Fee and price-history fields in the Inventory Listings Feed, about 30% more accurate NeoVIN decodes, MarketCheck Price and MarketMatch for Canada, and 5,369 new dealers.
This release adds fee-aware pricing, a 90-day price history and alternate VIN support to the Inventory Listings Feed, improves NeoVIN decode accuracy by about 30%, launches MarketCheck Price™ and MarketMatch™ for Canada, and adds 5,369 dealers with active listings.
Inventory Listings Feed:price is now the most prominent advertised price, with new fees, is_price_inclusive_fees, price history and alternate_vin fields.
NeoVIN & Enhanced VIN Decode: options arbitration engine, better trim, version and equipment group identification, dealer installed options, new ref_msrp and ref_adopted option fields and new listing_confidence tags.
MarketCheck Price™ and MarketMatch™ for Canada: new marketcheck_price, msrp, vin_score, dealer_ymm_score and vin_rank fields.
Dealer coverage: 5,369 new dealers across cars, auction, FSBO, motorcycle, RV, boat and YME markets.
Following FTC guidance on advertised vehicle pricing, dealers across the industry have been redesigning the price section of their vehicle pages to disclose their fees and to make the price a shopper will pay the most prominent figure. Adoption has been gradual, and dealer pages now present price in several different layouts,
Over the past quarter, MarketCheck has aligned the price field with these layouts as dealers adopted them. price now carries the price the dealer presents as its most prominent advertised price, which is increasingly inclusive of dealer fees.
price is always a value shown on the dealer’s page. MarketCheck does not calculate it.
Where the page shows no total or final price, price is the price the dealer advertises.
Use fees and is_price_inclusive_fees to see whether price already includes the dealer’s fees.
Impact
price reflects the price a shopper sees first on the dealer’s page. As more dealers move to all-inclusive pricing, the share of listings where price includes fees will keep growing. When comparing prices over time, use is_price_inclusive_fees to separate all-in prices from prices before fees.
We have modified the options and features. The updates include:
Addressing missing option packages.
Improving the alignment between features and options.
Updating extraction logic where the data was available but not being mapped correctly.
Impact
While we may not see a major change in overall listing coverage, we expect a significant improvement in options and features quality and accuracy. Fields that may be impacted: options, features.
The total of the dealer fees shown on the vehicle's page, such as documentation or processing fees. Empty when no dealer fee is shown.
Only valid fee values are counted: a fee is not counted more than once, even if it appears several times in the page's price sequence.
Fee-label variations are recognised, including labels without standard fee keywords, labels attached directly to the amount (for example, DocFee, DealerFee), and plural or misspelled forms.
Shows whether the listed price already includes the dealer fees.
true: the dealer fees are already part of the price.
false: the dealer fees are charged on top of the price.
Empty: no dealer fee is shown, or the page does not make clear whether the fees are included.
Impact
Feed consumers can tell whether a listing's price is all-in or excludes fees. This supports all-in price comparisons in line with the FTC advertised-pricing framework.
Shows the past prices of a vehicle (VIN) at the same dealer over the last 90 days.
Each listing record now carries the earlier prices that the same dealer listed for that VIN during this window, so you can see how the price changed before the current listing.
Only prices from the same dealer are included. Listings of the same VIN by other dealers are not part of this history.
Impact
Feed consumers can track price changes for a vehicle at a dealer, such as price drops or increases, without having to rebuild the history from earlier feed files.
The vehicle identifier shown on the page when it is not a valid 17-character VIN, such as a partial VIN or chassis number. Filled only when vin is empty.
This field follows the same maintenance cadence as all other vehicle attributes in the dataset.
Impact
Adds alternate_vin support for vehicles without a manufacturer VIN, enabling alternate VIN lookups.
Previously, NeoVIN stopped at the first listing text that identified any option, and it could only price options that existed in the spec provider's catalog.
Multi-candidate arbitration: every unique option-text group is now scored against the vehicle's reference price, and the decoder keeps the one that reconciles best.
Cross-field arbitration: options text, features text, and seller comments compete in one pool, ranked by how much each field can be trusted.
Corroborated option borrowing: after a winner is picked, options corroborated by other evaluated groups, including used-inventory listings of the same car, can be borrowed to close an undershoot.
Sibling disambiguation with a transactional guard: each stage is checkpointed and reverted if it makes the decode worse.
Price-correction passes: sticker line prices, base price, and destination charge are adopted where the sticker corroborates them. Phantom engine prices, double-counted package siblings, and duplicated paint charges are zeroed. Glued multi-price strings are split into separate options.
Results (9.36M VINs, production vs new logic)
Metric
Before
After
VINs improved / regressed
-
24.5% / 1.13%
Median gap to reference
$200
$0
Exact matches
39%
52%
VINs at confidence 98.5 or 100
-
~843K more
Impact
Improves the accuracy of combined MSRP and options identification across window-sticker, listing, and OEM-feed decodes. Fields that may be impacted: combined_msrp, installed_options_msrp, installed_options_details, record_confidence, listing_confidence.
The trim fallback now uses Chrome specification data served from Redis, and the fuzzy matching fallback for trim is live.
The response records the trim confidence and the listing the trim was matched from.
The undetermined-trim handling was also extended to the GM group and to Nissan and INFINITI.
Impact
Resolves the trim on vehicles previously reported as undetermined, and shows what the match was based on. Fields that may be impacted: trim, trim_confidence.
When several specification records share a squish VIN, the correct one is now selected using make-specific fields instead of generic filters alone:
Wheelbase filter for Ford and Lincoln trucks (sticker wheelbase compared against the specification within a tolerance).
Model-code filter for nine makes, extended to the GM group and to Nissan/INFINITI.
Option-code filter for BMW.
The filters run in a set order, with the most discriminating first. Cases with several candidate versions that previously resolved to none now resolve to one.
Impact
Selects the correct specification record where the squish VIN alone is ambiguous. Fields that may be impacted: version, version_confidence, trim.
The high-precision keyword step now runs first, and the broader catalog match is used only as a fallback. This removes false-positive package code identifications.
The fallback validates each candidate code against the JATO manufacturer codes and picks the most specific match.
The fallback also scans equipment text from window sticker, OEM, and dealer sources, which recovers package codes that carry no package keyword.
Impact
Improves the accuracy of Equipment Group identification and removes one cause of incorrect version selection. Field that may be impacted: equipments_code.
Options with a negative price (credits, delete options, not-equipped adjustments) are now identified and applied as discounts. The JATO catalog does not supply them.
Discount extraction runs against the primary build source only, so the same credit is no longer applied more than once when it appears in Build Specs, OEM, and dealer data under different names.
Discount names are normalised before being matched against installed options, so the same amount is not deducted twice.
Impact
Decoded pricing now matches the window sticker on vehicles that carry credits. Fields that may be impacted: discounts, total_discount, combined_msrp, installed_options_msrp, record_confidence.
Previously, original_msrp was always taken from the earliest qualifying new listing. NeoVIN now evaluates three candidates and selects the best available value:
MSRP from the earliest qualifying new listing (the previous logic).
MSRP from the earliest qualifying new listing, excluding inferred MSRP.
The most frequent MSRP across all qualifying new listings, excluding inferred MSRP.
Impact
Improves the accuracy of the original MSRP reference and, with it, record confidence. Fields that may be impacted: original_msrp, mc_msrp, record_confidence.
Dealer window sticker as a source: dealer window sticker records are now used in the decode alongside build specs and OEM sources. Sources are fetched in parallel and merged in a configured priority order.
Build Specs gap reprocessing: vehicles that had build specs available but were not decoded from them are queued for re-decode, and the historical backlog has been cleared.
No-history VINs: VINs that return "No History Found" are now queued for a window sticker fetch instead of being dropped (limited to the makes the window sticker setup covers, with at most four attempts per VIN).
Impact
More vehicles are now decoded from a verified window sticker instead of from listing data. Fields that may be impacted: record_source, record_confidence.
Priced lines on the window sticker, dealer window sticker, or OEM inventory feed that match nothing in the spec provider's catalog are now installed as options using the source's own price. Examples are port- or dealer-installed accessories such as TOYOGUARD, spray-in bedliners, dealer wheels, and BMW M Performance parts.
DIO installation runs as the last correction pass. Install a line only if it closes or reduces the remaining gap to the reference price.
Rows installed this way carry a new boolean DIO flag in installed_option_details.
Closes MSRP gaps caused by dealer accessories that were previously invisible to the decode. Fields that may be impacted: installed_options_details (new dio flag), installed_options_msrp, combined_msrp, record_confidence.
24 standard-equipment fields that previously came from version-level defaults are now superseded at decode time with JATO equipment data for the exact build.
Equipment for the exact build is used first. The model-level aggregate is used only as a fallback.
The JATO ingestion job now reads the extended equipment table, so the superseding data is available to every decode.
Impact
Corrects specification fields where the build's actual equipment differs from the version default. For some fields, up to 50% of records are corrected. Consumers should expect value changes in the superseded specification fields.
ref_msrp: the window sticker's own line price for a matched option, captured at match time.
ref_adopted (boolean): set when the option's effective price was replaced by that sticker line price. This is the row-level counterpart of the |Opt Ref Price Used tag.
Example
A catalog-matched row after adoption. MSRP is still the catalog price ($2,945). The sticker for this build listed $1,545, and that amount was adopted into the totals.
Consumers can see both the catalog price and the sticker price for each option, and tell which one was used in the totals. Fields that may be impacted: installed_options_details (new ref_msrp, ref_adopted), installed_options_msrp, combined_msrp.
listing_confidence now carries extra tags that describe how the winning options were arrived at. A real value: mc_listings|used|dealer|franchise|2|fld:ftr|brwd:ftr+ulst|sibling:upg
Anyone parsing or comparing lc should either strip these tags or expect them.
Tag
Meaning
sibling:rlx / upg / exc / nrm
A sibling-resolver action survived: relaxation (unique minimal catalog reading), upgrade (attested token-superset swap), exclusivity resolution, or the BS+MSRP name-guard override.
fld:ftr / fld:sc
The winning option text came from the features field / from seller comments.
brwd:lst / oem / ftr / ulst
Options were borrowed into the winner. The token names the donor: marketplace listings, the OEM build record, other listings' features field, or a used-inventory listing (ulst, new).
clrmsrp
Priced paint resolved from exterior_color was added back through the catalog charge carrier.
psplit
A glued multi-price option string was split into per-option prices.
pkginc
An option priced inside a package on the same decode was zeroed through Includes-rule resolution.
paint:dedup
A duplicated generic paint charge was dropped in favor of the matching specific paint.
Impact
Makes every listing-based decode traceable to its evidence. Field affected: listing_confidence.
Predicts the current market price of a vehicle in Canada, using proprietary machine learning models trained on Canadian dealer retail sales data. Each prediction is VIN-specific.
Methodology
Trained on 360,000+ recently sold vehicle listings from 7,000+ Canadian dealerships.
Based on the last 3 months of sales data and retrained weekly.
Reflects current market conditions.
Features used for prediction
Vehicle specifications: year, make, model, trim, body type, transmission, drivetrain, fuel type, powertrain, colors, and features.
Location: city and province.
Condition: mileage, certification status, and seller category.
Market trends: recent pricing patterns for comparable vehicles.
Accuracy
95%+ R² on holdout datasets.
78%+ of predictions within a 10% margin.
96%+ of predictions within a 25% margin.
New fields
Field
Description
marketcheck_price
The vehicle's predicted market price in CAD, generated by the MarketCheck Price™ model.
msrp
The Manufacturer's Suggested Retail Price for the vehicle in CAD. Blank when an MSRP is not available.
Provides prescriptive vehicle–dealer matching for the Canadian market, with quantitative market fit scores and rankings.
Rankings are based on dealer inventory patterns, sales velocity, and year-make-model affinity, and are refreshed daily.
New fields
Field
Range
Description
vin_score
400–3000
The vehicle's market position, based on pricing, mileage, and market conditions. Higher scores indicate newer, lower-mileage vehicles at higher price points. Lower scores indicate older, higher-mileage vehicles at lower price points.
dealer_ymm_score
400–3000
The dealer's affinity for the vehicle's year, make, and model segment. Higher scores indicate the dealer typically handles vehicles in that segment with stronger market positioning.
vin_rank
1 to N
How well the vehicle matches the dealer, where 1 is the strongest predicted match. It considers inventory alignment, turnover rates, and current stock levels.
vin_score, dealer_ymm_score, and vin_rank are null when a vehicle could not be matched.