Skip to main content
Trackly SMS automatically selects the best-performing creative for each send, maximizing clicks and revenue.

​
How It Works

Instead of manually choosing which message to send, automated selection:
  1. Tracks performance for every automated creative (sends, clicks, conversions, offer payout)
  2. Scores each creative by expected revenue per send
  3. Samples a winner for each send, favoring proven performers while still testing others
  4. Updates the statistics as new results come in

​
Enabling Automated Selection

​
Campaign Setup

  1. Create a campaign in Automated Mode
  2. Configure the automated percentage:
  • 80% of contacts get ML-selected creatives
  • 20% get the default creative (control group)

​
Available Creatives

The ML system chooses from creatives marked as Automated type:
  1. Go to Creatives
  2. Mark creatives as “Automated” type
  3. These become candidates for ML selection

​
The Selection Model

Trackly SMS uses Thompson Sampling with Beta distribution posteriors as the selection algorithm.

​
How Thompson Sampling Works

Thompson Sampling is a Bayesian algorithm that balances exploration (trying less-tested creatives) with exploitation (favoring proven winners). For each send:
  1. Sample a score from each creative’s Beta distribution posterior
  2. Select the creative with the highest sampled score
  3. After delivery, update the posterior with observed results

​
Two Selection Modes

​
Metric Optimized

The algorithm optimizes Revenue Per Send (RPS):
For CTO (cost-to-operator) offers, conversion rate is omitted:
Offers with a payout below 20arescoredasiftheypaid20 are scored as if they paid 120, so low-payout, high-volume offers aren’t penalized in scoring against higher-payout offers.

​
Pre-Computed Parameters

Selection statistics are recomputed for your account every six hours, so selection adds no delay at send time.

​
What Drives Selection

Automated selection uses Thompson Sampling, which tracks aggregate performance per creative and per offer — not individual contact history. For each creative, the system tracks: Per-contact scoring on features like send history, time of day, or custom-field values is not yet available. If a selection method offering it is configured in SMS Settings > Automation, traffic routed to that method is selected at random.

​
Scoring

Selection always scores creatives on Revenue Per Send (RPS), as described above. On schedules flagged CTO (cost-to-operator), the conversion-rate term is dropped and creatives are scored on click-through rate × offer value.

​
Configuration

Optimization settings are configured at the account level under SMS Settings > Automation. Campaigns only control automated_percent.

​
Cooldown Settings

Cooldown is enforced per-contact, per-creative and is configured at the account level under SMS Settings > Automation, in the Cooldown Rules section. When a contact receives a specific creative, that creative becomes ineligible for that contact until the cooldown period expires.

​
Selection Methods

The system supports several selection methods, configured at the account level: Multiple methods can be configured simultaneously with traffic_pct weights, routing different percentages of traffic to different algorithms. Thompson sampling naturally balances exploration and exploitation: creatives with uncertain performance get tested more frequently, while proven performers are selected more often. This replaces the need for a fixed exploration rate.

​
Performance Tracking

​
A/B vs Automated

Compare automated selection against manual:

​
When to Use ML Selection

  • Large contact list (5,000+)
  • Multiple creatives to choose from (3+)
  • Enough historical data (10,000+ sends)
  • Measurable conversion goals

​
Cold Start / Learning Pool

Creatives are placed in a learning pool until they reach your account’s configured minimum-sends threshold (default 300 sends, configurable from 50 to 10,000). A configurable percentage of traffic (default 10%) is allocated to learning creatives, which are selected randomly to gather baseline data. The remainder uses Thompson Sampling scoring. Once a creative crosses that threshold, it exits the learning pool and competes on its sampled RPS score.

​
Graduation Warmup Ramp

The learning percentage you configure applies only once your account holds a real competitive field. Until 20 creatives have graduated past the min sends threshold, the platform automatically raises the effective learning share:
Without this ramp, the first creative or two to graduate would immediately capture ~90% of all sends — your contacts would receive the same one or two messages repeatedly while hundreds of other creatives starved. Each additional graduate shifts 5% of traffic back to Thompson exploitation, so the transition from warmup to full optimization is gradual and self-accelerating. Accounts with 20+ graduated creatives are unaffected — the configured learning percentage applies exactly. The ramp also applies when learning percentage is set to 0: with a small graduated field, exploration continues regardless, because without it no additional creative could ever graduate.

​
New Accounts

Without historical data:
  1. All creatives start in the learning pool
  2. Traffic is split randomly until enough data accumulates
  3. Once a creative passes your configured minimum-sends threshold, Thompson Sampling takes over for it
  4. Selection statistics refresh every six hours as more data comes in

​
New Creatives

New creatives automatically enter the learning pool:
  • They receive guaranteed exposure from the learning allocation
  • Performance data accumulates until the minimum-sends threshold is reached
  • Strong performers graduate to Thompson scoring and rise to the top

​
Monitoring

​
Dashboard Indicators

Watch for:
  • Model health: Is the model performing well?
  • Creative diversity: Are all creatives getting selected?
  • Performance trends: Is click rate improving?

​
Alerts

Set up alerts for:
  • Model degradation
  • Single creative dominating (may indicate overfitting)
  • Performance drops

​
Best Practices

Keep 5-10 active automated creatives. Too few limits ML effectiveness.
Add new creatives periodically. Stale content loses effectiveness.
Always maintain some manual sends to measure ML lift.
A creative stays in the learning pool until it reaches your minimum-sends threshold. Judge its performance after it graduates, not before.

​
Next Steps

Campaigns

Set up automated campaigns

Reporting

Analyze ML performance