Free calculator · no email required
What does your support queue cost?
Cost per ticket at a fully-loaded rate, your mix by category, and an honest deflection figure — including the two categories where an automated answer makes things worse and the rate is set near zero on purpose.
How to cost a support ticket
Cost per ticket is the minutes an average ticket consumes multiplied by the fully-loaded hourly cost of the person handling it — salary plus employer overhead, divided by hours genuinely available. Deflection is the share of tickets an automated answer resolves without a human, and it varies enormously by category rather than being one number.
Your queue.
Prefilled with the mix we observe. Export a month of tickets, tag them, and replace every share — the answer changes a lot between a subscription brand and a considered-purchase one.
Fully loaded — the ops cost calculator works it out from salary and overhead.
Support queue, a year
$74,480
$4.43 per ticket × 1,400 a month.
- Honestly deflectableweighted by category, not one blanket rate
- 69%
- Tickets a month
- 967
- Hours returned a month
- 113 h
- Worth a year
- $51,466
Where it comes from
Hours returned are capacity, not redundancy. The same team handles more volume and spends the time on cases where a person changes the outcome.
The categories, and why two are near zero
A deflection rate applied evenly across every category is a business case, not a plan. These are the rates we use.
| Category | Share | Deflectable | Why |
|---|---|---|---|
| Where is my order | 34% | 90% | The answer exists in the order record before the question is asked. The residual 10% is genuinely lost parcels, which need a person. |
| Returns and exchanges | 18% | 70% | Policy-compliant requests close themselves. Outside the window, damaged or high value goes to a human — which is most of the value of splitting them. |
| Product questions | 16% | 65% | Answerable from product data, if the product data exists. Stores that never filled in their specs cannot deflect these, which is the same gap the readiness checker measures. |
| Subscription changes | 12% | 85% | Skips, swaps, pauses and date changes are rules against known data. Also the highest-value category to deflect, because the alternative to a fast skip is often a cancellation. |
| Billing and payment | 11% | 40% | Explaining a charge is easy; refunding one is not. Deliberately low — refunds without a human are where automation stops paying. |
| Complaints and damage | 9% | 10% | Deliberately near zero. A person who has had a bad experience and is answered by a machine has had two bad experiences, and the second one is the one they tell people about. |
Shares and rates are our observation, not a benchmark. They come from DTC helpdesks at this size and are offered so the form is not empty; the figures that matter are the ones you replace them with after tagging a month of your own tickets. metric to confirm. Reviewed 2026-09-09.
What to automate first
By volume × deflectability, not by what irritates the team most.
- Where is my order A third of the queue and the answer already exists in the order record. If you do one thing, do this — and consider that a proactive delay notification removes the ticket rather than answering it.
- Subscription changes Rules against known data, and the highest-value category to resolve instantly: the alternative to a fast skip is frequently a cancellation.
- Returns, within policy Compliant requests close themselves; exceptions route to a person. The split is most of the value, because it is what stops the automation making the wrong call on the hard ones.
- Product questions — but only if your data is good Answerable from product data, if it exists. Stores that never filled in their specs cannot deflect these, which is the same gap the readiness checker measures.
- Not complaints, and not refunds Both are marked low here for the same reason: a wrong answer costs more than the minute it saves, and the person on the other end is already unhappy.
What this leaves out
The tickets you prevent entirely. A proactive delay notification removes a WISMO contact rather than deflecting it, and is cheaper than both. The best support automation reduces the queue rather than answering it faster.
Quality. A deflected ticket that leaves someone unsatisfied costs you later, in a return or a review or a customer who does not come back. Track resolution satisfaction alongside deflection or you are optimising one number into another one's expense.
The platform cost. Helpdesk automation is licensed per seat, per resolution or per message. Subtract it before treating the annual figure as saving.
Escalation. A deflection that fails and then reaches a human has cost more than a ticket answered directly, because the customer has now explained themselves twice.
Definitions
- Deflection
- A contact resolved without a human touching it. Distinct from self-service, where no contact happens.
- Cost per ticket
- Handling minutes × fully-loaded hourly cost. Rarely under a dollar, rarely over ten.
- WISMO
- Where is my order. The largest category in almost every DTC helpdesk.
- Fully-loaded cost
- Salary plus employer overhead over hours genuinely available.
- Escalation
- A deflection that fails and reaches a person. More expensive than never having deflected it.
Questions about deflection
How do I calculate cost per support ticket?
Minutes spent on an average ticket, times the fully-loaded hourly cost of the person handling it. Fully loaded means salary plus employer overhead divided by hours genuinely available — around 1,725 a year rather than 2,080. Most brands underestimate this by a third by using salary alone.
What share of tickets can actually be deflected?
It depends entirely on the mix. Where-is-my-order is the easiest category in support because the answer already exists in the order record; complaints are the hardest and should barely be deflected at all. The overall figure falls out of your own distribution, which is why this calculator lets you edit every share rather than quoting one number.
Why is the complaints deflection rate so low?
Because it should be. Someone who has had a bad experience and is answered by a machine has now had two bad experiences, and the second is the one they tell people about. We set it near zero deliberately. A tool that claimed 80% deflection across every category would produce a better-looking business case and worse outcomes.
Is deflection the same as self-service?
Close, and the distinction matters. Self-service is a customer finding an answer without contacting you. Deflection here means a contact that arrives and is resolved without a human touching it. The second is easier to measure and harder to fake, because the ticket exists either way.
Are the category shares a benchmark?
No. They are what we observe across DTC helpdesks at this size, offered so the page is not an empty form. Export a month of tickets, tag them, and replace every share — the answer changes substantially between a subscription brand and a considered-purchase one.
Does deflection reduce headcount?
Usually not, and it is more honest to say so. It absorbs growth: the same team handles more volume, and the time returns to work that needs a person — proactive outreach, complex cases, the ones where a good reply changes a customer's mind. Presenting it as a redundancy plan is how these projects lose the support of the team that has to make them work.
Related
Find out what you’re losing.
Before you commit to anything, we tell you exactly what you’re losing and what it costs to stop it. Two weeks. Fixed fee. Credited in full against any build you go ahead with.
- Fee
- $1,500–$3,000, fixed
- Duration
- Two weeks
- Credited
- In full, against any build
- You supply
- Read access + one 45-minute call