Programmatic SEO site turning USASpending.gov micro-purchase data into actionable sales intelligence pages for government contractors.

Every year, federal agencies make millions of purchase card transactions under the $10K micro-purchase threshold. These transactions are individually reported to USASpending.gov with NAICS codes, PSC (Product & Service Codes), awarding agency, recipient name, and dollar amounts — but nobody aggregates them at the product/agency/vendor intersection level.

The micro-purchase layer is the federal government’s “impulse buy” signal. It reveals what agencies need urgently, what they’re buying repeatedly, and which vendors have incumbent relationships — all without the 6-18 month procurement cycle overhead of formal contracting.

The gap: Small and mid-size government contractors have no tooling to monitor micro-purchase trends by product category + agency. They’re flying blind on the $50B+ annual micro-purchase market.

The product: Programmatic pages that answer the query pattern: ”[Agency] [Product Category] contracts” — e.g., “Department of Defense IT cable purchases,” “VA medical supply vendors,” “DOE lab equipment spending.”

Primary: Small business government contractors doing <$5M/year in federal sales. Typically 5-50 employees. Searching Google for phrases like “how to sell to the VA” or “DOD purchase card vendors” or “GSA Schedule IT equipment.”

Secondary: Business development reps at mid-market GovCon firms ($5M-$50M) tracking agency buying patterns for capture planning.

Tertiary: GovCon consultants and GSA Schedule advisors who need data to justify their fees.

/agency/{agency-slug}/{psc-category-slug}

Examples:

• /agency/department-of-defense/it-equipment
• /agency/veterans-affairs/medical-supplies
• /agency/department-of-energy/laboratory-equipment

Target queries: “VA medical supply contracts,” “DOD IT equipment spending,” “DOE lab equipment purchases”

Estimated addressable pages: ~1,200 (24 major agencies × 50 meaningful PSC categories)

/vendor/{vendor-slug}

Examples:

• /vendor/grainger-inc
• /vendor/amazon-business
• /vendor/mcmaster-carr

Target queries: “Grainger government contracts,” “Amazon Business federal sales”

Estimated addressable pages: ~5,000 (top vendors by micro-purchase volume)

/industry/{naics-code}/{agency-slug}

Examples:

• /industry/334111/department-of-defense (Electronic Computer Manufacturing → DOD)
• /industry/339112/veterans-affairs (Surgical and Medical Instrument Manufacturing → VA)

Target queries: “NAICS 334111 government contracts,” “medical device manufacturers selling to VA”

Estimated addressable pages: ~2,400 (120 relevant NAICS × 20 agencies)

/insights/{year}/{topic-slug}

Examples:

• /insights/2025/fastest-growing-micro-purchase-categories
• /insights/2025/top-100-micro-purchase-vendors
• /insights/2025/agency-spending-shifts

Estimated pages: 20-30 evergreen + annual refresh

Base URL: https://api.usaspending.gov

No authentication required. No rate limit published, but practical limit is ~100 requests/minute based on community reports.

The critical filter for our niche: contract awards ≤ $10,000. Applied via award_amounts:

{
  "filters": {
    "award_type_codes": ["A", "B", "C", "D"],
    "award_amounts": [
      { "lower_bound": 0, "upper_bound": 10000 }
    ],
    "time_period": [
      { "start_date": "2024-10-01", "end_date": "2025-09-30" }
    ]
  }
}

Award type codes A-D = procurement contracts (excludes grants, loans, financial assistance).

USASpending updates contract transaction data daily (File D1 from FPDS). Practical refresh cadence for this product:

• Aggregate tables (agency/PSC/NAICS rollups): Weekly via Cron Trigger
• Vendor profiles: Weekly
• Trend/insight pages: Monthly
• Full historical backfill: Quarterly (bulk download endpoint)

┌─────────────────────────────────────────────────────┐
│                   Cloudflare Edge                     │
│                                                       │
│  ┌──────────┐   ┌──────────────┐   ┌──────────────┐ │
│  │  Worker   │   │  Worker      │   │  Worker      │ │
│  │  (Hono)   │   │  (Cron)      │   │  (Sitemap)   │ │
│  │  Web App  │   │  Data Ingest │   │  Generator   │ │
│  └────┬──┬──┘   └──────┬───────┘   └──────────────┘ │
│       │  │              │                             │
│  ┌────┴──┴──────────────┴────────────────────┐       │
│  │              D1 Database                   │       │
│  │  (Drizzle ORM)                             │       │
│  │                                            │       │
│  │  Tables:                                   │       │
│  │  - agencies                                │       │
│  │  - psc_codes                               │       │
│  │  - naics_codes                             │       │
│  │  - micro_purchases (partitioned by FY)     │       │
│  │  - agency_psc_rollups                      │       │
│  │  - agency_naics_rollups                    │       │
│  │  - vendor_profiles                         │       │
│  │  - vendor_agency_rollups                   │       │
│  │  - page_metadata (SEO titles/descriptions) │       │
│  └───────────────────────────────────────────┘       │
│                                                       │
│  ┌───────────────┐   ┌────────────────────┐          │
│  │  Workers AI   │   │  KV Namespace      │          │
│  │  (Page copy   │   │  (HTML cache,      │          │
│  │   generation) │   │   sitemap cache)   │          │
│  └───────────────┘   └────────────────────┘          │
│                                                       │
│  ┌───────────────┐                                   │
│  │  R2 Bucket    │                                   │
│  │  (Bulk CSV    │                                   │
│  │   archives)   │                                   │
│  └───────────────┘                                   │
└─────────────────────────────────────────────────────┘

// schema.ts

import { sqliteTable, text, integer, real, index } from 'drizzle-orm/sqlite-core';

export const agencies = sqliteTable('agencies', {
  id: integer('id').primaryKey({ autoIncrement: true }),
  toptierId: integer('toptier_id'),          // USASpending toptier agency ID
  toptierCode: text('toptier_code'),         // e.g., "097"
  name: text('name').notNull(),              // e.g., "Department of Defense"
  abbreviation: text('abbreviation'),        // e.g., "DOD"
  slug: text('slug').notNull().unique(),     // e.g., "department-of-defense"
});

export const pscCodes = sqliteTable('psc_codes', {
  id: integer('id').primaryKey({ autoIncrement: true }),
  code: text('code').notNull().unique(),      // e.g., "7035"
  description: text('description').notNull(), // e.g., "ADP Input/Output Equipment"
  categorySlug: text('category_slug'),        // e.g., "it-equipment"
  categoryName: text('category_name'),        // e.g., "IT Equipment"
  parentCode: text('parent_code'),            // PSC hierarchy
});

export const naicsCodes = sqliteTable('naics_codes', {
  id: integer('id').primaryKey({ autoIncrement: true }),
  code: text('code').notNull().unique(),
  description: text('description').notNull(),
  slug: text('slug').notNull(),
  sectorCode: text('sector_code'),            // 2-digit sector
  sectorName: text('sector_name'),
});

export const microPurchases = sqliteTable('micro_purchases', {
  id: integer('id').primaryKey({ autoIncrement: true }),
  awardId: text('award_id').notNull(),
  agencyId: integer('agency_id').references(() => agencies.id),
  pscCode: text('psc_code'),
  naicsCode: text('naics_code'),
  recipientName: text('recipient_name'),
  recipientUei: text('recipient_uei'),        // Unique Entity ID (replaced DUNS)
  amount: real('amount').notNull(),
  actionDate: text('action_date'),            // YYYY-MM-DD
  fiscalYear: integer('fiscal_year'),
  description: text('description'),
  placeState: text('place_state'),            // Place of performance
  placeCity: text('place_city'),
}, (table) => ({
  agencyPscIdx: index('idx_agency_psc').on(table.agencyId, table.pscCode),
  agencyNaicsIdx: index('idx_agency_naics').on(table.agencyId, table.naicsCode),
  recipientIdx: index('idx_recipient').on(table.recipientUei),
  fyIdx: index('idx_fy').on(table.fiscalYear),
}));

// Pre-computed rollup tables (materialized by cron worker)
export const agencyPscRollups = sqliteTable('agency_psc_rollups', {
  id: integer('id').primaryKey({ autoIncrement: true }),
  agencyId: integer('agency_id').references(() => agencies.id),
  pscCode: text('psc_code'),
  categorySlug: text('category_slug'),
  fiscalYear: integer('fiscal_year'),
  totalAmount: real('total_amount'),
  transactionCount: integer('transaction_count'),
  uniqueVendors: integer('unique_vendors'),
  topVendorName: text('top_vendor_name'),
  topVendorAmount: real('top_vendor_amount'),
  avgTransactionSize: real('avg_transaction_size'),
  yoyGrowthPct: real('yoy_growth_pct'),       // Year-over-year change
  updatedAt: text('updated_at'),
}, (table) => ({
  slugIdx: index('idx_rollup_slug').on(table.agencyId, table.categorySlug, table.fiscalYear),
}));

export const vendorProfiles = sqliteTable('vendor_profiles', {
  id: integer('id').primaryKey({ autoIncrement: true }),
  uei: text('uei').notNull().unique(),
  name: text('name').notNull(),
  slug: text('slug').notNull().unique(),
  totalMicroPurchaseAmount: real('total_micro_purchase_amount'),
  totalTransactions: integer('total_transactions'),
  agencyCount: integer('agency_count'),       // How many agencies buy from them
  topAgencyName: text('top_agency_name'),
  topPscCategory: text('top_psc_category'),
  firstSeen: text('first_seen'),              // Earliest transaction date
  lastSeen: text('last_seen'),                // Most recent transaction
  updatedAt: text('updated_at'),
});

// src/index.ts

import { Hono } from 'hono';

const app = new Hono();

// Homepage
app.get('/', (c) => { /* Landing page with search + top agencies */ });

// Agency + PSC Category pages (primary SEO target)
app.get('/agency/:agencySlug/:pscSlug', async (c) => {
  const { agencySlug, pscSlug } = c.req.param();
  // Query agency_psc_rollups + top vendors for this combo
  // Render programmatic page with:
  //   - Total spend, transaction count, vendor count
  //   - YoY trend (table or simple bar)
  //   - Top 10 vendors by amount
  //   - Recent transactions sample
  //   - "How to sell [category] to [agency]" generated copy
});

// Agency overview
app.get('/agency/:agencySlug', async (c) => {
  // Top PSC categories for this agency
  // Total micro-purchase volume + trends
  // Top vendors
});

// Vendor profile pages
app.get('/vendor/:vendorSlug', async (c) => {
  // Which agencies buy from this vendor
  // What PSC categories
  // Transaction history
  // Competing vendors in same categories
});

// NAICS industry pages
app.get('/industry/:naicsCode/:agencySlug?', async (c) => {
  // Industry spending by agency
  // Related PSC categories
});

// Insight/trend pages
app.get('/insights/:year/:topicSlug', async (c) => {
  // Pre-computed trend analysis
});

// Sitemap
app.get('/sitemap.xml', async (c) => { /* Dynamic sitemap from KV cache */ });
app.get('/sitemap/:segment.xml', async (c) => { /* Segmented sitemaps */ });

export default app;

// src/cron.ts — runs weekly

export default {
  async scheduled(event: ScheduledEvent, env: Env) {
    switch (event.cron) {
      case '0 3 * * 1': // Monday 3am UTC
        await ingestRecentTransactions(env);
        await recomputeRollups(env);
        await invalidatePageCache(env);
        break;
      case '0 4 1 * *': // 1st of month
        await generateTrendInsights(env);
        await regenerateSitemaps(env);
        break;
    }
  }
};

async function ingestRecentTransactions(env: Env) {
  // 1. Query USASpending /api/v2/search/spending_by_award/
  //    Filter: award_amounts ≤ $10K, last 7 days
  //    Paginate through all results (limit: 100 per page)
  // 2. Upsert into micro_purchases table
  // 3. Log transaction count for monitoring
}

async function recomputeRollups(env: Env) {
  // SQL aggregation queries against micro_purchases
  // INSERT OR REPLACE into agency_psc_rollups
  // INSERT OR REPLACE into vendor_profiles
  // Compute YoY growth percentages
}

Each programmatic page needs enough unique, useful content to rank. Here’s the template for the primary page type (Agency + PSC):

<h1>{Agency Name} {PSC Category} Spending — Micro-Purchase Intelligence</h1>

<p class="lead">
  In FY{year}, {agency} made {count} micro-purchase transactions
  totaling ${amount} in {category}. Here's what they're buying,
  from whom, and what it means for vendors.
</p>

[Summary Stats Bar]
  Total Spend | Transaction Count | Avg Order Size | Active Vendors | YoY Change

<h2>Top Vendors</h2>
[Table: Vendor Name | Total Amount | # Transactions | Avg Order | % of Category]

<h2>Spending Trend</h2>
[Simple HTML bar chart: quarterly spend for last 3 fiscal years]

<h2>Recent Transactions</h2>
[Table: Date | Amount | Vendor | Description (from award description field)]

<h2>How to Sell {Category} to {Agency}</h2>
[Workers AI generated paragraph — practical guidance on:]
  - Relevant GSA Schedule categories
  - Purchase card ordering procedures for this agency
  - SAM.gov registration requirements
  - Set-aside eligibility (SDVOSB, 8(a), HUBZone)

<h2>Related Categories</h2>
[Internal links to adjacent PSC categories for this agency]

<h2>About This Data</h2>
[Boilerplate: Source attribution, update cadence, methodology note]

Use @cf/meta/llama-3.1-8b-instruct for generating the “How to Sell” section per page. Template prompt:

You are a government contracting advisor. Write 150 words of practical
guidance for a small business that wants to sell {psc_category_name}
to {agency_name}. Include: relevant GSA Schedule numbers, the agency's
typical purchasing process for items under $10,000, and any relevant
set-aside programs. Be specific and actionable. Do not use marketing language.

Cache the output in KV with a 30-day TTL. Regenerate monthly.

• Goal: 8,000+ indexed pages, 5K-15K monthly organic visits
• Revenue: $0 (traffic accumulation)

• Product: “Agency Buyer Profile” PDF downloads (gated)
• Content: Expanded version of page data + procurement calendar + buyer contact info guidance
• Revenue: Email list building for Phase 3

• Tier 1 — Free: Basic page views (drives SEO)
• Tier 2 — Pro ($49/mo): CSV exports, email alerts when new transactions match saved filters, 12-month historical data
• Tier 3 — Agency ($199/mo): Full API access, custom agency watch lists, vendor competitive alerts, bulk data exports

• Product: “GovPurchase Intel for Capture Teams” — white-label reports for BD teams
• Revenue: $2K-5K/month per enterprise account
• Channel: Direct sales to GovCon BD consultants (referral fees)

• Month 6: $500/mo (early Pro subscribers)
• Month 12: $3K-5K/mo (Pro + Agency mix)
• Month 24: $10K-20K/mo (enterprise accounts + self-serve)

• Set up Cloudflare Worker project (Hono + Drizzle + D1)
• Create D1 schema, run migrations
• Build reference data loaders (agencies, PSC codes, NAICS codes from USASpending API)
• Build initial bulk download pipeline (FY2023 + FY2024 micro-purchases)
• Store bulk CSVs in R2, parse and load into D1

• Build Hono routes for agency + PSC pages
• Build rollup computation queries
• HTML templates (server-rendered, minimal CSS — TailwindCSS via CDN)
• Agency overview pages
• Vendor profile pages
• Basic internal linking structure

• Dynamic sitemap generation (segmented by page type)
• Meta tags, canonical URLs, Open Graph
• robots.txt, structured data (Dataset schema.org markup)
• Workers AI “How to Sell” copy generation for top 200 pages
• Submit sitemaps to Google Search Console

• Cron worker for weekly data ingest
• KV caching layer for rendered pages
• Error monitoring, data freshness checks
• Landing page with search functionality
• GA4 or Plausible analytics integration

• Email capture (ConvertKit or Buttondown)
• “Agency Buyer Profile” PDF generation (first 5 agencies)
• Stripe checkout for Pro tier
• Basic alert system (weekly email digest of new transactions matching filters)

Our wedge: Nobody is doing programmatic aggregation of the micro-purchase layer. The incumbents are all focused on contracts >$250K. We own the bottom of the market.

Based on RecallRadar experience, the critical milestone is getting 100+ pages indexed within 60 days of launch. Tactics:

1. Submit sitemap to GSC on day 1
2. Manually request indexing for top 50 pages
3. Internal linking density: every page links to 5+ related pages
4. Schema.org Dataset markup on every page
5. One genuine link-magnet insight page (e.g., “Top 100 Federal Micro-Purchase Vendors 2025”) promoted on LinkedIn/Reddit/Twitter

Suggested: govpurchase.intel or micropurchase.report or fedbuys.io

Worth checking availability. The name should signal “government purchasing data” to both users and Google.