Valuation, labor economics, and revenue framework (external research)

Product context: Estym8 is built from scratch as an AI-first construction preconstruction platform—not a legacy takeoff stack with AI bolted on. AI runs across the product: bid-package ingestion and classification, multi-model takeoff and vision, plan intelligence, cross-file synthesis, Estee, estimate-to-submittal draft review, and optimization recommendations. Canonical framing: AI-first positioning.

Purpose: Ground market size, labor cost, time-to-estimate, and rough company valuation in public statistics and third-party sources—not internal product docs. Use with legal and financial advisors; figures and multiples change with market conditions.

Last updated: May 2026 — includes §5A (~$5B/yr 20% of loaded labor framing) and §5B (hours-based savings with explicit f = plan-takeoff share of annual hours). Pitch deck: estym8.ai/docs/pitch-deck. Still not investment advice—re-validate before a live fundraise with counsel and current market comps.

---

1. Disclaimers

• This is not investment, tax, or legal advice.
• Valuation for private companies depends on term sheets, cap table, growth, churn, margins, IP, and buyer strategy—tables below are illustrative scenarios only.
• Web and BLS data lag reality; always verify before relying on numbers in fundraising or diligence.
• Pre-revenue and ARR multiple examples are conversation brackets for sensitivity analysis; they do not represent an internal fairness opinion, a third-party valuation, or an offer.

---

2. How many cost estimators (US)?

Occupation: Cost Estimators, SOC 13-1051 (BLS).

Re-verified June 2026 against the live BLS Occupational Outlook Handbook (May 2024 cycle): employment 221,400, median $77,070 / $37.05/hr, 10th/90th $46,330 / $128,640 — all figures above confirmed accurate.

Demand-side read of the -4% projection: BLS projects a modest decline in the estimator headcount over the decade. We treat this as a tailwind for automation, not a shrinking market — a labor-supply gap (fewer estimators against steady/rising bid volume) is precisely the work software absorbs; the value pool is loaded labor cost + the cost of unfilled estimating capacity, not headcount growth.

Important nuance: This SOC includes estimators in manufacturing and other industries, not only building construction. Construction-related slices are still large—BLS also publishes median wages by industry for the same occupation, for example:

(Source: BLS Cost Estimators page, industry table.)

Serviceable market for plan-based trade estimating: A subset of this 221k. Many hours also sit in preconstruction, project engineers, and PMs who estimate part-time; official counts understate “people doing takeoffs” if you only count the job title “cost estimator.”

Go-to-market focus (Estym8): Early revenue and product depth emphasize commercial, institutional, and larger multi-family pursuits with full drawing sets and bid-folder behavior aligned to multi-trade estimating and documentation—not a single-discipline niche. MEP remains the deepest automation and calibration surface today; other trades and GC scopes share the same folder pipeline where drawings support them, with vertical expansion on the roadmap. Purely small residential-only remodel bids are a weaker initial fit until the roadmap deepens that slice. This is positioning, not a revision to the BLS occupation counts above.

---

3. Pay, benefits, and “fully loaded” employer cost

3.1 Government benchmark (cash wages)

• Median annual wage $77,070 and median hourly $37.05 (BLS, above).

3.2 Broader construction-estimator compensation (trade / HR commentary)

Industry and recruiting studies often report higher average base for “construction estimator” titles (experience- and region-driven), and highlight bonuses and benefits as material parts of total compensation. Treat these as directional, not BLS-official:

• Example industry commentary: “Chief estimators” / senior roles with bonus-heavy packages; regional variation (e.g. high cost-of-living markets). See e.g. trade salary commentary and CFMA compensation trends and aggregator pages (verify date and methodology).

3.3 Employer “load” on salaries (rule of thumb)

BLS publishes the Employment Cost Index for employer costs of employee compensation, useful for thinking about benefits + employer shares alongside wages:

• BLS Employment Cost Index (current releases).

Practical modeling for fully loaded cost:

Example (illustrative, not a recommendation):
$77,070 median × 1.30 burden ≈ $100,000 loaded per year.
÷ 1,850 hours ≈ $54/hour employer cost for a median-cost estimator if all hours were billable estimating (they are not—meetings, admin, and non-bid work reduce “available” hours).

---

4. How long do humans take to take off and estimate?

There is no single BLS statistic for “hours per MEP bid.” Use industry ranges and assume larger jobs run to multi-day or multi-week calendar time, especially when takeoffs span many sheets, trades, reviews, and addenda.

4.1 Takeoff share of bidding

• Takeoff Monkey (construction takeoff service blog, 2024): States that a material takeoff can take from minutes to several workdays, and that for large or complex commercial projects, takeoffs may take several days or even weeks. Also states takeoff-related work can be on the order of 50–80% of the bidding process (their narrative benchmark—not a government statistic).

4.2 Electrical / commercial estimating (order of magnitude)

Trade and training content commonly bands electrical / commercial work by project size—for example, many hours on small jobs, multiple days on medium/large, and a week or more when complexity is high (multi-story, heavy gear, incomplete documents). See discussions in contractor media and training sites (e.g. commercial electrical estimating guides; Electrical Contractor Magazine and similar trade sources often discuss productivity and bid timing qualitatively).

4.3 “Days to weeks” on the largest pursuits

For very large or phased pursuits (healthcare campuses, major industrial, multi-package GC bids), calendar time often stretches because of:

• multiple addenda,
• discipline coordination,
• peer review,
• and re-bid cycles.

So the user intuition “days to weeks on large projects” is consistent with published takeoff-service and trade commentary, even when a single estimator’s “touch time” is hard to isolate.

4.4 Consolidated bands for modeling (planning only)

Use these only for ROI storytelling and sensitivity analysis—prove your own numbers with pilot timers.

---

5. Economic value created (buyer ROI sketch)

Illustrative labor value per bid, if the product saves a known number of estimator hours multiplied by a reasonable fully loaded cost per hour:

Value (labor only) ≈ (hours saved) × ($/hour loaded)

Example: 12 hours × $60/hour = $720 per bid in labor-equivalent savings—before counting missed bids, rework from errors, or extra bids won from faster turnaround.

---

5A. Industry-wide efficiency value pool — the “~$5B at 20%” framing (illustrative)

This section reconstructs the arithmetic behind a headline like: “If the industry saved even ~20% of addressable estimator labor cost, the annual value pool is on the order of ~$5B.” It is not an Estym8 revenue forecast, not an enterprise valuation, and not BLS-official beyond the cited headcount and wage anchors.

5A.1 Anchors (from this doc + BLS)

5A.2 Total loaded labor pool (before any “20%”)

Using N = 221,400 and loaded annual cost C:

5A.3 Applying 20% — where ~$5B / year comes from

If—only as a thought experiment—20% of that loaded labor could be redeployed or eliminated through better tools, process, and automation without losing output quality:

Rounding: ~$4.2B–$4.9B / year brackets ~$5B in normal investor rounding. That is an annual societal / buyer value-unlock story, not money that accrues to one vendor.

5A.4 Connection to takeoff duration and errors (qualitative)

• Time (§4): Commercial takeoffs commonly run hours → weeks of calendar and focused time; if a meaningful share of loaded estimator hours sits in counting, reconciliation, and rework, then a 20% “efficiency wedge” is a storyline for TAM slides, not a measured industry statistic.
• Human error: Rework, missed scope, and late addenda add cost beyond pure hours; we do not add a separate dollar layer here because credible industry aggregates vary by segment—treat error/rework as upside optionality on top of the labor pool, not a second precision claim.

5A.5 Critical caveats (read before repeating externally)

1. SOC 13-1051 is not “MEP plan takeoff only.” It includes estimators outside building construction (e.g. manufacturing). Using the full 221k overstates a construction-only TAM unless you apply a haircut (illustrative only):

   Construction-oriented haircut	Headcount used	@ $100k loaded	20%
   100% of SOC (upper bound)	221,400	~$22.1B	~$4.4B/yr
   50% (illustrative)	110,700	~$11.1B	~$2.2B/yr
   40% (illustrative)	88,560	~$8.9B	~$1.8B/yr

2. Not every hour is takeoff. Estimators spend time on meetings, vendor quotes, site visits, and non-PDF work—so 20% of all loaded labor is more aggressive than 20% of takeoff hours only unless you define scope narrowly with data.

3. Value pool ≠ vendor revenue. If buyers paid software fees equal to 2% of a $4.4B annual pool, that would be ~$88M ARR at full theoretical capture—still enormous for a startup, but orders of magnitude below $5B. A 1% fee on the same pool is ~$44M ARR. These are illustrative fee-on-value stories, not forecasts.

4. Company enterprise value remains in §7 (pre-revenue bands and ARR multiple math)—do not confuse §5A labor-value pools with Estym8 valuation.

5A.6 Cross-reference — hours-only model

For a model that applies savings to a takeoff-like hour bucket (instead of “% of all annual loaded labor”), see §5B immediately below. §5A and §5B are alternative framings; do not double-count them in the same slide.

---

5B. Industry time saved — hours-based model (addresses “not every hour is takeoff”)

This section estimates annual hours that could be compressed if Estym8-class workflows (single PDF or bid-package folder upload → classified run plan → takeoff + verbatim harvest + plan/cross-file intelligence + narrative export + estimate) cover a growing share of commercial estimator work. It is still illustrative—prove with pilot stopwatch studies on real bids.

5B.1 Definitions

Annual hours saved (if magically full adoption on modeled scope):

Hours saved ≈ N × H × f × η (multiply the four factors in the table above).

Dollar value (labor only) ≈ Hours saved × loaded $/hour. Use $55–$60/hr as quick loaded hourly anchors from §3.3 (median cash ÷ ~1,850 h × ~1.3 burden lands in that band—calibrate for your buyer).

5B.2 Scenario table (illustrative — do not cite as measured industry fact)

How to read this vs §5A: §5A’s “20% of all loaded labor” (~$4.4B/yr at $100k loaded) lines up with the Hero row’s dollar outcome only if you assume both (i) full SOC headcount and (ii) a very large share of the year is plan-takeoff-like and (iii) very high automation capture in that bucket. The Base row is a more defensible story until measured: ~$1.4–$1.5B/yr labor value at half the SOC headcount and mid f, η.

5B.3 How Estym8 maps onto f and η (product truth, not new claims)

What the product shortens today is mostly the manual loop from upload → structured quantities + harvest + intelligence artifacts + estimate body, compared to days of manual counting and retyping context. It does not remove: relationship management, scope meetings, non-uploadable owner directives, trade knowledge calls, or final bid assembly outside the app. That is why f stays well below 1.0 and η stays below 1.0 even when the demo feels like “BAM.”

5B.4 Relation to “50–80% of bidding is takeoff” (industry blog, §4.1)

Some vendor commentary bands takeoff as half or more of the bidding process—that is not the same as half of every paid annual hour for SOC 13-1051. Using 50–80% as f would explode the model; we do not do that here. If you want an aggressive deck appendix, you can show a sensitivity column with f = 0.40 only with a footnote “upper theoretical bound for plan-desk-heavy estimators—requires validation.”

5B.5 What to measure next (to replace illustration with data)

1. Stopwatch pilots: wall-clock before/after on matched pursuit types (folder vs manual).
2. Decompose tasks: coding sheets for takeoff vs admin vs pricing so f is evidence-based per segment.
3. Adoption curve: none of the tables include penetration; multiply any row by 0.1%, 1%, 5% adoption to sanity-check vendor revenue, not just social value.

---

6. SaaS valuation multiples (context for an exit / sale)

Public and private B2B SaaS is often discussed on ARR or revenue multiples, highly sensitive to growth, NRR, Rule of 40, and profitability. Recent industry summaries (verify dates):

Construction-specific vertical SaaS transactions vary; strategics may pay for strategic fit + team + tech (sometimes revenue + premium), while financial buyers anchor more on margins and retention. No single public “construction estimating multiple” should be trusted without comparables.

---

7. Estimated valuation for Estym8 (scenarios—illustrative only)

Before you cite a number externally: Replace or stress-test the pre-revenue band and the ARR × multiple tables with your own cap-table math, dilution, and comparables; have legal and financial advisors sign off on anything used in a term sheet, SAFE, or slide deck.

These do not reflect an actual offer or fairness opinion.

7.1 Pre-revenue / minimal revenue (“technology + team”)

Early sales to a strategic buyer might combine:

• Acqui-hire value for team,
• IP / product premium for differentiated AI + workflow,
• Option value of the vertical.

Wide band (illustrative): low single-digit $M to ~$10–25M+ depending on proof (customers, LOIs), IP, and competitiveness of process. Treat as conversation bracket, not a formula.

7.2 Revenue-based quick ranges (assume modest vertical SaaS discount vs hottest horizontal SaaS)

Let R = ARR ($ millions). Illustrative enterprise value ≈ R × M, where M is a multiple chosen for growth and quality.

Why multiples widen: net revenue retention, yoy growth, gross margin, CAC payback, and strategic scarcity (e.g. unique plan-intelligence position) can move you above this band; high churn, services-heavy COGS, or key-person risk can push below.

7.3 Strategic competitor sale

An acquirer in construction software / preconstruction may pay:

• Revenue multiple plus
• synergy NPV (distribution, data, bundling), and/or
• defensive value vs competitors.

That can exceed pure financial-buyer math—but only with negotiation and alternatives.

7.4 Second product (Bodi) — option value (planning note)

The company also builds Bodi, an ICT / low-voltage “project brain” product (sources → claims → Basis of Design). For modeling:

• Do not assume Bodi ARR in the same cells as Estym8 until you have separate pipeline assumptions; treat it as incremental TAM (different buyer thread on the same jobs) and integration upside (Estym8 plan intelligence feeding Bodi claims—see Bodi + Estym8).
• Strategic buyers may ascribe portfolio value (two preconstruction narratives, shared AI discipline) beyond a simple sum of parts—or may only value Estym8 until Bodi has independent proof. This is conversation-dependent, not formulaic.
• Any combined valuation range requires its own spreadsheet: blended churn, separate CAC, and clarity on whether GTM is unified or parallel.

---

8. Plan: testing → revenue → attention → optionality

1. Paid pilots with documented time savings (stopwatch vs status quo).
2. Reference customers and short case studies (even one page).
3. Pricing anchored below fully loaded internal cost for marginal bid and below many outsourced takeoff quote bands—tune with win rate.
4. Content and SEO around takeoff time, RFI risk, and subcontractor bid workflow (keywords match pain).
5. Partners (associations, manufacturers, integrators) for reach without overbuilding sales.
6. Data room early: cap table, contracts, IP assignments, security posture—reduces friction for strategic talks.

---

9. Beta partners and comped access (commercial)

Some closed-beta companies may participate under an explicit agreement: structured product feedback (and reference potential) in exchange for a comped lifetime commercial license (unlimited plan-set / estimate usage in product terms).

In the product: At admin approve time, toggle Lifetime beta (comp) so that when the user completes invite signup, User.betaLifetimeAccessAt is set. Trial / paywall logic treats that like unlimited access regardless of testing mode—separate from time-bound approvedForTestingAt for general testers. Document the commercial terms (scope, IP, support level) in your own contracts; this doc only names the product hook.

---

10. Subscriber-scale MRR (illustrative)

Purpose: Quick monthly recurring revenue (MRR) and annual run-rate (ARR = MRR × 12) from paying subscriber counts. For planning and investor slides only—not a forecast.

Pricing anchors (aligned with lib/pricing/public-plans.ts, May 2026):

“Subscriber” in tables below means paying seat unless noted.

10.1 MRR by seat count — blended ~$170/seat/mo

All-Solo sensitivity (1,000 seats @ $179): MRR $179,000, ARR $2,148,000.

Studio-heavy example (300 seats, avg 3 seats/account @ $149): 100 accounts × $447 ≈ $44,700 MRR (illustrative mix—not additive with table above).

10.2 Should we offer annual prepay?

Yes, it’s usually worth offering for B2B SaaS like Estym8, with clear terms.

• Cash upfront — Improves liquidity vs monthly; helpful before heavy inferencing COGS months.
• Retention signal — Buyers who commit for a year often have clearer ROI and onboarding buy-in.
• Lower admin — Fewer monthly failed-card events; can align to customer fiscal years.
• Pricing — Common band: ~10–15% off vs 12× monthly (Estym8 uses 15% — ANNUAL_DISCOUNT_PERCENT).
• Risks — Refunds / pro-rata policy must be explicit; annual can delay churn feedback if product fit is wrong—pair with strong onboarding and a mid-year check-in for large deals.

Practical default: Offer monthly and annual side by side; annual labeled as “Save X%” with the same entitlements as monthly. Enterprise or multi-seat deals can stay invoice + annual contract even if the self-serve card UI stays monthly-only at first.

---

11. Source list (bookmark)