How Clamp builds estimates: Researched-rate methodology
Every number Clamp publishes can be traced: what work was assumed, what rate was applied, when it was observed, and how the math runs. This page is the trace.
Claims
Clamp’s researched baseline rates are documented in the published bath-5x8/v1 snapshot.
- A supplier bid
- A site inspection
- A guaranteed price
Inputs
- Scope
- dimensions
- assembly versions
- researched rate observations
Immutable reviewed snapshot bath-5x8/v1, published 2026-09-07; rate observation window 2026-05-16 to 2026-09-05.
Quantities
- Your description supplies the parameters; assembly takeoffs compute the rest. "5 by 8 foot bathroom" becomes 40 sf of floor and a computed surround area; "replace the tub" selects a tub set with one unit; each line item's quantity derives from those parameters with its arithmetic stated (net area, waste on material rows, rounding rule). Quantities in published examples show their derivation, so a reader can check the math, not just read the total.
Method
- 1. A fixed assumption frame.: Each rate is priced against the same stated scenario: mid-market residential contractor, no specific locality, standard occupied property with truck access, typical remodel conditions. Rates are comparable because the frame does not move.
Input: Mid-market residential contractor, no specific locality, standard occupied property with truck access, typical remodel conditions.. Output: Rates are comparable because the frame does not move.. Caution: The frame does not move. - 2. Three separately executed research runs.: The work is priced separately by three AI research systems, each pricing the same stated scenario rather than reviewing one another's output. Their cited evidence may overlap, so separate runs are not independent market observations.
Input: Current market evidence.. Output: Separately executed runs; cited evidence may overlap.. Caution: Not one model checking itself. - 3. Reconciliation.: The spread between sources is examined and reconciled into one baseline rate. Where the sources disagree, the reasons they cite are kept as the rate's variance drivers: the specific jobsite realities that move this price (wall composition, metro labor premiums, batch size, access).
Input: The spread between sources.. Output: One baseline rate.. Caution: Keep the rate's variance drivers. - 4. Grading.: Every rate records a confidence level and a volatility grade. Volatile rates (lumber, fuel-exposed work) get re-observed sooner; stable labor rates hold longer. Nothing is undated.
Input: Confidence level and volatility grade.. Output: Nothing is undated.. Caution: Volatile rates get re-observed sooner; stable labor rates hold longer. - 5. Versioning.: When rates are re-observed, published examples version forward with a changelog; the old version stays downloadable. A year label is never updated without the rates underneath it being re-examined.
Input: Re-observed rates.. Output: Published examples version forward with a changelog; the old version stays downloadable.. Caution: A year label is never updated without the rates underneath it being re-examined.
Traced row
Bathroom Tile Install Labor: From snapshot bath-5x8/v1: 93 sf at $12.50 per sf for labor.
Published row bathroom-tile-install-labor in immutable snapshot bath-5x8/v1.
Quantity: 93 sf. Derivation: (28 floor sf + 65 surround sf) × 1 difficulty × 1 standard-format factor = 93 sf. Net 93; purchased 93; waste 0; deductions 0; rounding: No additional rounding.
Published quantity contract for bathroom-tile-install-labor.
Rate: $12.50 per sf labor. Observation window 2026-05-16 to 2026-09-05; row observed 2026-05-16.
Published component rate and row provenance for bathroom-tile-install-labor.
Assumption frame: typical project scale, complexity, access, and occupied baseline conditions. Assumptions include: Existing floor joists and subfloor are structurally sound with no water damage; All rough plumbing and electrical connections remain in current locations; Wall framing and drywall behind tile are intact with no concealed rot or mold; Electrical circuits have adequate capacity for new fixtures without panel upgrade; Standard 8-foot ceiling height and no concealed structural obstructions; Transcript did not specify a finish tier (basic/standard/premium); assumed standard given no descriptors of high-end or budget materials.; Attic or ceiling cavity access exists within 10 feet of fan location; Electrical circuit capacity exists for fan load without panel upgrade; Exterior wall or soffit is accessible for termination within 10 linear feet of fan; Ceiling framing is standard 16-inch on-center joists with no obstructions
Published baseline conditions and assumptions in bath-5x8/v1.
Math: 93 × $12.50 = $1,162.50 → Tile trade subtotal $6,050.63 → direct job cost $19,867.38.
Published row total, trade subtotal, and direct cost in bath-5x8/v1.
Scope selection, prose accuracy, arithmetic and provenance mapping reviewed; not independent market-rate verification.
Validation
documented
- Quantity derivation
- Component reconciliation
The published bath-5x8/v1 snapshot contains 51 rows and a direct job cost of $19,867.38.
Limitations
- Modeled examples are not bids, completed-job records, or national averages.
- Rates are researched baselines with stated windows; local quotes vary.
- Concealed damage, permits beyond stated scope, and unstated finish tiers are excluded unless itemized.
- AI-assisted editorial checks are recorded as such; they are not independent market validation. A named human reviewer approves every published example.
- Download the immutable bath-5x8/v1 snapshot: /data/estimates/bath-5x8/v1.json, /data/estimates/bath-5x8/v1.csv, and /data/estimates/bath-5x8/v1.dictionary.md.
- Your result is a new draft, not a reproduction of this frozen example
- Published bath-5x8/v1 on 2026-09-07 after signed owner approval.
Questions
Can I reproduce a published example?
Yes: the snapshot holds every quantity, rate, and calculation step; the estimator accepts the same scope description.
Why do two runs differ slightly?
Extraction reads natural language; assumptions are listed on every draft for review.
Do you publish your whole rate library?
No. Published examples disclose their own rows fully; the library, formulas, and pipeline stay private.
Sources & provenance
- 5×8 Bathroom Remodel v1 reviewed snapshot Clamp Research (owner-supplied)