Slide 1 of 4
Health is the action surface
Retention and Expansion explain the picture. Health bins every customer so the at-risk list can be attacked directly, rather than over a Slack thread.
Renewals and expansion with the Customer agent
DoLesson 3 of 9Read the health picture
Finding your at-risk book
Produce the specific, owned list of at-risk accounts you will work this week, and establish the cadence that keeps it current.
23 minIntermediate
Part 1 of 4
The surface built for doing something
The three reports divide by purpose. Retention is about keeping the revenue you have, Expansion about growing from the base; both are for understanding. Customer Health is the working list: it bins every customer into Healthy, Neutral or At-Risk so the at-risk set can be worked directly.
That division is worth respecting in how you spend time on each. Retention and Expansion reward being read carefully once a month, when the levers and segment mix are the question. Health rewards being worked weekly, because its output is a list of accounts somebody has to do something about, and a list of accounts nobody walked is indistinguishable from not having the list.
Part 2 of 4
Filter before you look
Start by narrowing to the book you actually own. Fiscal year and quarter set the period; the team filter narrows to Enterprise, Mid-Market, SMB or Strategic. Each of those segments tells a different retention story, and a company-wide view is an average across all of them, which is precisely the reading the previous lesson argued against.
It is also a practical matter of scale. The at-risk list for a whole company is too long to walk in a weekly meeting, so it does not get walked. The at-risk list for one segment is a manageable number of accounts, which is the difference between a cadence that holds and one that quietly lapses after three weeks.
Part 3 of 4
Walk it, and close each one out
The distribution supports a direct working motion: click a slice or its legend row to open the first account in that segment, then move through the rest with the left and right arrows. That turns the at-risk set into a queue rather than a chart, and walking it once a week is the cadence the surface is designed around.
One rule makes it work: every account leaves the review resolved. Anything else and the same accounts reappear next week a week older, and the meeting slowly becomes a status ritual instead of a save motion.
- Escalate — a named owner and a date, recorded on the account
- Close out — with the reason it is not actually at risk
- Never "keep an eye on it", which is neither and produces next week's identical list
Part 4 of 4
A worked walk
What a real weekly pass looks like, narrated. Filter: current quarter, Mid-Market team. The At-Risk slice shows nine accounts. Click the slice; the first account opens.
Account one: score dropped on support history; the escalation closed last week and the record has not caught up. Close out, with the note "escalation resolved 14th, expect recovery next cycle." Account two: single-threaded and the champion has gone quiet, so escalate to the AE by name, date Friday, note "breadth before value: get a second contact." Accounts three through seven: worked the same way with the arrow keys, two escalations, three close-outs. Account eight: genuinely ambiguous, with flat usage and no obvious cause. That one gets escalated too, because "I could not tell from the record" is precisely what escalation is for; keep-an-eye-on-it is the only forbidden outcome. Account nine: closed out, seasonal customer in their off-season, same note as last year.
Nine accounts, roughly twenty minutes, every one resolved to a name-and-date or a reason. Next week's list will be different accounts, which is the proof the walk is working.
Do this in the product
Build and walk the list
- Open Customers and set the fiscal year and quarter you are operating on.
- Filter to your team so you are looking at your book rather than the company average.
- Open the Customer Health tab and click into the At-Risk segment.
- Walk every account in the segment using the arrow navigation.
- For each, either escalate it to a named owner with a date, or close it out with the reason it is not at risk.
- Put the walk in the calendar as a weekly item, and add the Retention tab to your monthly business review instead.
Key questions
You should be able to answer each of these from memory before opening it. Recalling the answer is what makes it stick; recognizing it when you read it does not.
Why filter to your team before opening the at-risk list?
Segments tell different retention stories, so a company-wide view is an average across all of them. It is also too long to walk in a weekly meeting, which is how the cadence quietly lapses.
What are the only two acceptable outcomes for an account in the weekly walk?
Escalated to a named owner with a date, or closed out with the reason it is not actually at risk. "Keep an eye on it" is neither, and produces the same list next week one week older.
What to take away
- Retention and Expansion are for understanding. Health is the working list. Give them different cadences.
- Filter to your segment first — a company-wide at-risk list is too long to walk, so it does not get walked.
- Every account leaves the review escalated with a name and a date, or closed out with a reason. Nothing else.
Teach this lesson
The argument in 4 slides, for presenting it to your team
Slide 2 of 4
Filter to your book first
Enterprise, Mid-Market, SMB and Strategic each tell a different retention story. Operate on yours, not on the company average.
Slide 3 of 4
Walk the list weekly
- Click a slice or legend row to open the first account in that segment
- Arrow keys walk the rest
- Escalate or close out each one — no account stays open unresolved
Slide 4 of 4
Every account leaves with an owner and a date
A review that ends in shared concern and no assignment produces the same list next week, one week later.
Demo
Try the demo.
See agents carry the repeatable work of GTM across sales, marketing, customer success, and RevOps. Every action prepared, reviewed, and recorded. Fictional data, real product.
Explore the demo