Q3 in Review
Sydney, Australia
Sam Russell, Product Lead

Last quarter ended with a platform that attributes, tracks, plans and explains, and this quarter went into giving that attribution more depth, starting with the structure itself. A group can now hold one attribution model for the whole engagement, each business unit can expand into a submodel that keeps its own showback rules, cost structure, and its total rolls up to the root, so a group controller sees the sum of the parts while a unit lead still works inside their own estate.
Once the structure holds a whole group, budgets have somewhere to cascade, an allocation set at the top becomes the envelope for the nodes beneath it, a budget linked to its parent cannot exceed it, a budget set up independently says so, and a node that falls out of line with the level above it is flagged, with the reason on drill-down. The nodes became the way in to everything else as well, each one carries its budgets, goals and unit economics and opens them from its own menu with the context carried across, so nobody has to re-select an entity on the way to the answer.

Showbacks carry shared cost on to the teams that use it, and Analytics now has a Spend Flow view that draws the movement. Pick an attribution node and the Sankey chart lays out what the node carries directly, the volume moving through it and what it carries in total, with a table built from the same grouping underneath, so the gap between direct spend and total is a shape on a page, easy to point at in a meeting.
Cost per customer is half a ratio, and the other half is what the customer pays, so revenue now lives in the attribution model beside the cost. Stripe revenue is spread over its service period and attributed to reporting entities by rules you write, an item, a customer list, a share apportioned down from a parent, each line carrying the rule that put it there, and every entity shows its technology margin, revenue less technology cost, with whatever is not yet attributed called out and a control total to reconcile against the ledger. Revenue unit economics sit on top, revenue and margin per customer, per GB stored, per 1,000 forecasts, on any entity.
Attribution only reaches as far as the spend it can see, so coverage kept widening. Azure now connects for both Microsoft Customer Agreement and Cloud Solution Provider accounts through a short guided setup, in beta, which brings the estates that buy through a partner into the same model as everything bought direct. More providers meant more currencies, and a total spanning AWS, Anthropic, GCP and Azure is only usable when the figures share one, so OPTIMAZE now holds a single account currency, set under Settings, and every financial figure in the platform is expressed in it. Each line converts once at ingest, at the daily market rate by default or at a fixed rate you set for any currency that needs one, the native bill is never rewritten, and the working, original amount, converted amount, rate applied, is there on drill-down.
A structure that holds a group, budgets that cascade through it, revenue beside the cost and figures in one currency are what make the data worth asking for from anywhere, which is what MCP is for. OPTIMAZE is now reachable through it, so an MCP client can query attributed cost data directly, without anyone needing to be in the platform at the time, and the first set of tools covers the Finance domains: cost and billing, budgets, goals and unit economics.
Most of this quarter sits underneath the screens, in the structure, the revenue, the coverage and the currency, and the reach on top of them follows from that. The changelog is where each piece is written up as it ships.
Subscribe to our blog.
Never more than once a week.
