You can estimate the labor recorded for support work by separating first contacts, repeat contacts, reopened conversations, and supervision, then applying your own labor rates. That gives you a direct labor estimate, not proof of the cost caused by bad support. The calculation measures work in your records. Investigating avoidable contacts or lost customers requires separate evidence.
Use one consistent period and definitions, and mark estimates clearly. For another method focused on assessing conversation outcomes, see how to review your AI Agent’s conversations. For context on support Agents and deployment, read AI Agents for Customer Support.
Define the measurement before collecting numbers
Choose a period your staffing and contact records cover, such as a calendar month. Decide which channels and roles are in scope. Write down how you will identify multiple contacts about the same issue. Use the same period for every record you combine.
Fill in this scope note:
- Period measured:
- Channels included:
- Roles included:
- Time unit: minutes or hours:
- Rule for matching contacts to the same issue:
- Known gaps or unusual conditions:
A conversation thread is not necessarily the same thing as an issue. Someone may follow up in another channel, or reopen a conversation you considered finished. If your records cannot reliably connect contacts to one issue, mark the relationship as unclear instead of assuming they match.
Note changes that could make the period hard to compare with another one, such as different staffing, support hours, channels, or record-keeping practices. You can still calculate a total, but preserve those notes so the number is not presented without its context.
Record staffing time and its source
List each role that handles support work and the time spent during the period. Use time records that separate support from other duties when available. If you use scheduled hours, label them as scheduled hours; they do not necessarily represent active handling time.
| Role | People included | Support minutes | Record or estimate? | Source and notes |
|---|---|---|---|---|
| Example role | 1 | 120 | Recorded | Time log |
For each role, keep a note of the labor rate you plan to use and what that rate includes. Count only time you can reasonably assign to support when a person has other responsibilities. Keep roles separate if they have different rates.
If the time is estimated, write down how you estimated it. For example, note whether the figure came from a sample of shifts or a staff estimate. A number with a clear limitation is more useful than an exact-looking figure with no explanation.
Separate first contacts, repeats, and reopenings
Use definitions that fit your records and apply them consistently. For this worksheet, a repeat contact is another contact about the same continuing or unresolved request. A reopened conversation is work that resumes in a conversation previously treated as finished. Record their counts and time separately where your records allow.
| Issue or conversation reference | First-contact minutes | Repeat-contact count | Repeat-contact minutes | Reopened count | Reopened minutes |
|---|---|---|---|---|---|
| Example A | 8 | 2 | 10 | 1 | 5 |
In this illustrative count example, the recorded time is 8 + 10 + 5 = 23 minutes. It is an arithmetic example, not a customer result. The two repeat contacts and one reopening are separate counts; adding them does not tell you the number of unique issues. One issue can have both repeat contacts and a reopening.
If you cannot classify a follow-up with confidence, put it in an unclear group or report a combined repeat-and-reopen figure. State what the combined figure includes. Do not force a distinction your records cannot support.
Count supervision without counting time twice
Record time spent reviewing support work, coaching on specific cases, handling escalations, or correcting an answer or process. Use separate lines for different activities if that helps you see where the time went.
- Activity and role:
- Minutes spent:
- Related case or review sample, if known:
- Recorded or estimated:
Check whether supervision time is already included in a staffing total. If it is, do not add it again. Similarly, keep routine meetings out of the support-work total unless you can identify the portion that relates to support review. If you estimate supervision from a sample, record how many items you reviewed and how you chose them; a sample is not a full-period count.
Calculate the direct labor estimate
If each category is distinct and the time does not overlap, add the minutes:
Total support minutes = first-contact minutes + repeat-contact minutes + reopened-conversation minutes + supervision minutes
Use a complete staffing total instead if it already includes those activities. Do not add the category totals to that staffing total. Where possible, compare the staffing view with the issue-level view: the staffing figure can summarize time, while the issue-level records help describe how contacts were classified.
To convert minutes into a labor estimate, calculate each role separately:
Direct labor estimate = sum across roles of (support minutes for that role ÷ 60 × that role’s labor cost per hour)
Use the labor rate that fits your accounting method. Decide whether it includes wages alone or also benefits and employer costs, then use that definition consistently. If you do not have a reliable rate, report time rather than inventing a cost. This calculation covers the labor in your chosen scope, not every possible support expense.
Include human handover time when it is in scope
If a website Agent is part of your support process, decide whether and how you will include human-takeover conversations. Count active staff time spent reading the conversation, replying, and following up. Keep elapsed waiting time separate from active labor unless you are deliberately measuring both.
Authorized Workspace members can take over a conversation and reply as a human. Waiting or human-takeover conversations do not receive an AI response from the widget. Human handover is a route for help, not a response-time guarantee. See the human handover feature for related product information.
Do not automatically classify a handover as a repeat contact, a failure, or an extra issue. Apply your definitions and record the work. Whether a conversation was handled appropriately is a separate question from how much staff time it took.
Check the worksheet before comparing periods
Review these points before you use the total:
- Same dates: Staffing and contact records cover the same period.
- Stable definitions: First contacts, repeats, reopenings, and supervision were classified using the same rules.
- No duplicate time: The same minutes are not included in both staffing and category totals.
- Visible estimates: Estimated time and unclear classifications are identified.
- Consistent rates: You state what each labor rate includes.
- Comparable records: Changes in staffing, hours, channels, and recording practices are noted.
If you compare two periods, place the counts and time totals side by side and include any changes to how work was recorded. A difference in recorded labor is a difference in measured workload. By itself, it does not establish the reason for the difference or show that a support change caused it.
Write a conclusion the records can support
Finish the worksheet with three short notes: which category accounts for the most recorded time, which figures are estimated or incomplete, and what record would improve the next calculation. You can report the direct labor estimate and the method used to calculate it.
Keep the interpretation narrow. Repeat contacts or reopenings may be useful reasons to review conversations, but this worksheet alone cannot establish that poor support caused them. It cannot establish churn, lost revenue, or indirect costs that are not part of your records. Describe what you counted, identify the gaps, and treat explanations for the counts as questions to investigate rather than measured results.
Worked calculation with explicit assumptions
Suppose a fictional monthly record contains 20 hours of first-contact work, 5 hours of repeat-contact work, and 3 hours of supervision, all at an assumed loaded rate of $25 per hour. The direct labor estimate is (20 + 5 + 3) × $25 = $700. If a two-hour handover review is already included in supervision, do not add it again.
This does not establish lost revenue or prove that all repeat contacts were avoidable. Keep those questions separate. Use the conversation review example to classify the work, then the future support investment guide to choose a specific improvement to evaluate.