Kit Home Index › How we grade
How we grade, and what we refuse to grade
Rubric v1.0, August 2026. Three pillars, weights published below, and you can recompute any grade on this site by hand.
This grade covers what a maker publishes about buying, pricing and materials. It deliberately says nothing about whether the kit works on your land, because that depends on your address and not on the kit. Site fit is a set of pass and fail gates, run against your land in the Finder, and it is never averaged into anything. An average would let a good price bury the fact that a kit cannot carry your snow load.
The three pillars we grade
| Buyability, weight 25 | Can you place an order today, is there a published price at all, the lead time, and where they will deliver. A beautiful render with a waitlist and no price scores near nothing here, and should. Rated on 120 of 514 kits. |
|---|---|
| Cost honesty, weight 25 | Not cheapest wins. Cheapest is usually the most dishonest number on the page. This scores how completely the published price describes what you will actually pay, across fourteen build stages, and whether the exclusions are stated plainly. A maker publishing $49,000 that covers a third of the true cost scores below one publishing $180,000 to $240,000 all in. Rated on 160 of 514 kits. |
| Materials and carbon tier, weight 15 | A category level embodied carbon class, from the structural system. This is a property of the material class, not of the individual maker, and it is cradle to gate only, which means it ignores that a high embodied carbon envelope with a very good R value can win on lifetime carbon in a cold climate. We say that here rather than letting someone else say it for us. We do not print a per model kgCO2e figure, because almost no kit maker has an EPD. Rated on 408 of 514 kits. |
What we refuse to grade, and why
| Site fit | Never in the grade. Wind, snow, seismic, flood and soil requirements come from your address, not from the kit, so a kit cannot carry a site fit score around with it. It runs as pass and fail gates in the Finder and it is never averaged into anything. An average would let a good price bury the fact that a kit cannot carry your snow load, which is the exact failure IIHS designed conjunctive scoring to avoid. |
|---|---|
| Energy performance | We report the wall and roof R values a maker publishes and score nothing, because the honest comparison is against the IECC requirement for your climate zone, and that needs your address. We label published values as nominal where that is what we have, since makers rarely publish tested whole assembly values including thermal bridging. |
| A HERS index or a green certification | Never computed and never assigned. Both require accredited modelling and third party field inspection of a built, sited home. LEED, NGBS, Passive House, HERS and ENERGY STAR cannot be attached to a model from a spec sheet. Estimating one would be wrong often enough to destroy the only asset this site has. |
| Needs fit | Bedrooms, size, accessibility and whether it can be permitted where you are. That depends on you, so it belongs in the Finder rather than in a grade. |
Why so much of the Index is not rated
A pillar with nothing published comes back unrated, never zero. A maker who publishes nothing is unmeasured, not bad. If silence scored as a bad grade then the grade would be a completeness score wearing a costume, and any maker could correctly accuse us of punishing them for something we never measured. NHTSA shows "Not Rated" and says plainly that it does not mean unsafe. Morningstar withholds a rating entirely below thirty six months of data. We do the same: a kit needs at least two of the three pillars rated before it gets a grade at all.
| A | 20 |
|---|---|
| B | 39 |
| C | 40 |
| D | 32 |
| F | 21 |
| Not rated | 362 |
Completeness is reported separately on every page and never rolls into the grade. A well documented mediocre kit must never outrank an excellent quiet one, so the two are two different axes and they are displayed as two different things.
The rules behind the rules
- The scoring code cannot read an affiliate rate. It is enforced in the database schema and in the build, not in a policy document, so the promise survives careless code.
- No kit maker has ever paid for placement here, at any price, and none ever will. No sponsored slots, no paid ranking, no kit maker sponsorships.
- Kits owned by this site's owner are flagged as ours on every card, get no forgiveness for missing data, and can be hidden from the Index with one switch and no account.
- Every score carries a rubric version. When the rubric changes you will be able to tell whether a kit's grade moved because the kit changed or because we did.
- Nothing here is verified by us in a laboratory. It is what the maker published, cited and dated on every page. We take no manufacturer supplied claim as verified.
Membership
Membership is for the part after the shortlist.
The narrowing is free and always will be. Membership is for the part after you have your shortlist, when you need to act on it.
- Delivered cost by build stage, so you can see what the published number leaves out
- See it on your land. Your parcel with the footprint dropped on it to scale, setbacks checked
- The questions to send makers who publish nothing, written from their own blanks
Everything membership includes
$47 once, for life
No subscription. No kit maker has ever paid for placement here, at any price, and none ever will.
Become a member