A cheaper, smarter alternative to BidNet.
One engine, two faces: a private radar that finds you a government niche worth entering, and the same intelligence sold as a SaaS that undercuts BidNet on price and beats it on product.
The core bet
BidNet's business is selling access to public data, priced as a per-state tax. Its only real moat is the exclusive right to host bid postings for specific agencies. We do not attack that moat. We win on what is already free (federal data) and on what BidNet does badly (intelligence).
Free data
SAM.gov publishes a full daily CSV of every active opportunity. No key, no rate limit. Our bulk ingest is a download, not a scraping war.
Intelligence gap
BidNet shows listings. It does not summarize a 200-page RFP, score your fit, check set-aside eligibility, or draft your capability statement. That gap is our product.
Flat pricing
BidNet multiplies per state. We sell one national price below a single BidNet state package.
The research method, turned into software
Winning small government contracts starts as a manual research loop. Every step is something software can automate. Automating it is the product. Here is the loop mapped to the feature that replaces it.
| What most contractors do by hand | What BidSight does for you | Where |
|---|---|---|
| Government creates demand via contracts, regulation, and the tax code | Framing for niche selection and content marketing | core |
| Map your offering to NAICS + PSC (federal) / NIGP (local) codes | Code lookup from a plain-English business description | AI |
| Pull award history by code to size the market and find buyers | Niche Finder: scores every code on money, fragmentation, competition | core |
| Catch "sources sought" / RFI notices during market research | Sources Sought Radar (ptype=r), surfaced not buried | core |
| Use set-asides (8(a), WOSB, SDVOSB, HUBZone) to cut competition | Eligibility explainer from your profile + certification steps | AI |
| Subcontract under primes first for fast revenue + past performance | Prime/subcontract signals on each award record | later |
| Favor boring recurring services | Filters for recurring buyers and small, repeated awards | core |
Verified market examples (public award data)
R706's profile (huge volume, tiny average award, mostly under the $250k simplified-acquisition line) is the textbook small-business niche. The Niche Finder is built to surface exactly that shape automatically.
Who we are up against
Verified pricing as of June 2026. The pattern: incumbents either charge per state (BidNet) or charge a lot for national (DemandStar), and none of them add intelligence on top of the listings.
| Provider | Price | Coverage | AI / intelligence |
|---|---|---|---|
| BidNet Direct | ~$9-45 / state / mo; premium region ~$1,500/yr | State + local (its moat), federal add-on | None |
| DemandStar | County ~$35/yr · state $35-$1,299/yr · national ~$2,599/yr | State + local network | None |
| SAM.gov | Free | Federal only | None (and brutal UX) |
| BidSight | $0 free / $29 / $79 mo, flat national | Federal now, local rolling out honestly | Summaries, fit, set-aside, drafts |
The wedge we attack
The per-state tax. A business wanting three states pays BidNet three times. We charge one flat national price that is below a single BidNet state-and-local package (~$36/mo).
The wall we avoid
BidNet's exclusive agency-hosting contracts. We do not try to out-host them. Federal stands on its own; local is upside layered in over time.
Data architecture
Federal sources (build now, all free)
| Source | Gives us | Limits | Role |
|---|---|---|---|
| SAM.gov daily CSV | Every active opportunity, full records | None. No key, no rate limit | Primary nightly bulk ingest |
| SAM.gov Opportunities API | Full description text, set-aside filter, Sources Sought (ptype=r) | 10/day basic, 1,000/day after entity reg (~2-3 wks) | On-demand detail + Sources Sought polling |
| USAspending.gov API | Historical awards by code, agency, geography | Free, no key | Powers the Niche Finder |
The nightly ingest, step by step
Pull the SAM.gov daily CSV via a scheduled Netlify function.
Compare against stored notice IDs; isolate new and changed records.
Write new/changed opportunities into Postgres (Supabase) with NAICS, PSC, set-aside, agency, place of performance, deadline indexed.
For opportunities matching a paid user's saved search, fetch full text via API and generate a Claude summary + fit score. Cache it.
Email each user the new matches since their last digest.
State + local (the hard, defensible part)
- Open-data portals first. Missouri, Kansas, KCMO, Johnson County to match your footprint, then biggest states by spend.
- Per-portal scrapers normalized into the same schema; LLM-assisted parsing makes this far cheaper than when BidNet built its moat.
- Honest coverage map. Show exactly which jurisdictions are live. Never pretend to national local coverage.
Stack (fits your existing world)
Store
Postgres on Supabase. Real filtering across thousands of rows. (Blobs is wrong here.)
Front + cron
Netlify + GitHub auto-deploy; scheduled functions for the nightly ingest.
Intelligence
Claude API for summaries, scoring, drafts, code lookup.
Start on your Gmail/Zapier wiring; move to Resend/Postmark once paying users exist.
The Niche Finder
This is the feature BidNet has no answer to, and the first product we prove with real data. It answers "where should my business even play?" by scoring every candidate code on the shape of a winnable small-business niche.
The eight signals scored per code
Money
Total dollars spent over the last 5 years. Bigger is better, but only paired with the signals below.
Fragmentation
Number of distinct contracts. More contracts = more entry points for a small player.
Award size fit
Median award and the share under $250k (the simplified-acquisition sweet spot where small firms win).
Competition
Number of distinct winning vendors. Many winners = open; a few winners = locked up.
Set-aside share
Percent of awards reserved for set-asides. High share + you qualify = a thin field.
Recurrence
Do the same buyers re-buy on a cycle? Recurring revenue is the whole "boring business" point.
Geography
Is the buying near Kansas City, or would you be chasing work across the country?
Barrier to entry
Licensing, bonding, clearances, capital. Lower barrier ranks higher for a first contract.
How the score is built
Weights are a starting point, refined against real niches small businesses actually pursue. Every input traces to a USAspending query the user can re-run, so no number is unverifiable.
Live Niche Finder output (real USAspending data, 2020-2025)
| Niche | NAICS | 5yr spend | Contracts | Avg award | Top-5 share | Score |
|---|---|---|---|---|---|---|
| Logistics & distribution | 541614 | $9.2B | 138,690 | $67K | 71% | 71 |
| Janitorial | 561720 | $7.8B | 20,480 | $379K | 16% | 62 |
| Landscaping / grounds | 561730 | $2.5B | 15,356 | $160K | 12% | 54 |
| Temporary staffing | 561320 | $5.6B | 21,044 | $264K | 35% | 53 |
| Technical & trade training | 611519 | $6.3B | 2,594 | $2.4M | 51% | 40 |
This is not a mockup. These rows came from the BidSight engine querying USAspending live. Note the tension the data surfaced: logistics wins on raw volume but is 71% locked up by the top 5, while janitorial is a genuinely open field. That weighting is exactly what the scoring model is designed to surface.
Feature spec
Core search & alerts Available
- Search + filter by keyword, NAICS, PSC, set-aside type, agency, place of performance, deadline, dollar range.
- Saved searches that become email alerts on new matches.
- Opportunity detail with full solicitation text pulled from the API.
- Sources Sought Radar - early-stage notices most tools bury. Catch demand before the RFP is written.
AI intelligence Expanding
- Plain-English summary of any solicitation: what they want, key dates, mandatory requirements, evaluation criteria, gotchas.
- Fit score against the user's saved business profile, with a short why / why-not.
- Set-aside eligibility explainer: which set-asides they likely qualify for and how to get certified.
- Capability statement draft: auto-generate the one-pager agencies and primes request.
- Niche Finder as a premium feature (Section 04).
Out of scope for v1
Full proposal-writing suite, teaming marketplace, post-award contract management. These are later upsells, not launch features.
Pricing
One flat national price, deliberately set below a single BidNet state package. The AI features justify the premium tier without needing more data.
- Federal search + view
- 1 saved search
- No alerts, no AI
- Unlimited saved searches + alerts
- Sources Sought Radar
- AI summaries + fit scoring
- All 50 states' federal data
- Everything in Pro
- Niche Finder
- Capability statement drafts
- Set-aside eligibility + award analytics
- Per region, as it goes live
- State + local coverage
- Sold honestly by jurisdiction
Go-to-market
1. Content as the funnel
Content marketing proved the audience is hungry. Publish auto-generated "which boring niches have the most money and least competition" reports. Each is marketing and a live demo of the paid Niche Finder.
2. Free federal search ranks
SEO on "[NAICS] government contracts", "sources sought [industry]", "is [business] eligible for SBA set-aside".
3. Target the video's viewer
Aspiring boring-business owners with no contract yet who find BidNet expensive and confusing.
4. Wedge communities
SDVOSB / WOSB / HUBZone groups (the eligibility explainer is a magnet) and trade associations for waste, trucking, training, janitorial.
Product roadmap
Foundation Data & Niche Finder
- SAM.gov daily CSV downloader + parser
- USAspending.gov client for award history by code
- Niche Finder scoring over your seed codes
- Ranked niche report from public award data
Live today Federal search & alerts
- Supabase schema + nightly ingest cron
- Search UI on Netlify with all filters
- Saved searches + email alerts
- Opportunity detail with full text
- Sources Sought Radar view
In product AI layer + billing
- Claude summaries + fit scoring (cached)
- Set-aside eligibility explainer
- Capability statement generator
- Stripe billing + Free / Pro / Edge tiers
Growing State + local coverage
- Missouri / Kansas / KC open-data ingestion
- Per-portal scrapers, normalized schema
- Honest live-coverage map + local add-on billing
Risks & caveats
- You still need a business to win contracts. The tool finds and qualifies opportunities; it does not perform the work. Use the Niche Finder to decide whether a code is worth pursuing before you invest in past performance and certifications.
- State/local is genuinely hard. No free firehose, lots of scraping, and it is exactly where BidNet's exclusivity bites. Federal must stand on its own.
- SAM.gov terms. The public CSV/API data is fine to redisplay; some sensitive API fields are restricted to internal government use. We ingest only public Contract Opportunities data.
- Entity registration takes 2-3 weeks to reach 1,000 API calls/day. The CSV path means we are not blocked, but start registration now.
- Verify every published number. The Niche Finder's credibility is the product. No invented stats; everything traces to a re-runnable query.
What we are building next
- Deeper state and city coverage beyond BidNet-dependent states.
- Richer fit scoring with set-aside eligibility and buyer recurrence.
- Package Intel summaries for more solicitations, nightly.
- Strong-fit digests and alert quality improvements.