Eight common labels. Calibrated to the evidence. Every verdict is a recommendation, not a command.
Every Carlarity report ends with a one-line verdict. The Deal Verdict Engine™ collapses your deal to one of four canonical outcomes (Buy, Negotiate, Walk, Insufficient Evidence), surfaced as one of the five labels below — five, and only five. The buying mode changes the advice under the label, never the word.
The asking price sits above the modeled midpoint, and the data supports specific negotiation targets. Every verdict reads this way — label, confidence, and range travel together.
Illustrative — not a live valuation
Five labels, and only five. At LOW confidence a Buy or Negotiate reads VERIFY FIRST; at INSUFFICIENT confidence the report says INSUFFICIENT EVIDENCE. Report Guard™ enforces the modulation.
This deal appears favorable across pricing, structure, and equity outlook, and data confidence supports the call. Action: proceed with the standard verification checklist (mechanical inspection, title check, financing pre-approval). Inside a return window, BUY means keep the car.
When you'll see this: HIGH or MEDIUM confidence, deal scores well across the board.
Change what you can: price, fees, add-ons or financing. The line under the word names what applies to your deal. On a fixed-price platform (Carvana, CarMax, Vroom, Tesla) the price is not negotiable, so it names the add-ons and the financing, and Carlarity never offers a price target there. Older reports may show NEGOTIATE, which is now called ADJUST.
When you'll see this: HIGH or MEDIUM confidence, and the problems are ones a buyer can fix.
The deal would read BUY or ADJUST, but pricing confidence is LOW. The direction is real; the precision is not. Action: verify with 3–5 comparable listings and a mechanical inspection before acting on it.
When you'll see this: LOW confidence on a deal that would otherwise be BUY or ADJUST.
The deal does not meet Carlarity's threshold and nothing the buyer controls fixes it — usually the price itself. Walking away is the recommended action, and confidence supports it. On a fixed-price platform WALK means compare other listings; inside a return window it means return the car.
When you'll see this: HIGH, MEDIUM or LOW confidence, deal scores poorly enough that no controllable change bridges the gap.
Carlarity cannot produce a confident recommendation: pricing confidence is INSUFFICIENT, or the engine could not reach a directional verdict. Action: do not rely on this report for the price call; verify independently or consult an independent expert before proceeding.
When you'll see this: INSUFFICIENT confidence — Carlarity recuses itself rather than guess.
The platform sets the price, so ADJUST never means haggling and the report never offers a price target. It means change what you control: decline the add-ons, bring outside financing, fix the structure of the deal.
When you'll see this: the online fixed price form of ADJUST. Low and insufficient confidence still modulate it the way the table below describes.
When the price itself is the problem and nothing you control fixes it, the verdict is WALK: compare other listings — the lever you have is a different car, not a counter-offer.
When you'll see this: the online fixed price form of WALK. Low and insufficient confidence still modulate it the way the table below describes.
You have already bought the car and can still give it back. BUY means keep it: the evidence supports the purchase you made.
When you'll see this: the inside a return window form of BUY. Low and insufficient confidence still modulate it the way the table below describes.
WALK means return the car: the evidence points away from keeping it, and the window is the remedy.
When you'll see this: the inside a return window form of WALK. Low and insufficient confidence still modulate it the way the table below describes.
The same underlying deal can produce different verdict labels depending on data confidence. Signal Integrity Layer™ caps confidence when evidence is thin — the verdict softens to match.
| Underlying signal | HIGH / MEDIUM tier | LOW / INSUFFICIENT tier |
|---|---|---|
| Favorable price + structure | BUY | VERIFY FIRST |
| Mixed signals, negotiation supported | ADJUST | VERIFY FIRST |
| Unfavorable across dimensions | WALK | WALK |
| Insufficient data | (rare at HIGH/MEDIUM) | INSUFFICIENT EVIDENCE |
Three evidence states you may see beneath a verdict — and what each one says about the data behind it.
The Open-Web Corroboration Engine™ fired on this report — your primary pricing providers couldn't carry a confident answer alone, so we widened the evidence surface with structured open-web listings from our whitelisted aggregators. The chip tells you (a) it fired, (b) how many open-web comps contributed, and (c) that the confidence ceiling for this path is capped at MEDIUM.
Below the verdict, the report lists every open-web comp used (price, mileage, trim, source domain, listing date) so you can verify the underlying evidence in one click. A confident HIGH verdict always requires primary provider agreement; the open web alone cannot earn it.
The engine found between one and two open-web data points — meaningful signal, but not enough to elevate confidence under our methodology table. Common on low-volume models where structured aggregator coverage is thin. The amber treatment is intentional: this is partial signal, not verdict-grade evidence.
The surfaced comp(s) are shown for transparency so you can verify the source yourself. Your verdict and confidence tier remain at whatever the primary providers reported — we don't pretend partial data is full data.
For some vehicle classes — luxury, rare, very-low-volume — structured listing data isn't reliably available on the open web. We surface this state explicitly rather than failing silently. The engine ran, found nothing usable, and is telling you so.
Your verdict comes entirely from the primary pricing providers in this case. The supply-limited disclosure is part of the report's trust-honest framing: we'd rather you know what we tried than imply we have more data than we do.
A confident-sounding wrong answer is worse than an uncertain right answer. When pricing data is thin, our verdicts get softer — not because the underlying analysis is weaker, but because we'd rather admit limited data than pretend our model is stronger than it is.
This is the same principle that governs the Methodology page: verdicts are calibrated to the evidence behind them. Read the verdict alongside the confidence label and the data provenance — they're three signals that travel together.
Carlarity verdicts are decision support, not decision authority. Override the verdict when:
Use Carlarity to surface what to investigate. The final call is yours.
Full methodology: the 12 systems, 4 confidence tiers, and the philosophy of trust-honest grading.
Read the methodology →
What we fully support, what needs manual entry, and what we don't support yet.
See coverage →
Before you sign
One report before you sign — pricing, structure, and a recommendation that tells you how much to trust it.
We use our own analytics to improve Carlarity. With your permission we also let Google, Meta, TikTok and Reddit set cookies to measure whether our ads work. Decline and nothing is set — your reports work either way. Change your mind any time under Privacy choices.
