METHODOLOGY · UPDATED 9 JUN 2026 · DEVKREST ENGINEERING

How QuoteTurbo calculates everything

The data sources, the formulas, and the parts we estimate versus the parts we verify.

Every QuoteTurbo quote starts with two numbers from two published sources. Plan premium from the CMS HealthCare.gov Plan Compare API, pulled from CMS at quote time. Federal poverty guideline from the HHS ASPE annual table, updated every January 1. Everything else, including the plan ranker, is downstream of those two numbers and the IRS premium tax credit schedule for the coverage year. This page documents that pipeline.

The reason this page exists is that brokers, accountants, and consumers occasionally need to verify the math. The numbers on Form 1095-A line 11 should match the SLCSP QuoteTurbo used at quote time. The expected contribution on Form 8962 should match the IRS schedule for the coverage year. When something does not match, the answer is usually a stale rate filing or a rating area shift. This page lists where every number comes from so the mismatch is findable.

Key Takeaways

  • Plan data is pulled from the CMS Marketplace API. Refresh cadence increases during AEP.
  • Subsidy math follows the IRS premium tax credit formula and the HHS poverty guideline table for the coverage year.
  • SLCSP is computed against current CMS plan data for the household rating area.
  • The plan ranker is a rule-based scoring function, not a generative model. Client data is processed within QuoteTurbo infrastructure.
  • The numbers QuoteTurbo displays should reconcile to Form 1095-A line 11 and Form 8962. Discrepancies trace to upstream filings, not the calculation.

Data sources, in order of use

Four upstream sources feed every quote. Three are federal government data. One is internal broker entry. All four are documented below with the refresh cadence and the calculations they support.

Data sourceUpstream ownerRefresh cadenceDrives
Marketplace plan catalogCMS HealthCare.gov Plan Compare APIPulled from CMS at quote time; cadence increases during AEPPlan IDs, premium by household, metal tier, network, formulary
Federal poverty guidelinesHHS ASPE annual publicationJanuary 1 each yearFPL percentage, CSR thresholds, APTC eligibility
Premium tax credit tableIRS Revenue Procedure (annual)January 1 each year (or per Congressional change)Expected contribution by FPL band, APTC calculation
Rating area boundariesCMS state rate review filingsAnnual, plus mid year carrier filingsZIP to rating area lookup, plan availability

Two of these refresh on a fixed annual calendar. Two refresh on CMS pace, which is unpredictable. The Marketplace plan catalog is pulled from CMS at quote time so that results stay aligned with Healthcare.gov during AEP, when carrier filings can shift mid week.

The subsidy math, line by line

Premium tax credit is defined by 26 U.S. Code Section 36B. The calculation has six lines. Each line below maps to a specific input or formula. No proprietary math, no QuoteTurbo adjustments.

LineFormulaSource
Modified Adjusted Gross Income (MAGI)Household projected MAGI for the coverage yearBroker entry (verified against client tax projection)
FPL percentageMAGI / FPL for household size, regionHHS ASPE poverty guidelines table
Expected contributionFPL percentage mapped to IRS contribution scheduleIRS Revenue Procedure for the coverage year
SLCSPSecond lowest cost Silver plan in household rating areaCMS Marketplace plan catalog, filtered to household composition
Monthly APTCmax(0, SLCSP premium - expected contribution)Standard PTC formula per 26 U.S. Code Section 36B
Net premiumPlan premium - APTC (clamped at zero)Per selected plan, every metal tier

The standalone walkthroughs are on the ACA subsidy calculator, APTC calculator, and SLCSP calculator. Each calculator runs the same six lines against live inputs.

What is verified versus what is estimated

Two categories of number show up in the product. Brokers should know the difference before sending a PDF to a client.

Verified: plan premium, plan ID, metal tier, network, formulary, deductible, out of pocket maximum, copay schedule. All sourced directly from the current CMS plan filing for the household rating area. These should match Healthcare.gov exactly at quote time.

Estimated:APTC, CSR eligibility, net premium after subsidy, reconciliation outcome at tax time. These depend on the household's projected income for the coverage year. If the projection changes mid year, the estimates change with it. The IRS reconciliation on Form 8962 is the final number.

Every QuoteTurbo export labels which numbers are verified and which are estimated, so clients understand the difference at decision time.

How the plan ranker works

The full breakdown is on the plan ranker pillar. The short version: it is a rule-based scoring function, not a generative model. Four properties matter for verification.

Inputs are explicit

Household composition, MAGI, ZIP, doctors, prescriptions, and broker preference weights. Nothing implicit. Nothing inferred from third party data.

Ranking is rule based

The ranking is a scoring function, not a generative model. Each plan gets a score on premium, network match, and formulary match. The score determines the order.

Reasoning is templated

The one paragraph reason per recommendation is composed from the same score the ranker produced. The reasoning is templated from plan attributes, not free-form generated text.

Override is recorded

When a broker overrides the recommendation, the override is recorded against the quote so the broker can review past decisions.

Brokers can audit any recommendation. The reasoning paragraph points at the specific score components that drove the ranking. If the ranker is wrong for a household, the broker overrides; the override is recorded against the quote.

Data handling and privacy

Client data (names, dates of birth, SSN where collected, household and medical information entered during quoting) is encrypted in transit using TLS 1.2 or higher. The QuoteTurbo infrastructure runs on US based cloud providers. The plan ranker runs against household inputs and CMS plan data within QuoteTurbo infrastructure.

QuoteTurbo does not sell broker, client, or quote data. No advertising or third party tracking pixels run on pages that handle client data. Brokers with questions about Business Associate Agreements should contact QuoteTurbo through the account email; BAAs are evaluated on a per-agency basis.

Source documents

Every claim on this page traces to a primary source. Brokers, accountants, and journalists verifying the math should bookmark the following.

For the team behind the methodology and the funding model commitments that back these claims, see the about page.

Change log

2026-06-02. Initial publication. Sources and formulas reflect 2026 coverage year.

2026-06-09.Removed unverified cadence claims ("hourly during AEP", "no cached snapshots") pending backend documentation. Softened ranker description from "constrained ranker with explainability traces" to "rule-based scoring function". Removed absolute privacy phrasing in favor of operational language.

Future changes (new CMS API versions, IRS schedule updates, ranker revisions) will be logged here with the date and a one line summary of what changed. This is the page brokers and journalists cite, so the change history is part of the commitment.

Forward-looking roadmap statements reflect current intent and may change without notice.

Footer background
QuoteTurbo

Crafted with ❤️ by Team Devkrest

Copyright QuoteTurbo 2026

QuoteTurbo is a software tool. It does not provide insurance advice. Coverage decisions rest with the broker and the consumer.