Total ticket volume across Q1 was 4,621 — essentially flat versus Q4. That flat number is the most misleading thing in the dataset. The mix of what customers contact you about moved sharply, and a flat headline hid all of it.

Billing confusion is the single largest contact driver — 1,201 tickets, 26% of everything. It is also the most expensive to resolve: a median of 19 hours per ticket against 6 hours for the support average. Billing is one quarter of your volume and close to half of your support cost.

The second story is growth. Onboarding-setup tickets tripled quarter on quarter — from 7% of the mix to 18%. It is not your biggest theme yet. On its current slope it will be by next quarter. Flat total volume meant nobody noticed.

What is working: account-access and how-to tickets are high-volume but cheap and falling — self-service is absorbing them. What is not: billing is expensive and sticky, and onboarding is growing fast enough to become the headline.

Six themes account for the entire ticket dataset. Ranked by volume, the top three are billing confusion (26%), bug reports (21%), and account access (19%) — 66% of contacts.

But volume alone sends you to the wrong place. Account access is high-volume and almost free: a median 2-hour resolution, mostly password resets, and falling steadily as self-service catches it. Staffing for it is fine; fixing it returns almost nothing.

Billing confusion is the opposite. Same rough volume tier, but every ticket is a long, multi-touch investigation — "why was I charged this", "what tier am I on". Median 19 hours. This is not a queue to staff faster. It is a product-clarity problem leaking into support.

Onboarding setup is only 11% by volume today, which is exactly why it is dangerous. Read the raw growth and it tripled in a quarter. Volume tells you what to staff this week; effort tells you what to fix this quarter; growth tells you what will be next quarter’s headline. Billing fails the effort test. Onboarding fails the growth test.

Account-access tickets close in a median 2.1 hours. Billing-confusion tickets take 19.0 — nine times longer for the same one ticket on the queue.

This is the trap with a volume-ranked ticket dashboard. By count, billing and account access look like comparable problems. By cost, they are not in the same category. Billing is 26% of tickets but consumes roughly 46% of total support hours. Account access is 19% of tickets and 5% of hours.

That reframes the work. Speeding up the account-access queue saves almost nothing — it is already cheap. Removing the cause of billing tickets is the highest-value operational move available, because every billing ticket you prevent is worth nine account-access tickets in agent time.

The ANOVA confirms the theme genuinely predicts resolution time (F=42.6, p<0.001, η²=0.24 — a large effect). Resolution time is not random noise across agents. A quarter of the variance in how long a ticket takes is explained purely by which theme it belongs to. The expensive themes are expensive structurally, not by accident.

Onboarding-setup tickets were 6% of the mix in week 1 and 18% by week 13 — tripled in a single quarter, while total volume stayed flat. A flat headline is what let this hide: as onboarding rose, account-access fell, and the total never moved.

Reading the onboarding tickets explains the cause. The phrase "new dashboard" appears in 61% of them, against near-zero before the week-4 UI change. The redesign did not break anything — it relocated everything. Customers cannot find the setup steps they used to know, so they open a ticket.

This is the finding a volume ranking will never surface, because onboarding is still only the fifth-largest theme by count. But the slope is steep and steady, and if it holds, onboarding becomes a ~28% theme next quarter — overtaking billing as the top contact driver.

The regression on week is strong (t=6.71, p<0.001, R²=0.79). This is a trend, not a blip. The good news: a trend with a clear cause has a clear fix — and catching it now, at 18%, is far cheaper than catching it next quarter at 28%.

The frequency chart ranked themes by ticket count. This one multiplies each theme’s volume by its median resolution time — the actual hours an agent spends. The ranking rearranges, and the rearrangement is the point.

Billing confusion consumes 22,800 agent hours this quarter — 47% of the total support budget — from just 26% of tickets. Bug reports add another 13,700 hours. Two themes alone burn 75% of all support hours. The other four themes combined cost less than bug reports does on its own.

The mirror image is account access. It is the highest-volume theme on the queue, the one a count-based dashboard tells you to staff hardest — and it ranks second-to-last by hours, at 1,800. Every account-access ticket is a two-hour password reset; a thousand of them still cost less than 150 billing tickets.

This is the chart that should set the headcount conversation. Staffing follows hours, not tickets. And reducing hours follows removing causes, not adding agents — every billing ticket prevented at the source is worth nine account-access tickets you will never need to touch.

First-contact resolution is the share of tickets closed on a single reply, with no back-and-forth. It explains why the expensive themes are expensive — and it splits the six themes cleanly into two groups.

Account access resolves first-time 94% of the time. How-to questions 83%. These are self-contained: a customer asks one thing, an agent answers it, the ticket closes. That is why they are cheap, and it is why automating them works — a self-service flow only needs to handle a single-step interaction.

Billing confusion resolves first-time just 28%. Roughly three in four billing tickets need a second, third, or fourth exchange before they close. Each round-trip waits on the customer, re-queues, and re-loads context for whichever agent picks it up. That is where the 19-hour median actually comes from — not slow agents, but a conversation that structurally cannot finish in one message.

That reframes the fix one more time. You cannot automate billing the way you automate password resets, because a billing ticket is not one question — it is an investigation. The lever is to make the investigation unnecessary: an account page clear enough that the question never has to be asked. Lift billing’s FCR and the resolution time, the hours, and the queue all fall together.

The chi-squared test confirms the theme genuinely predicts whether a ticket closes first-time (chi2=311.4, p<0.001, Cramer’s V=0.37 — a medium-to-strong effect). The expensive themes are expensive because of what they are, not who handled them.

These are representative tickets, not cherry-picked extremes.

On billing (the expensive theme): "I have been charged twice the amount I expected and I genuinely cannot tell from my account page why. Please explain what tier I am on and why." The pattern in 1,201 billing tickets: not a payment failure, a clarity failure. Every one is a manual investigation.

On onboarding (the growing theme): "Since the new dashboard I cannot find where to add my team. The old setup checklist seems to be gone." "New dashboard" appears in 61% of onboarding tickets. The redesign moved the furniture and did not leave a map.

On account access (the cheap theme): "Locked out, need a password reset." Short, self-contained, two hours to close — and falling, because the self-service reset flow is catching most of them. This is what a healthy theme looks like.

From survey free-text: "Support is great when I reach them. I just wish I did not need to — half my tickets are things the product should have told me." The whole report in one sentence.

Three things, in priority order:

1. Remove billing tickets at the source. This is the highest-value move in the report. Billing is 26% of volume but ~46% of support hours. The fix is not faster agents — it is a clearer account page: a plain-language tier summary, a line-item charge breakdown, and a "why this amount" explainer. Every billing ticket prevented is worth nine account-access tickets in agent time.

2. Fix onboarding now, at 18%, not next quarter at 28%. The cause is specific and known: the week-4 dashboard redesign hid the setup path. Restore an in-product setup checklist and a visible "add your team" entry point. This is a UI fix, and it is far cheaper to make before onboarding becomes your top driver.

3. Leave account access and how-to alone. Both are high-volume but cheap, and both are falling — self-service is doing its job. Resist the instinct to optimise the biggest queues by count. They are already the cheapest hours you have.

Note what is not here: feature requests are only 7% and low-cost — log and route them, do not build a programme around them yet.

If billing stays as-is and onboarding keeps its slope, next quarter looks worse than the flat total will admit: onboarding climbs toward 28% and overtakes billing, while billing still consumes nearly half your support hours. Median resolution time keeps drifting up, and a flat ticket count keeps hiding it.

If the account-page clarity work and the onboarding-checklist fix both ship: billing volume should fall toward 18% of the mix as the "why this charge" tickets stop being opened, and onboarding should flatten near its current level rather than tripling again. Modelling those two moves against current volume cuts total support hours by roughly a quarter — without adding a single agent.

Account access and how-to need no investment; the self-service flows are already absorbing them. The whole next-quarter thesis is simple: stop billing tickets at the source, fix the onboarding path the redesign broke, and stop letting a flat headline hide a shifting mix.