Case studyNimbus SaaS (Postgres)May 10, 20264 min read

The $454K Hiding Behind Healthy Net New MRR

A typical weekly review would have moved past this. The drilldown stopped on it.

Headline finding

$454K

Contraction hiding behind a healthy Net New MRR headline

A typical weekly review opens like this. Someone shares a slide. The first line reads:

Net New MRR: $413,796. Growing steady.

The CEO nods. The board deck gets one more checkmark. The meeting moves on to the next slide.

Then someone clicks into the KPI. The breakdown appears. The number that should have been the lead is not the headline.

ComponentAmountWhat it means
Contraction MRR$454,412Customers staying on but spending less
New MRR$413,796New customer revenue
Expansion MRR$197,121Upgrade revenue from existing customers
Churned MRR-$56,598Cancellation activity
Contraction MRR$454,412downgrades
New MRR$413,796headline metric
Expansion MRR$197,121upgrades
Churned MRR-$56,598cancellations
Net New MRR's four components, sorted by absolute impact. Contraction is the largest single force on the business this period.

Contraction is bigger than acquisition. The largest single force on the business this period is not new logos. It is existing customers downgrading. The dashboard had this data. It just never put it on the same page as the headline number.

What a SaaS metric map should surface by default

Every SaaS company has roughly the same metric vocabulary. Net New MRR breaks into four standard components: New, Expansion, Contraction, Churned. This is what a SaaS metric map looks like. It is the canonical structure of the business, regardless of which specific tables a particular company stores its subscriptions in.

A dashboard built from scratch will display Net New MRR as one number and put the four components on a separate tab three clicks away. That is the default dashboard. The MindPalace default is different. Open the parent KPI, see the components on the same screen, sorted by absolute impact. Contraction landed at the top. New MRR, the metric the team had been chasing all quarter, came in second.

The next move was to drill into Contraction by plan tier. The concentration showed up immediately. Enterprise accounts were responsible for most of the downgrade dollars, and the concentration was in renewal-window months. Roughly 38 percent of Enterprise accounts coming up for renewal were taking a plan step down rather than churning or holding steady.

The question on the next page of the deck changed. It was no longer "how do we get Net New MRR to $500K next quarter." It became "why are 38 percent of Enterprise accounts downgrading at renewal, and which CSMs own those accounts."

That is a different operational priority.

How the system got there

Three pieces of MindPalace combined to produce the breakdown.

The Living Map encoded the SaaS metric map. Net New MRR was the parent. New, Expansion, Contraction, Churned were the children. That structure came from canonical SaaS research, not from this warehouse's schema. The vocabulary exists before the customer does.

Cartographer grounded each child to specific tables in this warehouse. New MRR mapped to fct_subscriptions filtered by type = 'new'. Contraction mapped to the same table filtered by type = 'downgrade'. The grounding was deterministic. No language model picked the columns.

KPI Drilldown ran the four child queries in parallel against Postgres, computed the totals, and returned the sorted breakdown in 15.5 seconds. No human authored any of the queries. No one asked for the breakdown. It was the default Act 2 of a Net New MRR drilldown.

The Living Map also proposed a fifth child, CAC. Cartographer found that the marketing-spend table has no foreign key to accounts. The engine's reachability check dropped CAC from the breakdown before any SQL ran. A dashboard built by hand would have shown CAC anyway, with per-account attribution that does not exist in the schema. The honest answer is "we cannot compute this for this grain." The engine said that quietly and moved on.

Why this matters

The dashboard tells you what happened. A SaaS metric map tells you which part of your business is moving and by how much. A typical weekly review would never have seen the Contraction number on the same page as the headline. The team would have spent another quarter chasing new logos while existing Enterprise customers quietly took a plan step down. We have written before about why most companies are not as data-driven as they claim. This is the operational version of that gap. A SaaS metric map closes it.

KPI Drilldown's default sort by absolute impact is what put Contraction at the top of the screen. Putting the breakdown on the same screen as the headline is not a feature. It is a default. The default is what gets seen in the meeting.

A note on the data

Nimbus is a representative B2B SaaS dataset built for product demos. The figures above are real outputs from the live engine running against that dataset on 2026-05-10. The numbers are not from a customer.

If you want to see your own SaaS metric map running against your warehouse, request a demo. The drilldown takes about fifteen seconds.