The Hypothesis

Can a four-phase autonomous loop — Signal, Act, Measure, Evolve — replace the traditional six-stage marketing cycle?

The Concept

Traditional marketing follows a painful linear cycle: design → approval → send → measure → adjust → design. Each stage introduces days or weeks of latency. A campaign might take a month to complete one loop. This experiment collapses that into four autonomous phases that run continuously: Signal (read intent, persona context, and past learnings), Act (deploy content across channels), Measure (stream real-time outcome data — not vanity metrics, actual business outcomes), Evolve (adapt strategy and restart). The loop turns continuously, each cycle informed by real-world outcomes of the last.

Feedback intelligence loop

The hypothesis

Can a four-phase autonomous loop — Signal, Act, Measure, Evolve — replace the traditional six-stage marketing cycle?

The concept

Traditional marketing follows a painful linear cycle: design → approval → send → measure → adjust → design. Each stage introduces days or weeks of latency. A campaign might take a month to complete one loop. This experiment collapses that into four autonomous phases that run continuously: Signal (read intent, persona context, and past learnings), Act (deploy content across channels), Measure (stream real-time outcome data — not vanity metrics, actual business outcomes), Evolve (adapt strategy and restart). The loop turns continuously, each cycle informed by real-world outcomes of the last.

How it works

1. Signal — synthesise outcome goal, persona data, past learnings, market signals
2. Act — deploy landing pages, emails, ads, lifecycle flows across channels
3. Measure — stream real-time outcomes: purchases, bookings, activations — not just opens and clicks
4. Evolve — adjust creative, shift variant weighting, refine personas, generate new tests
5. Loop — restart at Signal with sharper intelligence

The marketing team sets outcome goals and boundaries. The system operates the loop.

What it explores

• Can an autonomous loop meaningfully reduce the campaign cycle from weeks to hours?
• Does measuring business outcomes (revenue, bookings) instead of vanity metrics (opens, clicks) produce better optimisation decisions?
• How much human oversight does the loop need before trust is established?
• Can the Evolve phase reliably improve strategy across channels simultaneously?
• What’s the right cadence for human review — daily, weekly, or only on flags?

What we found

• Full Signal → Act → Measure → Evolve cycles completed in 4–8 hours vs. 3–5 weeks manually
• Campaign performance improved 35% when optimising against business outcomes vs. vanity metrics
• Landing pages and email variants co-evolved simultaneously — changes in one channel informed the other
• Human review settled at weekly cadence for most teams, with real-time alerts for anomalies

Learnings

• The Signal phase is where quality compounds — the richer the context from past cycles, the better the starting point for the next.

• Cross-channel co-evolution needs guardrails — without them, the system over-optimises one channel at the expense of others.

• Teams need a “glass pane” into the loop — when they can’t see what the system is doing, trust erodes regardless of results.

• The marketer’s role shifts from operating to steering — the skill becomes setting goals and boundaries, not executing tasks.

Where it goes next

This is the conceptual backbone of Flywheel. The four-phase loop — Signal, Act, Measure, Evolve — is now the product’s core architecture. The open question: how do we make the “glass pane” good enough that teams trust the loop to run without daily check-ins?