For Everyone · Updated Oct 20, 2026
Understanding upload and value thresholds
Three settings control what Scowtt sends to your ad platforms: which records are eligible, which leads are sent at all, and which leads carry a value. Together they balance how much the platform learns from against how clean that signal is.
Upload criteria: what can be sent
Upload criteria define which kinds of records are eligible to be sent, such as leads, scheduled consultations or opportunities. Sending earlier in the funnel, for example at the lead stage, gives the platform more signal and gets it there sooner.
- Too restrictive: Slow learning and unstable bidding.
- Too loose: Noisy signals and weaker efficiency.
Upload threshold: which leads are sent
The upload threshold is the minimum score a lead needs before Scowtt sends it to the platform at all. It controls how much data the platform sees, which trades learning speed against precision.
- Set too high: Low volume, so the bidder struggles to learn and leans toward only the most obvious leads.
- Set too low: Too much noise, which dilutes the model's influence.
In practice, the upload threshold is often set low, so almost every lead is sent except the very lowest quality.
Value threshold: which leads carry value
Of the leads that are sent, the value threshold decides which ones get a $0 value and which get a positive predicted value. A $0 value acts as a negative signal: it teaches the platform what not to chase, not just what to chase.
| Lead score | What's sent |
|---|---|
| Below the value threshold | Sent with $0 value |
| Above the value threshold | Sent with its predicted value |
How values are spread
Ad platforms optimize on the relative value of conversions, not exact dollar amounts. If every lead looks roughly the same, the platform can't tell where to focus. If the gap between weak and strong leads is clear, bidding leans toward the people most likely to become valuable customers. Scowtt can scale the values it sends so the platform gives them the right weight.
Platforms also compare total conversion value, so the average value Scowtt sends needs to stay in a sensible range. Too low, and the platform can underbid and miss good leads. Too high, and it can overspend. Your Scowtt team checks this after any model or threshold change.
Adjusting them together
Upload and value thresholds work as a pair, so they're adjusted together rather than one at a time. Early on, or with smaller budgets, they usually favor volume so the platform has enough to learn from. As campaigns mature and budgets grow, they shift toward precision.