What running support out of Gmail actually costs you
September 8, 2026 · 8 min read
Support feels free when you are the one doing it. Nobody sends you an invoice for the forty minutes you spent this morning telling six people where their orders were. It comes out of the same pot as everything else you do, so it never shows up as a line item.
That is exactly why it grows unchecked. Here is how to put a number on it.
The arithmetic
Three inputs.
Tickets per day. Count yesterday. Not a normal day — yesterday. Most merchants guess low by about half, because the quick ones do not register as work.
Minutes per ticket. Not typing time. Time from opening the message to closing the tab, including looking up the order, checking the tracking number, and re-reading what you sent last time so you do not contradict yourself. For a store without a helpdesk this is usually four to eight minutes, and the lookup is most of it.
What your hour is worth. If you would pay someone $20 an hour to do it, use $20. If you would rather be doing something that grows the store, use what that is worth instead — which for most owner-operators is higher.
Multiply:
tickets/day × minutes/ticket × 22 days ÷ 60 × hourly rateFifteen tickets a day, six minutes each, at $25 an hour is $825 a month — just under $10,000 a year. Thirty tickets at eight minutes is $2,200 a month.
Those are not dramatic numbers. That is the point — small enough to ignore individually, large enough to matter annually.
Find your row:
| Tickets a day | At 4 min each | At 6 min each | At 8 min each |
|---|---|---|---|
| 5 | $183 | $275 | $367 |
| 10 | $367 | $550 | $733 |
| 15 | $550 | $825 | $1,100 |
| 20 | $733 | $1,100 | $1,467 |
| 30 | $1,100 | $1,650 | $2,200 |
| 50 | $1,833 | $2,750 | $3,667 |
Two things stand out. Going from four minutes to eight doubles the bill without a single extra ticket arriving — which means the lookup, not the volume, is what you are actually paying for. And at fifteen tickets a day, a modest number for a growing store, support is roughly a $10,000 line item you have never seen written down.
Where a single ticket's minutes actually go
The number above is only useful if the minutes estimate is honest, and most people underestimate because they think of the typing. Here is a routine "where is my order" broken down.
| Step | Typical | Avoidable? |
|---|---|---|
| Read the message | 20 sec | No |
| Find the order in Shopify | 60–90 sec | Yes, if it is already attached |
| Check fulfilment and tracking | 45 sec | Yes, same |
| Re-read what you said last time | 30–60 sec | Yes, if the thread is in one place |
| Write the reply | 60–90 sec | Partly, with saved answers |
| Get back to what you were doing | Not counted, and not small | Yes, by batching |
The striking part is the ratio. Less than a third of a support ticket is answering the question. The rest is assembling the context needed to answer it, and then recovering from the interruption.
Which explains why hiring someone rarely halves the cost the way people expect. You are not buying more typing capacity. The lookup is still there, and now a second person is doing it.
What the number leaves out
Three costs the arithmetic misses, all of which are bigger than the arithmetic.
Context switching
A support question that arrives while you are doing something else costs more than the minutes it takes. You put down what you were doing, look up an order, write a reply, and then spend a while finding your place again. Six ticket interruptions is not forty minutes of support work. It is a day where you also did not finish anything else.
The ones you drop
Every merchant working out of a shared inbox has lost a message. It got read on a phone, marked as read, and never answered. You do not find out. The customer does not email again — they just do not come back, and possibly tell someone. There is no line in your accounts for this and it is probably the largest number on this page.
Answering the same question repeatedly
If a fifth of your tickets are the same five questions and you write a fresh reply each time, you are paying for that answer over and over.
The cheapest fix, which is free
Before any tooling: stop answering support as it arrives.
Six messages handled as they land is six interruptions, each costing the reply plus the recovery. The same six handled in one sitting is one interruption, and you get faster within the batch because you stay in the same frame of mind and often reuse a sentence from the previous reply.
Twice a day is usually the right cadence for a small store — morning and late afternoon. That keeps every reply comfortably inside a working day, which is all the responsiveness most customers want, while removing most of the switching cost.
What actually reduces it
Where the minutes go
Answering from a shared inbox
- Search the order in another tab
- Check tracking in a third
- Scroll the thread for what you said last time
- Guess whether a colleague replied
- Rewrite the same answer from scratch
Answering from a ticket
- Order already on screen
- Tracking already on screen
- Whole conversation in one thread
- Who replied, visible
- Paste from a saved answer
Cut the lookup. If the order, tracking number and customer history are already on screen when you open the message, ticket time roughly halves. This is the single biggest lever and it is mostly a tooling problem — you are not typing slowly, you are switching tabs.
Make waiting visible. You cannot drop a message that is showing you how long someone has been waiting. Less about speed than about nothing falling through.
Write the common answers once. Five questions covering a fifth of your volume is an afternoon of work. Some customers will find them. Some will not, and you will paste rather than compose.
Keep the conversation in one thread. Half the re-reading is reconstructing what was already said. When replies come back to the same place, that work disappears.
What fixing it costs
Worth putting next to the number, because the comparison is usually lopsided enough to make the decision obvious.
| Fix | Cost | Removes |
|---|---|---|
| Batch twice a day | Free | Most of the context switching |
| Write your common answers down | One hour, once | The repetition |
| Order context beside the message | $0–19/mo | Roughly half of every ticket |
| Threaded replies and visible status | Same | The re-reading, and the dropped ones |
| Hiring someone | $1,500+/mo | Your hours, not the underlying cost |
The last row is the one to look at twice. Hiring transfers the work rather than reducing it — the lookups still happen, someone else does them, and you have added a management job. Worth doing when volume genuinely demands it, but doing it before fixing the four-minute lookup means paying someone to do avoidable work. See managing your first customer support hire if you are at that point.
Run your own numbers
Yesterday's ticket count, an honest estimate of minutes, and a rate you would actually pay. If the monthly figure comes out under about $300, your current setup is fine and you should ignore all of this. Once it is over $500 — which is ten tickets a day at six minutes — the fix costs a fraction of the problem.
Related
This page counts the hours support costs you. What good customer service is actually worth does the other side — the repeat business it earns or loses.
If the number came out high enough to think about hiring, managing your first customer support hire covers the handover. If it came out high because of the lookup rather than the volume, the problem with Shopify's contact form for customer support is the other half of that.
Support that already knows what your customer ordered
HelpHQ turns messages into tickets inside your Shopify admin, with the order beside the conversation. Free for your first 50 tickets a month.
See how it works