An AI bidding engine that learns from every auction. Deterministic inputs. Model-weighted outputs.

How it works

Three steps. Deterministic throughout.

Simulate a bid through the auction engine.

This is a client-side simulator, not a live production auction. Configure a campaign on the left. Every input runs through the same scoring logic the production decision engine uses: vehicle segment, live dwell, venue, daypart, and deal type compound into a single score. The score on the right is what your bid would hit in an equivalent auction.

Priority Waterfall

Five tiers. One decision engine.

Every ad request walks five priority tiers in order. The first tier with an eligible campaign wins. If nothing qualifies up top, the request falls to the next tier, and the next, until a creative is served.

Audience signals

Six signal dimensions, read at bid time from the OCPP wire.

Live dwell time, vehicle MSRP, and make/model are not directly observable in traditional DOOH inventory; Vistar Media, OUTFRONT Media, and Gas Station TV run modeled audience products derived from mobile-location panels and survey data. PredictBid reads all six signals from OCPP telemetry in real time, before the auction fires. Each stacks independently; when multiple trigger together, their CPM lifts compound.

AI modules

Three engines. Running continuously.

Deal tiers

Four ways to buy. One intelligence layer.

PredictBid works across all four deal types and automatically routes your bid to the tier with the highest expected win probability.

Questions & Answers

Questions we get from media buyers.

Short answers on how PredictBid fits into an existing programmatic stack, which signals drive the score, and how VideoEV compares to Gas Station TV, Vistar Media, and OUTFRONT Media.

What is PredictBid? PredictBid is VideoEV's AI continuous-optimization engine for commerce-media bidding at EV charging sessions. It reads six deterministic OCPP 2.0.1 signals (vehicle tier, income proxy, live dwell, make/model match, location premium, weather/daypart), runs them through model-weighted scoring, and returns a recommended CPM plus a win-probability estimate before the auction clears. Weights are recalibrated from observed auction outcomes. Inputs are deterministic, weights are AI-tuned, and the score breakdown is visible to the advertiser before spend.

How does PredictBid differ from Vistar Media or OUTFRONT Media? Vistar Media and OUTFRONT Media run sophisticated DOOH audience products built on polygon catchments, panel-augmented mobile-location modeling, and third-party demographic overlays. The audience signal there is modeled from phone movement plus panel data. PredictBid reads hardware-derived signals directly from the OCPP feed and resolves each EVCCID against the Vehicle Identity Graph before the auction fires. Different category of signal, different confidence shape per impression.

Is PredictBid cookieless? Does it use PII? Yes. PredictBid is cookieless by construction and does not carry PII in the bid stream. VideoEV targets the physics of the charging session (kW draw, voltage, hardware identifier), not the person. No mobile ad ID, no email, no device fingerprint. The platform is IAB TCF v2.2 and GPP aligned, MSPA registered on all three VideoEV domains.

What signals does PredictBid use at bid time? Six contextual signals: vehicle tier, income proxy, live dwell time, make/model match, location premium, and weather/daypart. All six arrive deterministically from OCPP 2.0.1 telemetry and are resolved through the Vehicle Identity Graph before the auction starts. This is Deterministic Physics-Based Targeting, not probabilistic mobile inference.

What is the average yield lift from PredictBid? Aggregate yield lift is under measurement during pilot. PredictBid combines three mechanisms: dynamic CPM floors that adjust to live demand, audience-signal compounding when multiple premium signals fire together, and a fill-rate maximizer that routes bid weight toward stations with available inventory. Signal weights recalibrate weekly from pilot auction outcomes. Specific lift figures will be published once pilot campaign volume supports a stable benchmark.

Is PredictBid actually AI, or just a rules engine? Both. The production path is a deterministic scoring function with transparent multipliers per campaign. Advertisers can see and approve every weight. The AI layer is a weekly recalibration pass that tunes those weights from observed auction outcomes (win rate, completion rate, conversion proxy). That pattern matches the Skai 2026 State of Retail Media finding that 85% of brands are comfortable with AI-driven campaign optimization when inputs are visible. We publish the score breakdown rather than hiding it behind a black-box model.

How does VideoEV fit into the commerce media landscape? VideoEV is a non-retail commerce media network: a captive branded screen at the point of a commerce event (the charging session), with retailer-style 1P signal (OCPP telemetry) and clean-room export into Amazon Marketing Cloud. Commerce media is 19.3% of U.S. digital ad spend in 2025 per EMARKETER; non-retail commerce media compounds at 35.1% CAGR. VideoEV sits alongside Instacart, Uber, PayPal Ads, Chase Media Solutions, Grocery TV, and Cooler Screens in that category, not alongside Lamar or OUTFRONT.

Is PredictBid compatible with OpenRTB 2.6? Yes. PredictBid exposes an IAB OpenRTB 2.6 compatible endpoint for programmatic DOOH bid requests and serves VAST 4.3 responses with impression tracking, brand-safety categories, and universal ad IDs. VideoEV also supports OpenDirect 2.1 for direct booking and is AAMP discoverable by autonomous buyer agents.

How long is the average charging dwell on the VideoEV network? Roughly 28 minutes of captive attention per session, a physics-derived estimate for a typical DC fast charge from ~20% to ~80% State of Charge. That is about 9× longer than the ~3-minute distracted dwell at a gas pump. Aggregate video completion rate is read from VAST quartile events and will be published once campaign volume supports a stable figure.

What CPM range does PredictBid target? $14 to $24, depending on deal tier, sector, and audience match. Open auction floors start near $14, preferred deals sit around $20, and programmatic direct runs up to $24 for high-MSRP matches. Individual campaign CPMs are never exposed in public rate cards, only aggregated ranges.

Which EV charging networks does VideoEV integrate with? VideoEV reads OCPP 2.0.1 telemetry natively. Target CSMS platforms include Driivz, AMPECO, ChargeLab, and EV Connect; compatible charger hardware spans ChargePoint, Blink, EVgo, Electrify America, ABB, and BP Pulse. The integration is read-only and adds no new hardware to the site. Commercial partnerships are not yet contracted.

How big is the U.S. EV charging ad opportunity? 141 million U.S. DC fast-charging sessions are projected for 2025 by Paren, growing more than 30% year-over-year. The VideoEV Vehicle Media Network is building toward this pipeline through OCPP-compatible CSMS platforms including Driivz, AMPECO, ChargeLab, and EV Connect. The segment skews to drivers with an average vehicle MSRP above $65,000, inferred from the resolved make/model distribution, a demographic difficult to reach at comparable scale through legacy DOOH.