Box-Office Prediction

Opening weekend numbers, with an honest confidence band.

Project domestic and international opening windows for any wide release. Built on a model that learned from twenty years of weekend grosses, not a hand-tuned spreadsheet.

Start free with Box-Office PredictionSee pricing

How it works

The exact thing your team is trying to call.

Opening weekend gross by territory, plus full first-month decay curve, with explicit confidence intervals at the 80 and 95 percent levels.

  • Territory-level forecasts for the top 35 markets
  • Decay curve modeling, not just opening weekend
  • Calendar effects for holidays, sports, and tentpole conflicts
  • Trailer-impression and pre-sales ingestion
  • Marketing-spend elasticity simulation

Signal sources

What feeds the box-office prediction model.

  • 01Box Office Mojo and Comscore historical grosses
  • 02YouTube trailer view and sentiment
  • 03Fandango and AMC pre-sale signal
  • 04X premiere-week chatter
  • 05News RSS coverage and review embargo data
Sample entities tracked:Untitled Marvel Q3Indie Drama SpringAction Sequel 2027

Use cases

What teams actually do with Box-Office Prediction.

  • Greenlight reviews for box-office prediction

    Run scoring on a portfolio of candidates and surface the ones whose box-office prediction signals diverge most from their internal narrative.

  • Continuous monitoring

    Subscribe via webhook and receive a payload only when the score crosses a threshold you defined. No dashboards required.

  • Cross-vertical analysis

    Pull the same entity into other verticals to see whether the signals agree or disagree. Disagreement is often where the alpha is.

  • Backtesting before commitment

    Score the last 24 months of decisions against the model and compare hit rates. The calibration_bucket field tells you how much to trust the model in the bands you care about.

API

Calibrated, explainable, ready for production.

Every prediction ships with a confidence interval, the top feature drivers explaining the score, and the five nearest historical analogs the model leaned on.

Available via REST and a typed TypeScript SDK.

See full API reference
GET /v1/predictions/box-office/{entityId}
{
  "entityId": "film_kj91",
  "openingWeekendUSD": 78400000,
  "ci95": [
    62100000,
    96800000
  ],
  "territoryBreakdown": {
    "domestic": 0.46,
    "international": 0.54
  },
  "drivers": [
    {
      "feature": "trailer_view_z",
      "contribution": 0.14
    },
    {
      "feature": "presale_velocity",
      "contribution": 0.11
    },
    {
      "feature": "franchise_fatigue_index",
      "contribution": -0.06
    },
    {
      "feature": "competitor_window_pressure",
      "contribution": -0.04
    },
    {
      "feature": "marketing_spend_z",
      "contribution": 0.03
    }
  ],
  "analogs": [
    "tt9876123",
    "tt8443209",
    "tt7012891",
    "tt6611123",
    "tt5500098"
  ]
}

Compared to

How Box-Office Prediction differs from what is already out there.

VendorHow SignalGrid is different
Gower StreetOpen public signal model with explicit confidence bands per territory.
Boxoffice ProContinuous-update model, not weekly editorial forecasts.
ComscoreForward predictions and decay curves, not only reporting on what already happened.

Try Box-Office Prediction on your own entities.

Start free during the public beta. We will pre-select box-office prediction for you on signup.

Start freeTalk to us

Other verticals