of the catalog. The 70-74 tier is where operator-specific advantage (your audience, your authority, your funded runway) matters MORE than the score's universal measure.
65-69	Hypothesis with named issues that need work before launch. Often financial-upside is 1-3.	Worth Phase A only if you are running this as a strategic add-on to an existing business. As a standalone bet, the math is thin. Read the year-1 take-home carefully.
Under 65	Sketches, early-stage hypotheses, ideas the studio surfaced from market signal but did not develop fully.	Worth reading for category awareness. Worth Phase A only if you have specific reason to believe the studio missed something. The score is honest about the gaps.

Three ways to read the score that get operators in trouble

Mistake 1: treating "score" like "probability of success"

If you read an 80 as "80 percent likely to make me money", you will be disappointed. The catalog explicitly attaches a separate probability_of_meaningful_success number to every product (typically 5-25 percent). That number is the honest one for outcome estimation. The Adoptability score is the comparative ranking, which is a different measurement.

Mistake 2: filtering only by score without reading the Fermi summary

Every product's Fermi summary names the SPECIFIC risk that drove the lower axes (e.g., "HubSpot is shipping this same feature free to existing users" or "EagleView already has 80 percent of the contractor market"). The score reflects that risk; the Fermi summary names it. Reading score-without-Fermi is reading the conclusion without the reasoning. The two-second filter is fine for casual browsing; the 30-second read is required before unlocking a dossier.

Mistake 3: assuming polish equals fit

The 26 hand-polished product pages in the catalog have higher landing-page-quality axis scores (because they're polished). They do not have higher probability-of-meaningful-success. Polish reflects how WELL the page presents the idea, not how well the idea will work for you. A scrappier-looking page can still be the right idea for your specific access. Use the score, not the polish, as the comparative signal.

What to do with the score on your next browse

Concrete workflow:

1. Filter the catalog by your existing distribution access. Whatever vertical or buyer profile you already reach, filter to that. Use category, audience, or search.
2. Sort by Adoptability score within that filter. The top 3 in your accessible vertical are your best comparative reads.
3. Read the Fermi summary on each top-3. Look for the named risk. That risk is what Phase A would test.
4. Check the year-1 take-home for the math you actually need. If take-home is negative, you are signing up for a Year-2 scale play. If it is positive, you are signing up for a profitable Year 1.
5. Look at the probability_of_meaningful_success. 15 percent is normal. 25 percent is high. Below 10 percent is "you need an unusual edge."
6. Then unlock the $5 dossier. The dossier has 16 sections including the risk memo and 7 likely objections handled. If the dossier does not feel like it matches the score's promise, refund and move on. We would rather you walk away early than struggle through a wrong idea.

This workflow takes 5 to 10 minutes per product. It filters out 80 percent of the catalog quickly, then lets you spend real reading time on the 3 to 5 candidates that fit your specific situation. The score is the COMPARATIVE filter that makes this efficient.

The score itself, for the record

The Adoptability score is mechanically defined. It is the equal-weighted average of 10 axes, each scored 0-10. The axes are loaded from /factory/adoptability.json for every product. The formula is in /home/ubuntu/factory/director/adoptability-score.py on the studio's cron schedule. Re-scoring happens every 4 hours.

The formula does not change product to product. The axes do not get bonus weights. The scoring is heuristic but consistent, which is what makes the comparison meaningful even though the absolute predictive power is limited.

The honest framing: this number is the structural-credibility signal across the catalog. Use it as such. Layer your operator-specific judgement on top. Read the Fermi summary for the qualitative risk. And remember the actual probability-of-success number is much lower than the score, which is true of every operator launch ever.