11 metrics every TMC should know

Why it matters

Volume is the one number most travel teams already have, usually as a raw email count. On its own it explains nothing: 16,000 emails a month is a fact, not a decision.

It becomes useful when it is split three ways: per agent, per client, and per case type. Per agent turns it into a staffing question. Per client turns it into a service conversation. Per case type shows where the hours actually go, which in travel is not where the ticket count suggests: a schedule change takes two to three times as long as a routine change.

The trend matters more than the level. A step up that stays is a new client or a new channel. A spike that fades is a disruption day or a mailbox import.

Why it matters

Backlog is the earliest warning the desk has. It grows before response times slip and long before a client notices, so it is the number that tells you tomorrow's problem today.

In travel the tolerance is short. Corporate contracts expect same-day answers and often same-hour acknowledgements, so a backlog larger than one day of inflow means tomorrow's travelers are already queued behind today's.

It is also the metric that most exposes a desk with no closing discipline: tickets that end without ever being marked resolved pile up as a backlog that nobody is working.

Why it matters

It is the metric in every TMC contract and the first question buyers ask when assessing a servicing desk. Buyer guidance is blunt: a TMC that cannot state its average response time per channel is not monitoring it.

It is also the metric most easily gamed. Answering the phone in six rings and promising a callback meets the SLA and delivers nothing, which is why consultants call ring-count SLAs largely meaningless. Measured honestly, on email, it is the fastest read on whether the desk is staffed for its volume.

First response only means something with resolution behind it. A two-minute acknowledgement followed by a four-hour wait is a courtesy, not service.

Why it matters

This is the number the traveler experiences, and the one a travel manager would put in a quarterly review if anyone gave it to them. It separates a desk that acknowledges fast from a desk that finishes fast.

Travel runs on two clocks. Routine work (a quote, a straightforward change) should close the same day, and contracts say so. Airline-dependent work (refunds, waivers) can legitimately stay open for weeks because the airline holds the clock. Blending both into one average hides both.

Read against handling time, it also shows where the waiting happens: a wide gap between the two means the desk waits on third parties, not on itself.

Why it matters

Every case that needs a second reply is usually a case where the desk asked for something it should already have had: the traveler profile, the booking reference, the policy, the fare rule. First contact resolution is that manual lookup problem, measured.

It is the metric most tightly linked to satisfaction. Contact-center research finds each point of FCR is worth roughly a point of customer satisfaction, and that escalated or repeated contacts are where satisfaction drops fastest.

For a TMC it is also a capacity metric. A case that comes back consumes two handles instead of one, so a ten-point rise in FCR is roughly a ten percent rise in cases an advisor can carry.

Why it matters

Handling time is the metric the entire capacity story rests on, and the one fewest travel teams can produce. Cases per advisor per day is handling time turned inside out; without it, 'we saved time' is an adjective.

In travel the work is uneven. A simple time change on a flexible fare takes 8 to 15 minutes of lookup before any decision. A schedule change or a disruption takes 20 to 30 minutes. A tagged handling time shows which workflows are eating the day.

It is also where the value of automation becomes visible: a case the desk prepares before an advisor opens it should show a shorter handle, and a case resolved automatically should show none.

Why it matters

Every reassignment is a case that went to the wrong person first, or to a person without the skill: rail refunds that only senior agents can do, schedule changes that need a ticketing-qualified advisor.

It is expensive. Escalated tickets cost three to five times a first-line ticket in contact-center studies, and satisfaction drops by around 20 points when a case is passed on.

Where routing is automated, reassignment is the honest measure of routing quality: the desk made the first assignment, so every second one is a routing miss the advisor had to correct.

Why it matters

This is the cost multiplier hidden inside every other metric. Support research puts the average issue at about 2.3 contacts, so the true cost of a case is more than double the cost of a contact.

In travel the back-and-forth is worse than in software support, because the desk asks for information it should already hold and the traveler answers from an airport. A booking request needs at least a quote and a confirmation, so the natural floor is three to four messages, not two.

It is the cleanest read on whether context is reaching the advisor. When profile, booking and policy are in front of the person answering, the second and third messages disappear.

Why it matters

This is the number a client contract is written in. Corporate travel agreements are often phrased as 100% compliance, with persistent failure treated as grounds to end the contract.

Today most clients hear about it in a quarterly review deck the TMC assembled by hand, which is why travel managers complain of account managers pasting analytics with no recommendation attached. A per-client SLA number the TMC can show live is a different conversation.

It is a better metric than average response time for the same data, because it resists outliers: one weekend ticket moves an average, not a percentage.

Why it matters

Response SLA says the desk showed up; resolution SLA says it delivered. Buyers are moving toward outcome measures for exactly this reason, and pairing a process metric like answer speed with an outcome metric is the direction TMC service agreements are heading.

Because the target is per client, this is the metric that surfaces uneven service: a desk can run 95% overall while one demanding account sits at 70%. Read it by client before reading it in total.

Combined with the pending count, it is also a forecast. A large pending pile at the end of a shift is tomorrow's breach.

Why it matters

Automation rate is the one operational metric mid-market travel companies already ask for, often before they ask about time. It is also the number buyers want in the next RFP: how many manual tasks the TMC has automated, and what that freed up.

The two definitions diverge sharply. Industry analysis puts AI deflection above 45% of queries while genuine self-service resolution sits near 14%. A desk that reports only deflection is counting travelers who gave up.

For travel the right unit is the case type. Automation of a routine notice can run high; automation of a refund is constrained by the airline, not the desk. Industry-wide, only 40 to 45% of ticket changes and about 5% of refunds are completed with automation at all.