Smart Forecaster is Bear IQ's predictive registration model. It generates weekly scale goals and a forecast range based on your event's historical performance, configured weights, and anonymized platform-wide registration trends.
Weekly Scale Goals
Scale goals represent the target number of registrations for each week leading up to your event. They are calculated using two primary inputs:
-
Historical trajectory: Registration patterns from prior editions of this event within your Project, weighted according to your configured settings.
-
Market impact quotient: A slight adjustment factor based on how other events on the Bear IQ platform are currently performing. This helps account for broader market timing trends that may influence your registration trajectory.
Goal Locking
Weekly scale goals lock for 7 days. This prevents the target from shifting with every data refresh, which proved confusing and counterproductive in real-time mode. The "weeks out" interval is configurable down to the day in Settings within Bear IQ.
Time Horizon
The default forecast time series is 20 weeks. Forecast accuracy improves with more data, so 20 weeks provides a reliable baseline. For events with registration windows exceeding 30 weeks, the time horizon can be extended upon request. Contact your Customer Success Manager to adjust.
Forecast Range
The forecast range (e.g., 97k–100k) represents a 95% confidence interval. It reflects the model's prediction of final registration count based on the information available at that point in time:
-
Total registrants to date
-
Time remaining before the event
-
Historical and live market performance
-
Campaign momentum to date
Model Assumptions
The model assumes week-to-week campaign consistency. If something changes mid-cycle (e.g., a pricing increase that slows registrations), the model requires approximately one week of data to detect the shift. It will then adjust the momentum coefficient and revise the forecast accordingly. The model identifies the slowdown but not the cause.
Range Behavior Over Time
In most scenarios, the range narrows as you approach the event date. However, inconsistent performance can keep the range wide. Examples include outsized gains, bulk registration drops, or severe underperformance relative to traditional trajectories.
Momentum Coefficient
The momentum coefficient measures whether registrations are accelerating or decelerating. It serves two purposes:
-
Pacing ahead: Scale goals push to maintain or grow that lead. For example, if you are ahead of your historical pace, the model will forecast scale goals that allow you to continue that pace vs. slow down to just hit the goal. This ensures Bear IQ is reflecting an accurate trajectory.
-
Pacing behind: The model forecasts the catch-up pace required to return to the stated goal trajectory. For example, if your campaign is tracking behind the outlined pace to goal, the platform will scale your future weeks' goals to provide the best fit to “catch up” and thus give your team the ability to hit your pre-event registrant target.
Platform Learning
Smart Forecaster learns from recent events occurring across the Bear IQ platform. This data is always anonymized and aggregated. No individual event data is identifiable - even internally at Bear Analytics. Platform-wide trends influence the market impact quotient, helping the model account for timing shifts and broader registration patterns.
Projects
Understanding these terms is essential for interpreting how Smart Forecaster uses your historical data. A “Project” is the rollup of events in Bear IQ used for year-over-year analysis. For example, you might group the 2022–2026 editions of the same show into one Project. This allows you to compare performance across years, unify prospect data, and select which historical years to weight in your forecast.
Weight Configuration
You can configure how much each historical event influences the forecast. These settings are located in Settings > Forecast settings within Bear IQ.
Weights are assigned on a 0–100 point scale. You can weight toward recent events, events held in a specific city, or distribute weights individually across each historical event in your Project.
The model heavily prioritizes your historical performance based on these configured weights. Platform-wide (market) data provides a secondary influence on the momentum coefficient.