Customer Support Cost Benchmarks by Company Size (2026)

Cost per contact is the common choice and cost per resolution is the more honest one.

Per contact

Dividing by contacts is simple and rewards deflecting contacts regardless of whether the customer's issue was addressed.

It also worsens mechanically as automation absorbs simple contacts and leaves a more expensive mix.

Per resolution

Dividing by resolved issues accounts for repeat contacts, which is where the cost of poor first-time handling actually shows.

This is harder to calculate and is the version that supports decisions rather than presentations.

Per customer

Support cost per customer per period is the measure that matters for subscription businesses and connects to unit economics directly.

It also normalises for scale in a way that per contact does not.

Percentage of revenue

Useful for executive reporting and largely uninformative operationally, since it moves with revenue rather than with support performance.

It is worth reporting alongside rather than instead of an operational measure.

Per contact reason

Allocating cost by contact reason identifies which categories are worth eliminating at source rather than handling more efficiently.

This is where the largest available savings usually sit and a blended figure hides it entirely.

Structure, market and channel mix affect support cost considerably more than headcount does.

Where the team is

Labour cost varies enormously by market, which means two identically performing operations can differ several-fold in cost per contact.

Comparing across markets without adjusting for this produces conclusions about geography rather than performance.

In-house versus outsourced

Outsourced arrangements price differently and frequently exclude costs the client still bears, which makes the comparison misleading.

Management, quality oversight and knowledge remain internal costs in most outsourced models.

Channel mix

Phone, chat and email have very different cost profiles, and an operation weighted toward one is not comparable to one weighted toward another.

Concurrency in chat means it behaves quite differently from voice on a per contact basis.

Automation share

An operation with substantial automation has a different cost structure entirely, with more fixed and less variable cost.

Comparing per contact figures across different automation levels compares structures rather than efficiency.

Automation shifts cost from variable to fixed and makes per contact figures behave counterintuitively.

The mix effect

Absorbing simple contacts leaves a more complex mix, which raises cost per remaining contact even as total cost falls.

A rising cost per contact after successful automation is expected and is frequently misread as a problem.

Fixed platform cost

A platform cost that does not scale with volume changes how the total behaves as the business grows, which is the actual benefit.

Per contact figures obscure this, since the fixed cost divides differently at different volumes.

Maintenance as a real cost

Knowledge maintenance, review and configuration are ongoing costs of an automated operation and belong in the calculation.

Omitting them makes automation look cheaper than it is and produces business cases that do not survive examination.

Building a Baseline

Your own cost, calculated consistently over time, is the only comparison that supports a decision.

Document the scope

Writing down what is included, and applying it consistently, is what makes a figure comparable across periods.

Most unexplained movements in support cost turn out to be scope changes rather than performance changes.

Track the components

Reporting people, tooling and management separately shows which is driving a change, which a single figure cannot.

It also makes the effect of automation visible as a shift between components.

Segment by channel and topic

Cost by channel identifies where investment would help most, and cost by topic identifies which contact categories are worth eliminating.

Both are more actionable than a blended figure and require only allocating volume.

Published benchmarks are useful as a sanity check on the direction rather than as a target.

Spotting outliers

A cost far outside what is typical for a comparable operation suggests something worth investigating rather than a target to hit.

This is a legitimate use and does not require the comparison to be precise.

Supporting an argument

A comparison showing support cost per customer is unusually high can support an investment case, treated as indicative rather than authoritative.

Stating the comparability limitations makes the argument stronger rather than weaker.

Checking assumptions before scaling

An organisation planning growth benefits from a rough sense of how support cost typically behaves at larger scale.

This should be replaced with actual data as soon as the growth begins.

Chatboq reports the volume, resolution and channel data the calculation requires, segmented so cost can be attributed properly.

Resolution rather than contacts

Resolved issues, adjusted for repeat contact, is the denominator that produces an honest cost figure and is reported directly.

Volume by channel and topic allows cost to be allocated rather than blended.

The automation split

Volume handled automatically and volume reaching agents are reported separately, which is what the fixed and variable cost split requires.

This is also what shows whether growing volume needs growing headcount.

Trend rather than snapshot

Consistent measurement over time is what makes a cost figure useful, since the comparison that matters is against your own previous period.

Configuration changes are recorded alongside, so a movement can be attributed rather than theorised.