ShelfLife

1. One-liner

Watches every supplier quote inside a submitted bid and warns the estimator before an expired price becomes a signed loss.

2. Trend signal — why now?

Three things changed at once, and they all point at the same 90-day window between “bid submitted” and “contract signed.”

Material price volatility stopped being a 2021 story. Eaton executed price increases effective January 15 and April 1, 2026 in Electrical Americas, attributing them to “significant cost pressures driven by rapidly rising raw material and operational expenses.” The tariff backdrop is 50% on steel and aluminum and 50% on copper. Hubbell’s comparable 2026 increases ran from low single digits up to 10 percent. Transformer prices are up 111% since 2020 by one account and 77% since 2019 by another — either way, the direction is unambiguous.

Quote validity windows collapsed. Standard B2B quote validity is 30 days, but volatile markets have shrunk the window to 7–14 days. During periods of high volatility, many construction material suppliers only guarantee quoted prices for 15 to 30 days. Meanwhile the estimating guidance is blunt: “If your material pricing is from a quote that’s 60 days old, you could be off by 5–15% on wire alone.”

Lead times exploded, which stretches the gap the other way. Standard switchboards run 52 weeks from major OEMs. Power circuit breaker switchboards are 84+ weeks. Medium-voltage switchgear is 52–78 weeks standard, approaching 2–3 years at peak-demand manufacturers. Panelboards for light commercial sit at 28–48 weeks. When award decisions slip a quarter and equipment slots slip further, the contractor is holding a price they quoted from a document that expired months ago.

The killer detail is how price changes get communicated. They arrive through “authorized-distributor bulletins” from vendors like Womack Electric Supply, Crown Supply, and Geary Pacific — not press releases. Those bulletins “confirm the action and the effective dates… but state no percentage.” So the estimator’s exposure is buried in a pile of PDFs and emails from a dozen distributors, each with different formats, none of which reconcile against the quotes sitting in their own submitted bids. Nobody is doing that reconciliation. That’s the product.

Money is moving here too. Preconstruction ranked first as the project phase contech startups were building for in 2026 — a reversal from prior years when solutions targeted construction and operations. Brickanta raised $8M seed led by Northzone for AI preconstruction workflows including bid analysis and cost estimation. XBuild raised $19M Series A for AI estimating. Moab launched with $16M led by Elad Gil. Construction tech hit a $1B quarter in Q2 2026, with ~70% of H1 2026 deals between Pre-Seed and Series A.

Provenance:
  - Signal 1 (demand): Estimating guidance states a 60-day-old material quote puts a bid off by "5–15% on wire alone"; standard bid validity is 30 days while volatile-market supplier quotes shrink to 15–30 days — https://drawer.ai/blog/how-to-bid-on-electrical-jobs-step-by-step-guide-templates-strategies — 2026-08-12
  - Signal 2 (economic): Eaton executed 2026 price increases effective Jan 15 and Apr 1 citing raw material cost pressure, against 50% steel/aluminum and 50% copper tariffs; changes communicated via authorized-distributor bulletins that confirm effective dates but state no percentage — https://getuniqcli.com/news/eaton-price-increase-2026 — 2026-08-12
  - Signal 3 (feasibility/economic): Preconstruction became the #1 phase contech startups build for in 2026; Brickanta ($8M seed, Northzone), XBuild ($19M Series A), Moab ($16M) all funded for AI bid analysis and estimating — https://www.indexbox.io/blog/six-ai-construction-tech-startups-secure-major-funding-in-early-2026/ — 2026-08-12
  - Supporting: Switchgear/transformer lead times of 52–84+ weeks stretch the bid-to-award-to-order window far past any quote's validity — https://www.electronate.app/blog/switchgear-lead-times-2026-data-center-boom — 2026-08-12
  Category: Workflow automation

3. The opportunity

Every estimating package on the market — Accubid, ConEst, McCormick, and the new AI takeoff tools — is built around the moment before the bid goes out. They help you count devices, apply labor units, and produce a number. The instant you hit send, they stop caring.

But the loss doesn’t happen at bid time. It happens in the 30–120 days between submission and award, while the estimator is busy bidding the next six jobs. During that window:

• Supplier quotes embedded in the bid silently expire.
• Manufacturer price letters take effect on dates nobody diarised.
• The GC comes back to award, and the contractor either eats the delta or starts an awkward re-pricing conversation from a position of weakness.

The incumbent tooling in the adjacent space — AuraVMS, QuoteToMe, CalcuQuote, Luminovo — all sits on the procurement inbound side. They help you collect and compare quotes when you’re buying. AuraVMS does track validity periods and send expiry alerts, but for a procurement team managing an RFQ round, not for a contractor tracking the exposure inside bids they’ve already submitted. The direction of the arrow is different, and so is the buyer.

The gap: nobody owns the outbound bid’s material exposure after submission. Search for tooling here and you get bid tracking spreadsheets — which track whether you won, not whether your price is still real. As one comparison of available templates concluded, automated tracking of expiring quote dates and supplier repricing checks “would likely require custom spreadsheet development or specialized construction software beyond the generic templates found.”

A focused team can do this 10× better than a spreadsheet because the work is fundamentally document reconciliation across messy, unstructured inputs — supplier quote PDFs in a hundred layouts, distributor price bulletins in email, and the contractor’s own bid line items. That is exactly the shape of work that got cheap in the last 18 months and was miserable before.

4. Target market

Primary customer: The owner-estimator or lead estimator at a US electrical subcontractor with 5–50 employees doing commercial and light-industrial work — $2M–$30M annual revenue, submitting 10–40 fixed-price bids a month, with one to three people doing all the estimating. The same shape applies to mechanical/HVAC and low-voltage subs, which is the natural expansion.

Why they buy: Because they’ve been burned and they know the mechanism. The forum advice circulating among contractors is to quote “prices that are only good for 30 days or less” and to “never come down in price once quoted” — that’s a community that has already diagnosed the disease and is treating it with willpower and sticky notes. The recommended discipline is to “lock pricing with your supplier before bid day and note the expiration date.” Noting the expiration date is a manual act performed by a busy person under deadline. It fails constantly.

The math is why they pay. A $400K electrical bid with $180K of material exposure, off by 8% on an expired quote, is a $14,400 hole in a job whose target margin was maybe 5–20%. One catch pays for the software for years.

Rough TAM reasoning: Census Bureau data puts electrical contractors (NAICS 23821) at roughly 83,300 establishments employing 1.02M workers, with broader industry estimates as high as 180,000 businesses. 89% of establishments have fewer than 20 employees. The serviceable slice — commercial-focused subs with real bid volume and enough material exposure to care, call it the 5–50 employee band doing fixed-price commercial work — is plausibly 15,000–25,000 firms in the US. At $250/mo that’s a $45M–$75M ceiling. Small enough that no venture-scale player will fight hard for it; large enough to build a $3–5M ARR business.

Why now for them: Two consecutive years of unpredictable manufacturer price actions plus tariff-driven input costs have converted this from a background annoyance into the thing that decides whether a job is profitable. And lead times of 52–84 weeks mean the exposure window is no longer 30 days — it’s a year.

5. Product sketch (MVP)

• Quote intake — forward supplier quote PDFs and emails to a project inbox, or drop them in. The system extracts vendor, line items, unit prices, quote date, and the expiration/validity language wherever it’s buried in the fine print.
• Bid binder — group quotes under a submitted bid with the GC name, submission date, bid validity period, and expected award date. This is the object the whole product revolves around.
• Expiry radar — a single screen showing every open bid ranked by dollars of material exposure that is expired or expiring, not by date order. The estimator sees “$212K exposed across 4 bids” first thing Monday.
• Price-letter watch — monitors distributor and manufacturer bulletins for effective dates on the manufacturers present in your bids, and flags which of your open bids contain affected product categories.
• Exposure delta estimate — for expired lines, an estimated current-price range with the reasoning and sources shown, so the estimator knows whether to chase a re-quote or let it ride. Never a silent number — always a number with its evidence attached.
• Re-quote requests — one click generates a vendor-ready email asking for a refreshed price on specific line items, with the original quote referenced.
• Award-day brief — when a GC comes back to award, a one-page summary of what’s changed since submission, priced, with sources. This is the document the contractor uses in the re-pricing conversation.
• Escalation clause helper — flags bids where exposure has crossed the threshold at which an escalation clause should have been invoked, referencing the common 5–10% adjustment trigger structure used in US contracts.

6. AI angle — what’s load-bearing

Remove the AI and this product does not exist — it becomes a manual data-entry tool that no estimator will ever keep current, which is precisely the failure mode of the spreadsheets they already abandoned.

Three places the model does real work:

1. Quote extraction from chaos. Supplier quotes arrive as PDFs, scanned faxes, email bodies, and spreadsheet attachments, in a different layout per distributor. The validity term is prose buried in a footer — “prices firm for 15 days,” “subject to change without notice,” “pricing valid through 09/30.” Extracting line items and correctly interpreting the expiry language is the whole job, and it’s exactly what document models got good at recently.

2. Bulletin interpretation. Distributor price bulletins confirm effective dates but state no percentage. Reading them, resolving which product families they touch, and matching those families against line items in a contractor’s open bids is fuzzy semantic matching across two unstructured corpora. No rules engine survives contact with this.

3. Category matching. Mapping “3/0 THHN CU” on a supplier quote to a copper-tariff exposure category, or a manufacturer part number to the product family named in a price letter, requires domain-flavoured judgment at a scale that makes manual mapping pointless.

What the AI must not do is invent a current price. Every delta shown carries its source. In a business where the output feeds a contract negotiation, a confidently wrong number destroys trust permanently — the product’s credibility rests on always showing its work, and on saying “I don’t know, go get a re-quote” when that’s the honest answer.

7. Localization angle (if any)

N/A — this is a US-first play, with the UK/Australia/Canada as natural follow-ons.

The wedge is specifically the US market’s combination of tariff-driven input volatility (50% copper, steel, aluminum), a distributor-bulletin communication culture that keeps price intelligence unstructured and fragmented, and a large population of small fixed-price subcontractors. India and SEA have plenty of electrical contractors but the bid structures are more often rate-contract or item-rate rather than fixed-price lump sum with embedded supplier quotes, which removes the exact exposure this product manages. Forcing a localization angle here would weaken the product.

8. Business model — path to $1M–$5M ARR

Pricing: $199/mo Starter (up to 15 open bids, 1 estimator) · $349/mo Pro (unlimited bids, 3 estimators, price-letter watch) · $699/mo Multi-office (5+ seats, mechanical + electrical). Annual billing at 2 months free.

This is benchmarked, not invented. Contractors in this segment already pay for estimating tools: ConEst IntelliBid starts around $115/month per user, McCormick is listed around $300 per user per month, cloud estimating platforms publish rates from $299.90/month, and Accubid single-user licenses run ~$3,000 upfront with enterprise tiers at $15,000+/year per user. A $349/mo tool that protects margin sits comfortably inside an existing budget line and is trivially justified against a single $14K catch.

ACV: ~$3,400 blended (assumes majority on Pro, some Starter, a tail of Multi-office).

Rough math to $1M ARR: 295 customers × $283/mo average × 12 = $1.0M. That’s roughly 1.5% of the serviceable 20,000-firm segment. Achievable.

Rough math to $5M ARR: 1,200 customers at a slightly higher blended ACV ($350/mo) as multi-trade and multi-office accounts grow. This requires expanding beyond electrical into mechanical/HVAC and low-voltage — same product shape, same exposure problem, different product categories. It also likely requires a distributor or association channel rather than pure direct sales. Realistic in 3–4 years, not 18 months.

Expansion path: Seats grow with estimator headcount. Trades expand (electrical → mechanical → low-voltage → GC-level). The high-value upsell is the award-day brief becoming a paid change-order documentation module, since the same evidence trail that wins a re-pricing conversation also supports a change order — and change orders are where contractors already accept that documentation costs money.

9. Go-to-market wedge — first 100 customers

The buyer is findable, licensed, and publicly listed. That’s the whole advantage.

• Loss-autopsy cold outreach on public bid tabulations (customers 1–40). Public-sector bid tabs for school districts, municipalities, and state facilities are published with every bidder’s name and number. Pull tabs from 3 states, extract the electrical subs who bid, and reach out with a specific hook: “You bid the Jefferson County job in March. It awarded in July. Eaton’s April 1 increase landed in between. Did you re-price?” That’s not a cold email, it’s a diagnosis. Target 800 contacts, expect 8–12% reply on specificity alone, close 5%.

• State licensing boards + IBEW/NECA chapter directories (customers 40–75). Electrical contractor licenses are public records in every state. Filter to firms with commercial classifications, cross-reference against company size signals, and run a personalized sequence built around the specific manufacturers they carry. NECA chapter meetings are where these owners physically gather — a 20-minute talk titled “What the April price letter cost you” is a warm room, not a pitch.

• Distributor co-marketing (customers 75–100+). This is counterintuitive but strong: a distributor’s own interest is served when a contractor comes back for a legitimate re-quote instead of eating the delta and blaming the distributor. Approach 5 regional distributors with an offer to co-brand the re-quote workflow. One regional distributor relationship puts the product in front of hundreds of contractors with implicit endorsement. This is also the beginning of the moat.

• The estimator forums. Mike Holt’s forums and Contractor Talk have active, searchable threads on escalation clauses and volatile-market bidding going back years. These are not places to advertise — they’re places to be genuinely useful about a problem you’ve studied harder than anyone. A single well-regarded contributor account is worth more than a paid campaign in this niche.

The first 100 is a 4–5 month sprint with real sales effort, not a 2-week growth hack. That’s honest, and it’s the main reason distribution scores 12 rather than 14.

10. Build complexity — justification

Medium. The web app, bid binder, alerting, and email generation are all standard off-the-shelf work. The genuinely hard parts are two: reliable extraction of line items and validity terms from wildly inconsistent supplier quote documents, and building the price-bulletin monitoring corpus across distributors who publish to authenticated portals and email lists rather than open web pages. The second is as much a business-development problem as an engineering one — you may need distributor relationships to get reliable bulletin access, which is why the distributor channel in section 9 matters strategically and not just for leads.

A pair could ship a credible v1 in 12–16 weeks by starting narrow: electrical only, the top 6 distributor quote formats, and manual curation of price bulletins for the first 50 customers. Manually curating bulletins is the right call at small scale — it’s a few hours a week, it teaches you the domain, and it produces the training data that makes automation work later.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Processing the customer’s own quote documents on their behalf. Price bulletins are distributed to trade; respect any redistribution terms and never republish bulletin content across customers.
Ethical — no harm / dark patterns	✅	Helps a small contractor avoid unknowingly signing a loss. No deception, no dark patterns; the honest-uncertainty design rule is a feature.
Market exists (evidence above)	✅	83K+ establishments, existing spend of $115–$300/user/mo on adjacent estimating tools, documented 5–15% error from stale quotes.
1–5 person team can build this	✅	Pair in 12–16 weeks for a narrow v1.
Launchable with <$50K / ₹40L	✅	Two-person time plus modest inference and infra. Well under $50K.

All five pass.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Real money, and the mechanism is well understood by the sufferers — the community has already invented manual defences (30-day quote limits, escalation clauses). Not 18+ because the pain is episodic rather than daily; a contractor feels it a few times a year, sharply. Episodic pain is harder to sell against than daily pain.
Demand evidence	15	12/15	Strong indirect evidence: documented 5–15% error, confirmed 2026 manufacturer price actions, existing paid spend on estimating tools, funded contech in the adjacent preconstruction space. Not higher because I found no product being bought specifically for post-submission bid exposure — the demand is inferred from the pain, not observed as spend.
Build feasibility	15	11/15	Standard stack plus two genuinely messy problems: heterogeneous quote extraction and bulletin access. 12–16 weeks for a pair, with a manual-curation crutch early.
Distribution clarity	15	12/15	Public bid tabs and licensing boards make the list buildable and the outreach specific. Docked for a real 4–5 month sales cycle to 100 customers — contractors are not fast self-serve buyers.
Revenue mechanics	15	12/15	Pricing benchmarked against tools these buyers already pay for; $1M ARR needs ~295 customers, which the channel can plausibly deliver. $5M requires trade expansion, which is a real but unproven assumption.
Time to first revenue	10	8/10	Pre-sellable — the pitch works before the product is finished because the problem is diagnosable from public bid data. Realistically 6–8 weeks from launch to first paid, likely earlier via design partners.
Defensibility	10	5/10	Honest score. The extraction is copyable within a year. The real accumulating moats are the curated price-bulletin corpus with its history of effective dates, and distributor relationships — but neither exists at month 3. Execution and focus are the moat early.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · sales-heavy

Technical because the document extraction quality is the product — a 90%-accurate extractor is worse than useless when the output feeds a contract negotiation. Sales-heavy because contractors buy from people who demonstrably understand their trade, over the phone, not from a pricing page. A founder who can’t hold a credible conversation about labor units and switchgear lead times will bounce off this market. Bringing on an estimator as an advisor or co-founder is close to mandatory.

Key assumptions to validate (3–5)

1. Assumption: Estimators at 5–50 employee electrical subs genuinely lose money to stale quotes at a frequency they can recall and quantify — at least twice in the past 18 months. How to test: 25 phone interviews with estimators sourced from state licensing boards. Ask for specific jobs, specific deltas, specific dollars. If they can’t name an instance, the pain is theoretical and the score drops hard.
2. Assumption: Supplier quote documents can be parsed to ≥95% accuracy on line items, prices, and validity terms across the top distributor formats. How to test: Collect 100 real quote PDFs from 5 friendly contractors, run extraction, hand-verify. Anything below 95% on validity-term interpretation means a redesign toward human-in-the-loop confirmation.
3. Assumption: Distributor price bulletins are obtainable at a cadence and coverage that makes the price-letter watch feature real, rather than a demo that decays. How to test: Spend two weeks attempting to source bulletins from 10 regional distributors through direct relationships and customer forwarding. If coverage is unreachable, ship without this feature and lean entirely on quote-expiry tracking, which stands on its own.
4. Assumption: $349/mo clears the buying threshold without a procurement process. How to test: Price the pitch at $349 in the first 25 sales conversations and observe whether the objection is price or priority. If it’s priority, the problem is episodic-pain positioning, not the number.

Risk flags

1. Adjacent incumbent expansion: The funded preconstruction AI startups (Brickanta, XBuild, Moab) are building in the neighbouring room. Any of them could add post-submission quote tracking as a feature. Mitigation is depth — the price-bulletin corpus and distributor relationships are not a weekend feature — but this is a genuine race, and being acquired-or-crushed is a real outcome.
2. Data access dependency: The price-letter watch depends on bulletins distributed through authenticated channels. If distributors object to systematic collection, that feature degrades to customer-forwarded content only. The product survives this but loses a differentiator, which is why assumption #3 must be tested early rather than assumed.
3. Episodic pain / priority risk: This is the risk I’d worry about most. A contractor who hasn’t been burned in the last six months will agree the problem is real and still not buy this quarter. The counter is selling immediately after a documented price event — which makes the go-to-market timing-dependent in a way that’s uncomfortable but workable, since 2026 has produced price events reliably.
4. Market timing on volatility: If tariffs unwind and input prices stabilise for 18 months, urgency evaporates. The product becomes a nice-to-have. This is genuinely outside the founder’s control.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder comfortable with messy document extraction,
                        paired with someone who can sell to contractors by phone.
                        An estimator advisor is close to mandatory.
Time to revenue:        6–8 weeks from launch; design partners can pay earlier
Capital to launch:      $15–25K (two people's time plus inference and infra)
Top 3 assumptions to validate first:
  1. Stale-quote losses are recallable and quantifiable — 25 estimator interviews
     sourced from state licensing boards; demand specific jobs and specific dollars
  2. Quote extraction hits ≥95% on validity terms — 100 real PDFs, hand-verified
  3. Price bulletins are obtainable at useful coverage — 2-week sourcing sprint
     across 10 regional distributors
Kill criteria:
  - Abandon if fewer than 10 of 25 interviewed estimators can name a specific job
    where a stale quote cost them money in the last 18 months
  - Abandon if extraction accuracy on validity terms stays below 85% after two
    iterations on real documents
  - Abandon if fewer than 5 of the first 60 qualified outreach conversations convert
    to paid pilot within 90 days

15. Next step — 1-week validation sprint

Day 1–2 — Build the list and the hook. Pull public bid tabulations from 3 states for jobs bid in Q1 2026 and awarded in Q2 2026 — the exact window that straddles Eaton’s April 1 effective date. Extract the losing and winning electrical subs. This list is the asset; it also proves whether the outreach hook is mechanically constructible at scale.

Day 3–4 — 25 interviews, one question that matters. Call estimators from that list. Do not describe the product. Ask: “You bid this job in March, it awarded in July — what did you do about material pricing in between?” Then shut up and listen. Count how many describe a specific loss with a number attached versus how many say “we build in a contingency.” Simultaneously, collect real supplier quote PDFs from any contractor willing to share — target 100 documents — and run extraction against them.

Day 5 — Decide on two numbers. Go if ≥10 of 25 estimators name a specific job with a quantified loss in the last 18 months, and extraction on the collected PDFs correctly identifies the validity term on ≥85% of documents. Fewer than 10 recallable losses means the pain is real but not urgent enough to sell against, and the honest move is to shelve it. Below 85% extraction means the v1 must be redesigned around human confirmation rather than automation, which changes the pricing and the pitch — worth knowing before writing a line of production code.

Both thresholds are falsifiable, and both can be wrong in ways that stop the project. That’s the point.