Chapter 3 — Practice-on-your-own solutions

The Big Idea. «Tier-1 tickets resolve 22% faster since the March rollout; tier-2 and tier-3 held steady — renew the contract on the tier-1 case, or negotiate coverage where it earned it.»

The sentence names the specific tier where the improvement lands (tier-1), names the specific number the argument turns on (22% for tier-1, not the 14% blended headline), names the tiers where the tooling did not earn its keep (tier-2, tier-3), and implies the action (renew, but scope to what the tooling actually improved). The audience — the support director in a renewal meeting — can act on it in the next fifteen minutes.

The framing. Comparison. The chart’s job is to compare the post-rollout resolution-time distribution across tiers to the pre-rollout distribution across tiers — and to make it visible that only one tier moved materially. Change framing (before/after on the blended metric) would hide the whole finding; distribution framing would show the shape but bury the tier-by-tier signal.

The chart family. A three-panel small-multiple line chart, one panel per tier, with the pre-rollout baseline as a horizontal reference line on each panel and the rollout date as a vertical reference. The tier-1 panel is annotated; the tier-2 and tier-3 panels are muted. The reader’s eye lands on the panel where the story lives, and the muted panels prove the finding is specific to tier-1 rather than a blended average masking a broader effect.

Why not a single blended line chart? Because the 14% blended headline is the number the tooling vendor will lead with, and the director’s job in the renewal meeting is to negotiate the terms — not to accept or reject the whole contract. The chart’s argument has to survive the vendor’s narrative, which means the tier decomposition has to be visible in the first three seconds. A blended line does not survive that read.

The Big Idea. «Renewal rate is up 3 points on a smaller class of customers; the three largest accounts that did not renew took 12% of the renewal-revenue base with them — the count is winning while the revenue is losing.»

The sentence names both numbers and, critically, resolves the apparent contradiction: the rate is a count metric (how many renewed) and the revenue is a dollar metric weighted by account size. Both can move in opposite directions when the accounts that did not renew are disproportionately large. Naming the mechanism (three largest accounts, 12% of the base) is what makes the sentence action-implying — the reader knows to look at account-level renewal decisions, not at the customer-success playbook that drives count.

The framing. Comparison. Change framing (rate over time, revenue over time) would show two lines moving in different directions — a visible divergence, but not a visible cause. Distribution framing (renewal outcomes across account-size buckets) would show that the non-renewals cluster at the top of the size distribution — and that shape is the story. Distribution across account size is the framing that makes the tension not just visible but explanatory.

The chart shape (Ch 4 preview). A dot-plot or horizontal bar of accounts ordered by ARR descending, colour-coded by renewal outcome (renewed / did not renew / churned), with the top three non-renewals annotated by ARR. In one glance the reader sees the size distribution, the renewal outcomes, and the specific accounts driving the revenue-side story. The 3-point rate increase is a footnote annotation on the same chart, not a competing headline. Lina’s clustered-column failure (Ch 3) is the anti-pattern: two metrics competing for the same eye. The dot-plot solution here sequences them — distribution first, rate as a note — so the tension resolves in the reader’s head instead of hanging as ambiguity.