Blog · Forecast and pipeline · Customer service
How a customer service leader turns the ticket export into a weekly list of accounts whose contact volume has jumped against their own history, why the account's own baseline beats a global threshold, the surge score, and the two readings every surge has: a product problem or an account about to leave.
A support team sees tickets. A commercial team sees accounts. The ticket export joined to the account list shows one thing neither sees alone: the account that raised three tickets a month for a year and raised fourteen last week. This guide sets out the surge list built from the account's own baseline.
Per account, per week:
Baseline = median weekly tickets over the trailing twelve weeks Surge ratio = this week's tickets ÷ baseline Surge if ratio ≥ threshold and this week's tickets ≥ floor
Ranked by account revenue.
Account identifiers only.
| Pattern | Reading | Owner |
|---|---|---|
| Several accounts surging, shared category, category up across the base | Product problem | Engineering, with the account list |
| One account surging, multiple categories, nobody else | Account at risk | Account manager |
| One account surging, one category | Could be either; check the category trend | Support lead |
| Account | Revenue | Baseline | This week | Ratio | Categories | Reading |
|---|---|---|---|---|---|---|
| 4471 | $840,000 | 3 | 14 | 4.7 | 4 | Account at risk |
| 2210 | $610,000 | 8 | 21 | 2.6 | 1: exports | Product, with 9 others |
| 9034 | $95,000 | 1 | 6 | 6.0 | 2 | Account at risk |
Ten accounts surged on exports this week and the exports category is up threefold across the base; that is one problem for engineering. Account 4471 is alone, across four categories, and worth $840,000; that is a call from the account manager today, not a support metric.
Σ tickets per week = Σ accounts' tickets
A ticket with no account fails it and is listed. Unassigned tickets are usually trial users or a routing error, and either way they are not on any surge list until they have an account.
Global threshold. The big accounts are always surging and the small ones never are.
Baseline too short. Four weeks of history makes a normal fluctuation a surge. Twelve is the usual minimum.
Product problems worked as account problems. Ten account managers call ten customers about the same bug. Split by category first.
Silence ignored. An account whose tickets went to zero is the other half of the signal, and has its own list.
Mapped once, the ticket export and the account list produce the baselines, the surge list and the category split every week. Covirage builds this from the exports as they are. The customer service page describes the setup, and the silence and surge guide covers the paired signal on the other side.
Because a large account raises twenty tickets a week as a matter of course, and a small one raises one a month. Twenty is normal for the first and a crisis for the second. Each account against its own history is the only fair test.
Two to three times the account's median is the usual starting point, with a floor so that an account going from one ticket to three does not trigger. Both numbers are on the report.
By the category on the tickets. If the surging accounts share a category this week and that category is up across the base, it is a product problem. If one account's surge spans categories and nobody else is surging, it is that account.