How are Colorado contractors ranked?

What principles guide the rankings?

• Evidence, not opinion. Every figure comes from a public record we can link to.
• Signals, not causation. We show what correlates with winning. We never claim to know why a company won.
• Celebrate rising. We publish leaderboards of growth and new entrants, not lists of who is falling.
• Colorado work. Prime awards count when the work is performed in Colorado, wherever the company is headquartered. Subawards count when the subcontractor is based in Colorado, because USAspending.gov does not report where subaward work is performed.

What time periods are compared?

Every metric compares the trailing 12 months (the 12 months up to the snapshot date) with the prior 12 months (the 12 months before that). Rankings are recalculated weekly, every Monday, and each week's results are saved so trends can be charted.

What does each metric mean?

Obligations (12 months)

Sum of dollars obligated to the company on DoD prime awards performed in Colorado and subawards to Colorado-based subcontractors over the trailing 12 months. Includes modifications and deobligations, so it is a net figure.

Always obligated dollars, never "potential value" or ceiling.

Growth %

(last 12 months − prior 12 months) ÷ prior 12 months.

Only shown when the prior 12 months were at least $250,000. Otherwise the company is labeled New entrant or Emerging.

New wins

New prime awards (base awards, not modifications) plus subawards first reported in the last 12 months. A subaward re-reported later counts once.

Relationship breadth

Distinct prime contractors paying the company through subawards, plus distinct federal agencies (for example Air Force, Army, Space Force) awarding it prime contracts, in the last 12 months.

More relationships usually means a more resilient business.

First win

Earliest federal contract award, subaward or SBIR/STTR award in our data.

Drives the New entrant label: a company is a new entrant when its first win is in the last 12 months.

SBIR path

Highest SBIR/STTR phase the company has reached (I, II or III).

Phase III signals commercialization of research into production work.

How is the momentum score calculated?

Momentum is a score from 0 to 100. Each input is converted to a percentile within the Colorado cohort (0 = lowest, 100 = highest), then weighted:

momentum = 0.4 × growth percentile
         + 0.3 × new wins percentile
         + 0.2 × relationship breadth percentile
         + 0.1 × recency percentile   (how recent the latest win is)

Choices the formula needs that are worth knowing:

• Companies without a growth % (new entrants and emerging companies, whose prior year was under $250,000) get a neutral 50 for growth. Growth percentiles for everyone else are ranked among companies that have a growth %.
• SBIR/STTR awards count toward a company's first win, so an earlier SBIR award means the company is not a new entrant.
• The cohort is every company with a Colorado contract award in the last 12 months, or at least $250,000 in the prior 12 months.
• Rank 1 is the highest momentum. Ties share a rank.

These weights are version 1. We will revisit them after the first month of data and note any change here.

How are the “why they’re winning” signals written?

For the top 50 companies by momentum, each scorecard can show a short summary of signals that correlate with winning: growth, new wins, new prime relationships, how concentrated their customers are, SBIR phase and SAM.gov certifications. Summaries are filled in from fixed sentence templates using only the numbers on the scorecard, so they cannot state anything the data doesn't show, and they are published with each weekly update without hand editing. They describe what correlates with winning, never why a company won. Set-aside status is not a signal yet because the award fields we load don't include it.

Where does the data come from?

Companies are matched across sources by their SAM.gov Unique Entity ID (UEI). When a record has no UEI, we match on the exact company name only when exactly one company fits; otherwise the record is left unlinked rather than guessed.

What are the known limits of the data?

• Subawards are self-reported by prime contractors, often late and sometimes incomplete. A company's subcontract work may be undercounted.
• Small orders and commodities are left out. Prime transactions under $10,000 (the federal micro-purchase threshold) and food and fuel orders (product codes 89 and 91) are excluded. They are over half of all transactions but a fraction of a percent of the dollars, and would otherwise dominate counts of new awards.
• Place of performance is what the contracting office or prime reported. Work split across states may be attributed to one location.
• Recent weeks change. Late reports can raise recent totals after the fact; past snapshots are not rewritten.
• Classified work does not appear in public spending data.
• Certifications are as listed in SAM.gov; expired SBA certifications are excluded.

What if a number looks wrong?

Every figure links to its source record on USAspending.gov or SBIR.gov, so you can check it. If something is wrong, email info@machine94.com with the company and the record.