# Construction CFO firm evaluation: method (fixed 2026-10-31, before research)

Published study: https://constructioncfo.net/best-fractional-cfo-construction-subcontractors
Commissioned and published by SPM The Construction CFO, which is one of the firms scored. Not an independent third-party review.

## What "best" means
The firm a construction subcontractor or self-performing GC is most likely to get real financial results from. Findability is scored separately and is NOT part of "best".

## Firms (24)
SPM, CLA, Level, Civil CFO, Vessel Advisors, Salisbury Bookkeeping, Profitability Partners, James Moore & Co, Wiss, BPM, Remote Books Online, NOW CFO, Apparatus, LLUM, Local Fractional, CFO Advisors, Meru Accounting, Cherry Bekaert, Surety CFO, TGG Accounting, CBIZ (incl. Marcum), RedHammer, AmbitionCFO, Barbin Group.

## Rubric (each 0-5; every point needs a public URL)
- C1 Construction and segment fit (20%): 0 none; 1 construction listed among many; 2 dedicated but generic page; 3 construction practice with specifics (WIP, job costing, bonding); 4 construction-only with sub and/or GC focus; 5 construction-only + explicit trade-sub focus + stated range covering roughly $1M-$20M.
- C2 Evidence of client results (30%): 0 none; 1 testimonials without specifics; 2 qualitative case studies; 3 2+ quantified construction cases; 4 4+ quantified construction cases; 5 4+ quantified construction cases with named or verifiable clients.
- C3 Third-party trust (20%, points, cap 5): verified reviews across Google/Clutch/Trustpilot/G2/BBB (1-9 = 1, 10-49 = 2, 50+ = 3, minus 1 if average under 4.0); +1 construction industry ranking or award; +1 named team member with CPA or CCIFP.
- C4 Capability (15%): count of WIP, job costing setup, 13-week cash forecast, bonding support, bookkeeping, controller/AP-AR, monthly advisory meetings, construction software setup (0-1 = 1, 2-3 = 2, 4-5 = 3, 6-7 = 4, 8 = 5; 0 if no construction offering).
- C5 Transparency (15%, points, cap 5): pricing +2 (single range/starting price/hourly +1); scope per tier +1; onboarding timeline +1; stated revenue range or fit criteria +1.
Self-reported totals earn nothing in C2/C3. Anonymous case studies count as quantified, not verifiable.

best = 4*C1 + 6*C2 + 4*C3 + 3*C4 + 3*C5 (0-100)
subcontractor = best - 15 if the firm doesn't name subs, specialty or trade contractors as clients
GC = best - 15 if the firm doesn't name general contractors as clients
trade = 0.75*subcontractor + 25*(trade_fit/3); trade_fit 0 none, 1 listed, 2 page or real content, 3 trade case study with numbers
findability = 70*(queries in top 10 / 15) + 30*min(independent roundups, 3)/3; roundups written by an evaluated firm don't count

## Calibration pass (2026-10-31, after scoring, before totals, applied to all)
1. C5 pricing +2 when a specific package or tier price (or that tier's range) is visible; +1 for one overall range, starting price or hourly rate. Level, LLUM, Salisbury +1.
2. C2 counts construction-client evidence only. Wiss -1, CFO Advisors -1, BPM -2.
3. C3 credential point for any named CPA or CCIFP (public license lookup). RedHammer +1.
Raw agent scores are kept in firms.py and firms-evidence.json.

## Query set (one engine, standard mode, no location or login)
construction CFO; fractional CFO for construction companies; fractional CFO for subcontractors; outsourced construction accounting; construction bookkeeping services; best fractional CFO for contractors; fractional CFO for general contractors; CFO for electrical contractors; CFO for concrete contractors; CFO for civil contractors; accountant for HVAC contractors; bookkeeping for plumbing contractors; CFO for roofing contractors; construction controller services; outsourced WIP schedule for contractors.
Roundup searches: "best fractional CFO for construction 2026", "best construction accounting firms", "best bookkeeping services for contractors 2026" (up to 8 lists opened).

## Re-run (scheduled every 2 months, next 2026-12-28; each run adds the 4 most-seen construction finance firms not yet in the study and re-scores everyone)
1. Keep this file; date any rubric change and re-score everyone.
2. Re-check each firm from firms-evidence.json, score by the anchors, a URL per point.
3. Run the 15 searches the same day, record top 10.
4. Edit firms.py, run `python3 calc.py` -> scores.csv, trades.csv.
5. Publish the date, all changes, and corrections received.

## Correction log
- 2026-10-31: SPM C3 0 -> 1. Google blocks automated tools from opening Maps pages, so the research agents could not read SPM's Google count. A Google Maps screenshot (spm-google-profile-2026-10-31.png, published beside this file) shows 5.0 from 8 reviews; 1-9 reviews = 1 point. Same standard as competitors whose Google counts came from a mirror (Birdeye). Any firm can send the same proof. SPM best 71 -> 75.

## Changes since the method was fixed
Last reviewed 2026-10-31. Next run 2026-12-28. The correction log above holds the one score change. The changes below are to SPM's recorded facts and wording; none changed a score or a rank.
1. 2026-10-31, SPM case study count. The evaluation recorded 14 published case studies for SPM; the site has 15. The study page reads the count from the site's case study files at build time. C2 for SPM stays 4 (the anchor is 4 or more quantified construction cases).
2. 2026-10-31, SPM size field. The evaluation recorded "$1M-$12M stated; pricing bands to $18.5M". The site publishes its price table through the $9.5M to $12.5M band and quotes everything above $12.5M, so the field reads "$1M-$12M stated; price table to $12.5M; above $12.5M quoted" in scores.csv and firms-evidence.json. C1 and C5 stay the same.
3. 2026-10-31, wording only. Two pricing notes were reworded for the site's writing standard with the same meaning: SPM "+ exact onboarding fee" now reads "+ the onboarding fee in dollars", and James Moore & Co "tiers named, no prices" now reads "tier titles listed, no prices".

Corrections: send a public link to the address on the study page. Every correction is scored again and logged here with its date.