Concepts

Command Centre

The Command Centre is the main dashboard. It is the first screen you see after logging in. It shows a real-time summary of all connected ad platforms, customer segment health, and the latest agent recommendations.

The Command Centre has five sections:

• KPI strip: blended ROAS, total ad spend, conversions, and CPA across all channels
• Channel health table: every connected platform scored 0-100, with pacing status and trend
• Optimisation recommendations: specific budget, audience, creative, and bidding actions
• Agent activity feed: a chronological log of what each agent found and when
• Segment health panel: your customer segments with size and 7-day trajectory

AI Agents

Acera Labs runs ten specialist agents. Each agent has a defined scope: it observes a specific set of inputs, runs its analysis, and produces a defined output type. Agents do not make decisions without your approval.

The agents run on a scheduled cycle and also respond to events (for example, the Anomaly Agent fires whenever a metric breaks a statistical threshold, not just at a fixed time).

See Agents for the full list and what each one does.

Customer segments

Segments are groups of customers defined by behaviour and value. Acera Labs computes segments from your customer data and keeps them updated on each overnight cycle.

The default segments follow the RFM (Recency, Frequency, Monetary) model:

• Champions: high-value, recent, frequent buyers
• Loyal: consistent buyers with strong LTV
• Potential: show champion behaviour but have not yet reached that tier
• New: recent first purchases
• At Risk: previously strong customers showing declining engagement
• Hibernating: no activity in 90-plus days

You can view segment sizes, trends, and customer lists in the Segments section of the platform.

Customer Lifetime Value (CLV)

CLV is a predicted dollar value representing the total revenue a customer is expected to generate over their relationship with your brand. Acera Labs trains a CLV model on your transaction history.

The CLV module (under Models then CLV) shows:

• Predicted CLV distribution across your customer base
• Average predicted CLV
• Risk tiers (high, medium, low value)
• Health scores per customer

The Data Science Agent monitors CLV model accuracy and retrains when drift is detected.

Propensity scoring

Propensity models predict the probability that a customer will take a specific action: purchase, churn, upgrade, or refer. Each score is a number between 0 and 1.

The Propensity module (under Models then Propensity) shows scores by action type and by customer. Scores are refreshed on each overnight cycle.

Next Best Action (NBA)

Next Best Action is a recommendation for what to do with a specific customer or customer group right now. It combines segment membership, CLV, propensity scores, and channel availability to surface the highest-value action.

NBA recommendations appear in the Command Centre segment panel and in the Segments section.

Marketing Mix Model (MMM)

A Marketing Mix Model estimates the causal contribution of each marketing channel to revenue. It shows saturation curves (where spending more produces diminishing returns) and helps answer the question: where should the next dollar go?

The MMM module (under Marketing Mix Model) shows:

• Total media ROI weighted across channels
• Media contribution as a percentage of total revenue
• Per-channel response curves with saturation half-points
• Contribution waterfall showing baseline vs channel-attributed revenue
• Scenario planner for modelling budget reallocation

See BYOMM if you want to bring your own Marketing Mix Model.

Overnight cycle

The overnight cycle is the main processing window. It runs between 02:14 and 07:00 in your workspace timezone. During this window, the agents pull fresh data from all connected platforms, run their analysis, and produce the recommendations you see in the morning.

The cycle is not a single batch job. Each agent runs in sequence, and later agents read the outputs of earlier agents.

Approval gate

Nothing in Acera Labs activates without your explicit approval. Every recommendation requires a single-click approval before it is pushed to a connected ad platform. Every approval is logged with a timestamp, the approving user, and the full context of the decision. Approvals are reversible.