Product updates
Product updates
What changed, newest first. Each entry opens with the update itself; open Technical details where you want the implementation, the source file or the audit behind it. Ingest entries are written by the refresh pipeline; prior vintages of each file are archived under data downloads so old citations stay reproducible.
| Date | Type | Change | Vintage |
|---|---|---|---|
| August 28, 2026 | correction | Independent validation pass: the state ranking page and the market comparison were rebuilt on the CMS state records, and the national D-SNP capture proxy was restated from 54.3% to 51.8%.Technical detailsHealth plan share by state was computing every share by summing county rows, and its KPI row and both charts were still rendering the February 2026 series withdrawn by the August 24 audit — which showed Maine as one of the five lowest-penetration states when it is the fifth highest. Compare markets ranked twelve jurisdictions on the same withdrawn series. Both now read the CMS state records for April 2026, the same source the individual state pages and the homepage already used; 36 of 53 published percentages changed, none by more than 0.3 points on the current basis. The national D-SNP capture proxy included Puerto Rico, where CMS records 5,415 duals against 308,050 D-SNP enrollees — a basis the state table already suppressed as non-comparable. Summing over only the jurisdictions whose proxy is published gives 51.8% and 5,767,705 proxy uncaptured duals, replacing 54.3% and 5,465,070. The suppressed value no longer appears in the CSV export, which gains a Also: national ranks are now assigned on the exact share rather than the rounded one (Minnesota moves to 10th, Pennsylvania to 18th); the Opportunity Score no longer imputes zero for the two counties with no April 2025 record; nine superseded files were removed from the live asset path (the archived copies on data downloads are unchanged); and all twenty dataset panels now have machine-readable entries in the source registry. Full report: |
April 2026 · August 2026 |
| August 26, 2026 | feature | Marketing intelligence: a new section for market investment, D-SNP audience, positioning, policy and campaign planning.Technical detailsNine routes under /marketing/, built from data already ingested: the April 2026 CMS county file and the August 2026 CMS SNP Comprehensive Report. Adds a Marketing Opportunity Score (v0.1, MODELED) for all 3,199 counties in two modes: MA mode is computable on 5 of its 9 published components (applied weight 0.60), D-SNP mode on 6 of 6. Every component, weight, grain and missing input renders on the page, and the four missing MA components all describe competitive difficulty, so the MA score ranks market attractiveness rather than winnability. Also adds D-SNP market archetypes by state, carrier concentration computed on single-state plans with the unattributed share reported, a marketer lens over the tracked CMS actions already on the policy tracker (same dates, same statuses), a planning calendar in which every entry is marked statutory, CMS cadence or curated, an aggregate audience view, a business-case model built only from user-entered assumptions, and a 13-section strategy brief. No plan benefit, premium or star figure is estimated: those wait on the CMS Landscape, Plan Benefit Package and Star Ratings files and render as "coming soon" with the exact planned source. Methodology: docs/marketing-intelligence-methodology.md. Existing pages were verified byte-for-byte unchanged apart from the new navigation entries. |
April 2026 county file · August 2026 SNP report |
| August 24, 2026 | data | County data unified on one CMS source; a 76-county dataset withdrawn.Technical detailsAn audit found the extract behind the homepage, the state pages and the county page heroes
could not be reproduced from any CMS file: denominators differed from the April 2026 CMS
county file by 1.4× to 995×, penetration by a mean of 8.5 and up to 35.5 points,
one county carried the wrong FIPS code, and the year-by-year history was smooth in a way
observed data is not. Every county figure now comes from the April 2026 CMS Medicare Monthly
Enrollment file, and every key metric carries a “Why trust this number?”
receipt with its source, numerator, denominator, formula and vintage. Carrier counts, leading
carrier and the 2023–2026 trajectory were removed rather than restated, because no CMS
source for them exists here yet. State and national figures were, on that date, aggregated from the same county file, so national penetration read 51.6% on a stated basis rather than 57.8% on an unreconciled one. Superseded on August 26, 2026: the direct CMS record is now canonical for its own geography and the national share is 51.1%; county sums are published only as a reconciliation. Read the audit. |
April 2026 |
| August 24, 2026 | fix | Fixes to scripts and styles were taking up to a day to reach returning
visitors.Technical detailsStylesheets and scripts were cached for 24 hours under filenames that
never changed, so a corrected file could be live on the server while a browser that had
visited earlier kept running the previous one — which is why the county search
appeared unfixed after the fix shipped. Every script and stylesheet reference now carries
a content fingerprint, so a changed file is fetched immediately and an unchanged one stays
cached. Data files are now revalidated on every load, so a CMS refresh is never served
from a stale copy. |
— |
| August 24, 2026 | fix | The mobile menu was opening as a 33-pixel strip.
The sticky header’s backdrop blur made it the positioning context for the
drawer inside it, so the drawer sized itself against the 56-pixel header instead
of the screen.Technical detailsIt now fills the screen, carries the county search at the top, and
closes on Escape or the back gesture. Separately, 156 data tables across 136 pages
had no horizontal scroll container and were stretching pages sideways on phones;
they now scroll on their own. Tap targets in the header and on county action bars
were raised to the 44-pixel accessibility minimum. |
— |
| August 24, 2026 | site | Everything is free during the beta; a free account lifts the limits.Technical detailsPrices are removed from all three tiers. What changes what you can do is now whether you
have created a free account, not what you have paid: signed out you get 3 saved markets,
2 saved analyses, 1 saved alert rule (checked in this browser, not delivered) and a 50-row sample CSV; a free account raises those to
50, 50, 25 and the complete file. Creating one takes an email address and no password
— the site is static, so a password checked in your browser would protect nothing,
and we would rather say that than show a login that does not exist. Every county, state,
methodology and glossary page stays public and indexable either way; the gate is on
saving and exporting, never on reading. See what an account changes. |
— |
| August 24, 2026 | fix | The workspace page was not loading anyone’s saved work.Technical detailsIts inline script read the saved-markets store before the deferred script that defines
it had run, so every visitor saw “could not read browser storage” regardless of
their browser. Saved markets and analyses were never lost — they were in storage the
whole time and now render again. |
— |
| August 24, 2026 | site | Tutorials, a mobile-first header, and a readability pass.Technical detailsNew How to read the data — six lessons on the figures that
get misread most often: the two defensible MA & other health plan share bases, percentage points
versus percent, the dual denominator behind D-SNP capture, the three SNP products,
the two-year star-rating lag, and what each provenance tag licenses you to claim.
The header is rebuilt for phones and tablets: a full-height drawer with the search
field, 44 px targets, and the section menus reachable at every width rather than
hidden below 900 px. Heading hierarchy and the small annotation text were put on
one scale, and the vintage stamp repeated on every card in a KPI row is now a single
line beneath the row. No figures changed. |
— |
| August 24, 2026 | ingest | SNP Comprehensive Report refreshed (August 2026). 1,803 plan records: 8,437,333 people in a Special Needs Plan — 6,500,322 D-SNP, 1,806,608 C-SNP, 130,403 I-SNP.Technical detailsIntegration status now 788,394 FIDE, 2,316,408 HIDE, 3,395,520 coordination-only. The SNP page, its state and plan CSVs and the methodology entry all move to this vintage; the prior vintage stays in the archive so existing citations still resolve. |
August 2026 (duals April 2026) |
| July 31, 2026 | ingest | SNP Comprehensive Report ingested (May 2026). All 1,803 plan records now on the site: 8,315,532 people in a Special Needs Plan — 6,469,576 D-SNP, 1,717,346 C-SNP, 128,610 I-SNP — with integration status (FIDE / HIDE / CO), qualifying conditions, contract and inferred parent brand.Technical detailsThe SNP page now carries real enrollment rather than dual-eligible population alone, adds a plan explorer over every SNP in the country, and introduces a D-SNP capture proxy: D-SNP enrollment against the county-derived dual population, 54% nationally and ranging from 19% to 74% across states with a sizeable dual base. Published as plan CSV and state CSV. Plan rows reconcile to the CMS published total exactly. |
May 2026 (duals April 2026) |
| July 31, 2026 | ingest | County dual-eligible layer. Added DUAL_TOT_BENES and FULL_DUAL_TOT_BENES for all 3,199 counties from the April 2026 CMS Monthly Enrollment file: 12.0M dual eligibles nationally, 17.2% of Medicare, 71% of them full duals.Technical detailsNew D-SNP opportunity page with dual density mapped against MA & other health plan share, a 3D view, filters and CSV. Note: this is the dual-eligible population, not SNP plan enrollment — the CMS SNP Comprehensive Report (latest May 2026) is published only as a zip archive and is not yet ingested. |
April 2026 |
| July 31, 2026 | ingest | April 2026 county rebuild. Ingested the CMS Medicare Monthly Enrollment file at county geography for April 2026 (latest published) plus the matching April 2025 month for year-over-year comparison.Technical detailsCoverage goes from 131 curated markets to all 3,199 counties, with penetration, 12-month share change and MA enrollment growth per county. Published as CSV. |
April 2026 (vs April 2025) |
| July 31, 2026 | site | New county explorer: range filters on penetration, momentum and market size, four quick views, a sortable full-country table, CSV export of any filtered slice, a rotatable 3D penetration map and a 3D penetration × momentum × size scatter.Technical detailsNavigation regrouped into four dropdown sections with site-wide search across every page, all 52 state jurisdictions and all 3,199 counties, plus a proper mobile drawer. Plain-language “explain this” panels added to the market pages; glossary terms now show their definition on hover. The 41 state accordions on the homepage that previously dead-ended now link into the explorer. |
— |
| July 31, 2026 | correction | Penetration basis flagged. The county series uses MA_AND_OTH_BENES ÷ TOT_BENES, matching CMS's own published county-penetration definition — 51.6% national for April 2026.Technical detailsThe legacy 57.9% national figure quoted on the state pages could not be reproduced from the monthly file (the same calculation returns 51.1% for February 2026) and is flagged for re-derivation. Until that lands, the county explorer is the authoritative penetration source. See Methodology → penetration basis. |
— |
| July 20, 2026 | site | Second integrity pass: county penetration now shows a graded ±1σ reliability band (replacing the hard 20k cliff) with robust/moderate/thin tiers; the modeled growth-runway score is normalized to the 95th-percentile market so a single outlier no longer sets the scale; and compare mode respects the velocity noise-gate and updates its tab title/description per selection. | — |
| July 20, 2026 | site | Analytics-integrity hardening ahead of full county rollout: YoY velocity now carries a per-county stability band (sqrt(2p(1-p)/N)) and directional color/arrows are suppressed for moves within noise; thin or low-reliability county pages are gated to noindex,follow (and dropped from the sitemap) to protect domain authority; and the modeled growth-runway score now highlights each state's dominant driver by percentile so a high score is explainable, never a black box. | — |
| July 20, 2026 | site | County pipeline extended for full-coverage drill-down: refresh_cms_data.py --counties now layers Census ACS 5-year 65+ population onto every county, and optional CPSC (carrier count, top carrier, D-SNP share) and MA/PDP landscape ($0-premium availability) enrichment when the CMS source files are supplied.Technical detailsEach layer stamps its own provenance; county pages render the new fields only where present, so an unenriched county never shows an invented number. |
— |
| July 18, 2026 | site | County drill-down launched: every state page now links to per-county pages at /market/states/{state}/{county}/ with penetration trajectory, market structure, D-SNP share, So what / Now what, provenance, and neighboring-county links.Technical detailsNew homepage “Find your market” state→county browser (works without JavaScript), a multi-market “Your markets” watchlist, and a Compare-markets tool (/market/compare/) for side-by-side states and counties with shareable URLs. A new full-coverage county ingest (refresh_cms_data.py --counties, CMS Monthly Enrollment county file, registered as ma_county_penetration) extends county pages to every U.S. county on the next data run; small-population counties are flagged for reliability. |
— |
| July 17, 2026 | site | New Sales intelligence hub for payer sales organizations, pointed at AEP 2027 and plan years 2027-2028: a 2027 disruption pipeline (dated CMS events with an AEP 2027 play and a 2028 setup per event), AEP 2027 target markets with a CRM-ready CSV, growth runway index (modeled, labeled), whitespace counties, per-carrier vulnerability battlecards, and a D-SNP deployment map.Technical detailsAlso: downloadable state sales playbook one-pagers, a sales trigger watch in briefings, and glossary entries for AEP, OEP, SEP, LIS, ANOC, disenrollment rate, and runway score. |
— |
| July 17, 2026 | site | Site update: newsletter signup form, About/Privacy/Terms pages, glossary, bulk data-download index, per-vintage CSV archive, Atom feed for briefings, structured data, and social share cards. Star-movement page now shows official October ratings only. | — |
| July 1, 2026 | ingest | CMS policy developments reviewed and updated through July 2026. source → | Through July 2026 |
| June 15, 2026 | ingest | CMS Medicare Monthly Enrollment for February 2026 ingested — state penetration table, national basis figures, and all dependent pages refreshed. source → | February 2026 |
| June 15, 2026 | ingest | withdrawnSuperseded. National MA vs Original Medicare series extended with February 2026 enrollment. source → Superseded by the August 24, 2026 data audit after the county extract behind it failed reproducibility checks. Retained for historical transparency; it does not feed current production analytics. |
Plan year 2026 (Feb 2026 enrollment) |
| May 1, 2026 | ingest | withdrawnSuperseded. County-level market snapshot refreshed for all 3,199 US counties (CPSC files), with 2023–2026 penetration history. source → Superseded by the August 24, 2026 data audit after the county extract behind it failed reproducibility checks. Retained for historical transparency; it does not feed current production analytics. |
Plan year 2026 |
| May 1, 2026 | ingest | withdrawnSuperseded. Carrier footprint by state refreshed from March 2026 contract enrollment. source → Superseded by the August 24, 2026 data audit after the county extract behind it failed reproducibility checks. Retained for historical transparency; it does not feed current production analytics. |
March 2026 |
| May 1, 2026 | ingest | SNP share of MA enrollment by state refreshed from the SNP Comprehensive Report. source → | February 2026 |
| April 15, 2026 | ingest | Modeled PMPM benefit-value series re-estimated against CY2026 landscape files (modeled — directional only). source → | Plan years 2022–2026 |
| April 7, 2026 | ingest | CY2026 Part D benefit parameters and IRA negotiated-price list finalized and ingested. source → | CY2026 final parameters |
| March 15, 2026 | ingest | Dual-eligible density by state refreshed from MMCO 2025 Q4 data. source → | 2025 Q4 |
| February 15, 2026 | ingest | Regulatory reference library updated through the 2027 Advance Notice. source → | Current through the CY2027 Advance Notice (published January 26, 2026) |
| February 10, 2026 | ingest | 2026 Star Ratings (published October 2025) ingested — carrier trajectories and measure-level erosion updated. source → | 2026 Star Ratings (published Oct 2025) |
| January 20, 2026 | ingest | 2026 plan exits and January member-migration reconciliation ingested. source → | Plan year 2026 |
Get the 5 biggest shifts in Medicare data, monthly
One email per CMS release cycle: what changed in the new data, the five largest market shifts, and links to the updated answers. No product news, no filler.
No spam, unsubscribe anytime. See the privacy policy. Latest issue: Data release briefings →