Front

Front exposes several object types through its API: conversations, messages, contacts, teammates, inboxes, tags, teams, and analytics reports. This integration focuses on the conversation object, which is the fundamental unit of customer interaction and the most direct source of performance indicators for support and customer success teams. Messages and contacts were excluded: messages are too granular and carry high gaming risk, contacts are a segmentation object rather than a performance object. Five KPIs were retained, selected against three criteria: ability to attribute to an individual teammate, resistance to gaming, and balance between leading and lagging indicators.

Response discipline and SLA compliance

Average First Response Time measures how quickly a teammate sends the first reply after a conversation is created. SLA Breach Count records how many conversations exceeded a committed response threshold. These two indicators address the same underlying management concern — response discipline — from different angles. Average First Response Time provides continuous measurement and enables trend analysis; SLA Breach Count provides event-level measurement and quantifies the operational cost of missed commitments.

The two indicators are not redundant. A teammate can maintain a low average first response time overall while still accumulating SLA breaches on specific conversation types, inboxes, or time windows. Reading them together exposes structural weaknesses that the average alone obscures. Average First Response Time carries a moderate gaming risk — a brief placeholder reply stops the clock without resolving the customer's issue. SLA Breach Count is substantially harder to game, since the breach event is recorded automatically by the platform against defined thresholds. The combination offsets the weakness of each individual indicator.

Throughput and quality

Conversations Handled counts the number of conversations a teammate archived or resolved in a given period. Average CSAT Score measures the average satisfaction rating customers gave on conversations with that teammate. These two indicators form the core accountability pair for support performance: throughput without quality is activity; quality without throughput is insufficient capacity.

The tension between them is productive and intentional. A high Conversations Handled count combined with a declining Average CSAT Score typically signals that the teammate is processing volume at the expense of interaction quality. The inverse pattern — high satisfaction but low throughput — may indicate over-investment of time per conversation, whether from excessive thoroughness or from avoidance of the volume queue. Neither pattern is immediately alarming in isolation; both become managerially relevant when they diverge from baseline. Average CSAT Score has a small-sample sensitivity that limits its reliability for teammates handling fewer than ten rated conversations per period; this boundary should be respected when drawing conclusions.

Efficiency: Average Handle Time as a capacity signal

Average Handle Time measures the end-to-end duration of conversations managed by a teammate. It differs from Average First Response Time in that it covers the full lifecycle of the conversation rather than the initial reaction. A teammate with a long Average Handle Time relative to peers may be managing more complex cases, may be facing collaboration or tool inefficiencies, or may be over-communicating on straightforward issues. The metric does not distinguish between these causes, which is why it functions as a diagnostic prompt rather than a verdict.

Average Handle Time carries a meaningful gaming risk: a teammate incentivized to reduce handle time may close conversations prematurely or send fewer replies to artificially compress the duration. This risk is mitigated by reading it alongside Average CSAT Score, which deteriorates when conversations are closed before customers are satisfied. The two indicators constrain each other: improving Average Handle Time at the expense of Average CSAT Score is visible; improving both simultaneously is a genuine efficiency gain.

Scope and limits of the integration

Front records response times, conversation counts, and CSAT ratings, but does not measure the quality of reasoning or communication within a conversation. A conversation marked archived may have resolved the customer's issue completely or only partially, with the customer choosing not to reopen rather than having no remaining need. The API does not surface this distinction. Metrics derived from analytics report endpoints are only available for inboxes where the corresponding features — CSAT, SLA, ticketing — are enabled; teams that have not activated these features will see null values rather than zero, and the integration handles this distinction gracefully.

Attribution of conversations to a single assignee may undercount collaborative work: when a conversation is reassigned mid-resolution, only the final assignee carries the outcome in this model. Teams that rely heavily on handoffs or shared resolution practices will find the per-teammate figures less precise than teams where each teammate owns a conversation end to end. As with any integration of this type, the reliability of these KPIs depends directly on the discipline with which the team uses Front: accurate assignee management, consistent inbox routing, and active CSAT collection are preconditions for meaningful measurement.