Trace New

The situation

Capsule is a digital pharmacy delivering prescriptions directly to patients across major U.S. markets. The business runs across 10,000+ SKUs, multiple markets, and a wide mix of payors — which means almost nothing interesting about performance lives on the topline. The real story is always in the mix underneath.

Finance was bounded by what a spreadsheet and a few hours of analyst time could realistically produce: mostly topline, with limited segmentation and limited room to understand why the numbers moved the way they did.

Peter, Capsule’s Head of Data, saw the opportunity from a different angle. He wanted to support Finance — but not by pulling analyst hours into deeper slicing and dicing every month. Trace was brought in to solve both: give Finance a substantially richer business review, faster, without Data spending analyst cycles on manual decomposition.

What Trace automated

01 Instant MoM bridge with a waterfall view

Core KPIs render as a waterfall showing exactly how they moved from last month to this month — which drivers added, which subtracted, and by how much. The bridge that used to be built by hand is now an instant view.

02 Calendar-aware normalization

Months aren’t the same length, and performance shifts across days of the week. A February-to-March comparison can show a 5% lift that’s purely calendar. Trace adjusts for uneven days and day-of-week mix automatically — cutting hours of reconciliation and eliminating a recurring source of confusion in exec conversations.

03 Mix vs. rate decomposition

Gross profit per unit is one of Capsule’s most-watched metrics and one of the easiest to misread. GPU can swing from mix shift alone — more generics, fewer specialty fills — even when per-unit margin is flat. Trace separates rate change from mix shift by default, on every metric in the tree, every cycle.

04 Segmentation at full granularity

Trace handles the full 10,000+ SKU list natively and combines it with market and payor cuts — so when a GPU or revenue shift originates in a specific drug, market, or payor combination, it surfaces automatically. No manual hunting through filters in Looker.

Going live

Trace read directly from Capsule’s Redshift warehouse on top of the existing dbt models that powered Looker. Because the metric definitions were already formalized in dbt, onboarding was mostly column mapping — and the tree mirrored the spreadsheet, which made validation straightforward.

What's different now

The monthly review now starts with the drivers, not the spreadsheet. Darshan and the CFO can open Trace, see the MoM bridge, and move directly into what changed and why — with deeper decomposition available by default.

→ Rich driver-level insights, without adding analyst hours.

→ Calendar and mix effects controlled by default — two common sources of false signals, handled automatically.

→ The Data team stays focused on driving strategy and operations.

Modeled in Trace

Payors

Markets

10,000+ SKUs

Mix vs. rate

Gross profit per unit

Revenue

Beyond the business review

Trace’s AI analyst now operates on the same metric-tree model behind Capsule’s monthly review — proactively investigating material shifts, finding the driving SKU, market, or payor, and surfacing the explanation without requiring Finance to work through the analysis manually.

The metric tree provides the business context. The AI analyst does the analysis.

Capsule

Deeper, faster business reviews, every cycle.

10,000+

Instant

Redshift

Core KPIs render as a waterfall showing exactly how they moved from last month to this month — which drivers added, which subtracted, and by how much. The bridge that used to be built by hand is now an instant view.

Trace handles the full 10,000+ SKU list natively and combines it with market and payor cuts — so when a GPU or revenue shift originates in a specific drug, market, or payor combination, it surfaces automatically. No manual hunting through filters in Looker.

All customer stories