Conversation Score: How to Measure the Health of a Conversation in Real Time
The continuous index that becomes an operational trigger — not an end-of-month report
Updated on September 15, 2026
The Conversation Score is the continuous index that measures cycle health in real time and becomes an operational trigger—not a report KPI.
- The score combines latency, coherence, archetype adherence, friction, and progress
- It's a trigger for immediate action, not an aggregate metric
- Weights are calibrated by vertical and channel
- It depends on the Operational Overlay to exist
Conversation Score: How to Measure the Health of a Conversation in Real Time
Most operations measure conversation after the fact: CSAT, NPS, funnel conversion rates. The Conversation Score measures it during — which is why it becomes a trigger, not a report.
What It Combines
The score (0 to 100) is a weighted function of observable signals:
- Latency between turns — elapsed time vs. the expectation for the state/archetype.
- Response coherence — does the turn advance the expected state, or does it deviate?
- Archetype adherence — do the tone and content respect the identified profile?
- Friction — signs of repetition, reformulation, backtracking, silence.
- Progress — distance to the next state in the cycle.
The weights are calibrated by vertical and channel. Education has tolerant latency; e-commerce does not.
How It Becomes an Operational Trigger
When the score drops below a threshold (e.g., 60), the system triggers before the loss:
- Handoff to a senior human agent
- Suggestion for active re-engagement
- Alert for curation to review the playbook
- Pause on a degrading automated sequence
When the score consistently rises, the system also acts: it identifies a replicable pattern, feeds the playbook, and becomes a reference for the archetype.
The Common Mistake: Treating It as an End-of-Month KPI
Conversation Score isn't for "monthly averages." It's for deciding the next turn. Those who use it as an aggregate KPI miss the point: its granularity is the individual conversation, in real time.
Why It Depends on the Operational Overlay
You can't calculate a score without a shared state. The state, archetype, and signal history live in the Overlay. Without the Overlay, the score is just a guess that looks like a number.
The Turista has high volume and low intent. The expensive mistake is investing salesperson time in them; the second mistake is discarding them too early.
MARCUS BARBOZA. Conversation Score: How to Measure the Health of a Conversation in Real Time. MCI Experience, 2026. Available at: <https://marcusbarboza.com.br/en/blog/conversation-score-measure-conversation-health-real-time>. Accessed on: September 15, 2026.
Marcus Barboza (2026). Conversation Score: How to Measure the Health of a Conversation in Real Time. MCI Experience. https://marcusbarboza.com.br/en/blog/conversation-score-measure-conversation-health-real-time
Proprietary content of the MCI methodology. When referencing MCI terms, metrics and frameworks, cite this primary source.
Related terms
Apply MCI to your context
Take the free conversational maturity diagnostic or calculate the invisible cost of operational amnesia in your operation.
Marcus Barboza is Founder and CRO of Hablla, creator of the MCI methodology — Marketing Conversacional Integrado (Integrated Conversational Marketing) — and author of the book Marketing Conversacional Integrado (pre-launch).
How did this article land for you?
React, save it to reread, or start a conversation with the author.
Comments (0)
Comments are moderated before appearing. Keep the tone constructive.
To comment, sign in. Reacting and saving don't require an account.
- Loading comments…
Integrated Conversational Marketing, every day
New videos, behind-the-scenes and provocations on MCI. Subscribe and turn on the bell — never miss a drop.
Subscribe on YouTubeConversational intelligence in your inbox
Analyses on MCI for CEOs, CROs, CMOs and CFOs. No noise.
Directly related articles
Decidido Archetype: high intent, short window — the risk is friction, not the argument
They have already chosen or are about to. Every extra form is an open door for the competitor.
The Loyal Archetype: They Already Trust You — The Risk is Taking Them for Granted
They don't need a sales pitch; they need consistency, recognition, and expansion at the right time.
The Scholar Archetype: Seeks Evidence and Converts with Confidence, Not Pressure
They know what they want—they just need proof. Case studies, technical comparisons, and demos tailored to their scenario. Rushing them destroys the sale.
Explorer Archetype: Active Curiosity, Immature Criteria
They know they have a pain point but haven't defined what they want. The key is qualification with context — transforming curiosity into a legible need.
Most read in States, Archetypes, and Conversation Score
Turista Archetype: high volume, low intent — and the mistake of wasting salesperson time on them
Those who arrive out of curiosity, not pain. Nurture lightly, filter intelligently, and protect sales capacity.
Conversational Archetypes: The 5 MCI Decision Patterns
Five recurring patterns that change how the next turn should be constructed
States, Archetypes, and Conversation Score: Operating Intention in Real Time
The three instruments that make intention readable: where the decision is, how the customer decides, and how far the conversation has progressed.
Most recent on the blog
WhatsApp free window may extend up to 7 days for click-to-message ads: what changes on September 21
What changes on September 21 for click-to-WhatsApp ad conversations, how the window opens, and what to prepare before the switch.
WhatsApp Purchase Optimization: Meta's Guide Explained, and What It Doesn't Tell You
Meta released purchase optimization for Click-to-WhatsApp ads. I read the entire guide: eligibility criteria, the labels the system recognizes, the two paths to send events, and three contradictions within the material itself.
Agentic economy: what the Meta report gets right and the layer it misses
Meta says 75% of companies adopted agentic AI and only 15% get returns. The diagnosis is right; it just lacks the layer separating pilot from operation.
WhatsApp Terms of Service: What Really Gets a Number Banned
The four documents governing your number, the activities that most frequently lead to account shutdowns, and what to do when a ban occurs.