Which markets deserve the next marketing dollar?
A 0–100 Marketing Opportunity Score for all 3,199 US counties, in two modes. MA mode ranks general Medicare Advantage marketing opportunity; D-SNP mode re-weights the same market toward the dual-eligible opportunity. Every component, every weight and every input that is not yet computable is on this page. The score is arithmetic over published CMS counts, not a prediction and not a model output in the machine-learning sense.
D-SNP county rankings partially inherit state-level capture, concentration and integration values. Use them to prioritize markets for review, not as county-level estimates of those three measures.
Enrollment, penetration, momentum and dual counts are the April 2026 CMS county file. D-SNP capture proxy, competitive openness and integration readiness are state-grain figures from the CMS SNP Comprehensive Report, applied to the counties in that state and labeled as such.
Rank the markets
| # | County | Score | Band | Medicare | MA & other health plan share | 12-mo pts | Duals | Dual density |
|---|---|---|---|---|---|---|---|---|
| 1 | Merced County, CA | 90 | HIGH | 43,168 | 29.5% | +1.7 | 14,720 | 34.1% |
| 2 | Tulare County, CA | 87 | HIGH | 70,870 | 37.3% | +1.4 | 26,131 | 36.9% |
| 3 | Imperial County, CA | 85 | HIGH | 37,506 | 38.0% | +1.3 | 17,492 | 46.6% |
| 4 | Monterey County, CA | 84 | HIGH | 74,748 | 16.9% | +0.6 | 16,728 | 22.4% |
| 5 | Kings County, CA | 81 | HIGH | 19,718 | 37.8% | +1.4 | 6,409 | 32.5% |
| 6 | Suffolk County, NY | 80 | HIGH | 334,556 | 30.1% | +0.7 | 54,820 | 16.4% |
| 7 | Grant County, WA | 80 | HIGH | 18,194 | 37.3% | +2.3 | 3,554 | 19.5% |
| 8 | Allegany County, MD | 80 | HIGH | 17,391 | 18.8% | +1.1 | 3,294 | 18.9% |
| 9 | Okanogan County, WA | 80 | HIGH | 12,192 | 29.6% | +2.2 | 2,306 | 18.9% |
| 10 | Nassau County, NY | 79 | HIGH | 303,097 | 33.4% | +0.7 | 54,398 | 17.9% |
| 11 | District Of Columbia, DC | 79 | HIGH | 97,554 | 34.5% | +0.6 | 34,310 | 35.2% |
| 12 | Orange County, NY | 79 | HIGH | 72,788 | 38.2% | +1.0 | 13,792 | 18.9% |
| 13 | Benton County, WA | 79 | HIGH | 41,639 | 32.4% | +1.4 | 5,808 | 13.9% |
| 14 | Sullivan County, NY | 79 | HIGH | 18,350 | 38.5% | +1.4 | 4,880 | 26.6% |
| 15 | Marion County, IL | 79 | HIGH | 10,000 | 31.9% | +2.2 | 2,337 | 23.4% |
| 16 | San Francisco County, CA | 78 | HIGH | 159,389 | 51.3% | +1.3 | 55,648 | 34.9% |
| 17 | Caddo Parish, LA | 78 | HIGH | 52,706 | 55.3% | +4.4 | 13,691 | 26.0% |
| 18 | Wicomico County, MD | 78 | HIGH | 22,127 | 27.4% | +1.1 | 3,698 | 16.7% |
| 19 | Mckinley County, NM | 78 | HIGH | 12,558 | 33.7% | +3.1 | 2,126 | 16.9% |
| 20 | Westchester County, NY | 77 | HIGH | 198,268 | 39.1% | +0.7 | 37,921 | 19.1% |
| 21 | Rockland County, NY | 77 | HIGH | 60,258 | 32.6% | +0.5 | 12,944 | 21.5% |
| 22 | Lafayette Parish, LA | 77 | HIGH | 47,851 | 41.2% | +1.0 | 9,107 | 19.0% |
| 23 | Yuma County, AZ | 77 | HIGH | 46,259 | 52.4% | +1.4 | 15,090 | 32.6% |
| 24 | Calcasieu Parish, LA | 77 | HIGH | 40,212 | 45.6% | +1.5 | 7,845 | 19.5% |
| 25 | Franklin County, WA | 77 | HIGH | 12,733 | 37.3% | +2.1 | 2,487 | 19.5% |
Top 25 in MA mode, rendered without JavaScript. With JavaScript the table becomes the full 3,199-county explorer with filters, mode switching, a scatter view and export.
Opportunity against market size
How the score is built
Each available component is converted to a percentile rank against every scored county, ties sharing the mean rank. Percentiles are combined with published weights, and the weights are renormalized over the components that are actually available rather than treating a missing component as a zero. Percentile ranking rather than min-max scaling is deliberate: county MA enrollment growth spans roughly −89% to +132%, and min-max would let a handful of extreme counties compress everything else into a narrow band.
How this score is calculated
Percentile rank of every available component against all scored counties, combined with published weights renormalized over the available components. Deterministic arithmetic over published CMS counts. Not a prediction, not AI.
Version 0.1, generated 2026-08-26 by tools/build_marketing_data.py. Status MODELED. Bands: HIGH 75-100, MODERATE 50-74, LOW 25-49, MINIMAL 0-24. Counties below 2,000 Medicare beneficiaries are flagged low reliability, because a percentile rank there moves sharply on a small absolute change. They are flagged, not hidden.
| Component | Weight | Direction | Grain | Status |
|---|---|---|---|---|
| Addressable Medicare population Total Medicare beneficiaries in the county. Reach is bounded by this number before any creative decision. | 0.20 | higher is better | County | available |
| Health-plan share headroom Remaining share relative to the CMS MA-and-other-health-plan measure. This is a directional headroom indicator, not a count of MA-eligible beneficiaries available to convert. | 0.15 | higher is better | County | available |
| MA & other health plan share momentum Change in MA & other health plan share over twelve months, in percentage points. A market already moving costs less to move further. | 0.09 | higher is better | County | available |
| MA & other health plan enrollment growth Year over year change in MA & other health plan enrollment in the county. | 0.06 | higher is better | County | available |
| Dual / D-SNP opportunity Dual-eligible density. Duals hold year-round enrollment rights, so they change the shape of a media plan, not only its size. | 0.10 | higher is better | County | available |
| Competitive openness How concentrated the carrier field is. Needs the CMS CPSC contract/plan/county enrollment file, not yet ingested. | 0.15 | lower is better | County | not yet |
| Market disruption Plan exits and non-renewals displacing members in this county. Needs the CMS service-area and plan-year crosswalk files, not yet ingested. | 0.10 | higher is better | County | not yet |
| Product differentiation room How much benefit design space is unclaimed locally. Needs the MA Landscape and Plan Benefit Package files, not yet ingested. | 0.10 | higher is better | County | not yet |
| Quality / Stars context Star ratings of contracts serving the county. Needs the contract-level Star Ratings files, not yet ingested. | 0.05 | lower is better | County | not yet |
Any change to a weight, a component, the normalization method or a band threshold increments the version and requires a changelog entry. A score that changes without a version change is a bug.
D-SNP mode components
| Component | Weight | Direction | Grain | Status |
|---|---|---|---|---|
| Dual-eligible population Dual-eligible beneficiaries in the county, used as a market-size input. Actual D-SNP eligibility, plan availability and enrollment opportunity vary by state, eligibility category and plan design. | 0.35 | higher is better | County | available |
| Full-dual share Share of duals with full Medicaid benefits. Full duals are the core D-SNP eligibility group. | 0.10 | higher is better | County | available |
| D-SNP capture-proxy headroom Dual population minus reported D-SNP enrollment, as a share. Cross-vintage proxy: August 2026 SNP enrollment over April 2026 dual population. State grain: CMS publishes SNP enrollment by state, not by county. | 0.25 | higher is better | State, applied to county | available |
| D-SNP competitive openness Concentration of D-SNP enrollment across parent organizations in the state (HHI over single-state plans). State grain. | 0.15 | lower is better | State, applied to county | available |
| Integration readiness FIDE and HIDE share of state D-SNP enrollment. High integration markets reward a different message than coordination-only markets. State grain. | 0.10 | higher is better | State, applied to county | available |
| MA & other health plan share momentum Change in MA & other health plan share over twelve months, in percentage points. | 0.05 | higher is better | County | available |
D-SNP mode is fully computable today because CMS publishes SNP enrollment, integration status and plan detail. Three of its six components are state-grain, which is stated on the component and is the honest limit of the D-SNP view: applying a state capture rate to a county assumes the county behaves like its state.
What is missing from MA mode, and what it waits on
| Component not yet computable | Published weight | What it waits on |
|---|---|---|
| Competitive openness | 0.15 | How concentrated the carrier field is. Needs the CMS CPSC contract/plan/county enrollment file, not yet ingested. |
| Market disruption | 0.10 | Plan exits and non-renewals displacing members in this county. Needs the CMS service-area and plan-year crosswalk files, not yet ingested. |
| Product differentiation room | 0.10 | How much benefit design space is unclaimed locally. Needs the MA Landscape and Plan Benefit Package files, not yet ingested. |
| Quality / Stars context | 0.05 | Star ratings of contracts serving the county. Needs the contract-level Star Ratings files, not yet ingested. |
Weights are renormalized over the available components, so the score is a weighted average of what exists rather than a score with silent zeros. Coverage: 5/9 components, applied weight 0.60 of 1.00.
So what / Now what Interpretation
On the current inputs the highest-scoring large market is Merced County, CA at 90, with 43,168 Medicare beneficiaries, 70.5% of them not yet in a Medicare health plan and +1.7 points of twelve-month movement. That combination, large and still converting, is the profile the MA weights are built to surface. It is also exactly the profile most likely to be contested, and the score cannot see that yet.
- Filter to the counties you can actually serve before reading the ranking. A national top-25 is a conversation starter, not a media plan.
- Read the component bars, not the total. Two counties scoring 78 for different reasons need different campaigns: headroom markets need conversion messaging, momentum markets need share-of-voice against whoever is already winning.
- Switch to D-SNP mode if duals are in scope. The ranking changes substantially, because dual density and state capture proxy are not correlated with general MA opportunity.
- Save the filtered set as a marketing project, then take the same counties to AEP target markets so field and media agree on geography.
Limitations you should carry into the meeting
- Four of the nine MA components are not yet computable. The score therefore measures market attractiveness for marketing investment, not competitive winnability.
- D-SNP capture, competitive openness and integration readiness are published by CMS at state level. Applying them to a county assumes the county behaves like its state.
- Carrier concentration attributes only single-state D-SNP plans. Plans sold across several states carry one CMS enrollment figure that cannot be split by state; the unattributed share is reported per state.
- Counties below the reliability floor are flagged, not hidden.
- Nothing here describes an individual beneficiary. Every figure is a published geography-level aggregate.
- National D-SNP capture is computed on the same basis as /market/snp/ (all state rows, Puerto Rico included). Puerto Rico's own ratio is suppressed everywhere it would be displayed, because block-grant Medicaid makes its dual denominator non-comparable.
Sources and vintage
| Dataset | Source | Vintage | Status |
|---|---|---|---|
| MA and other health plan enrollment and share by county | CMS Medicare Monthly Enrollment (BENE_GEO_LVL=County) | April 2026 | live |
| Dual-eligible counts by county | CMS Medicare Monthly Enrollment (DUAL_TOT_BENES / FULL_DUAL_TOT_BENES) | April 2026 | live |
| SNP enrollment, integration and plans | CMS SNP Comprehensive Report (SNP_REPORT_PART_17) | August 2026 | live |
| Marketing Opportunity Score | Derived, tools/build_marketing_data.py v0.1 | 2026-08-26 | modeled |
Full methodology: Marketing Opportunity Score · site methodology.