Instrumenting a marketplace motion means deciding, before the year starts, which questions your reporting has to answer and making sure every one of them resolves to a dataset rather than to somebody’s spreadsheet.
The pattern is familiar to anyone who has run a marketplace number. A target is agreed in December. In April someone asks how it is tracking, and the answer takes four days, three exports and a reconciliation that nobody fully trusts. By the time the number arrives, the quarter it described is over.
That is an instrumentation failure, not an analytics failure. The data existed the whole time; nothing was set up in advance to turn it into an answer. The fix is cheap in December and expensive in April.
What “instrumented” actually means
A motion is instrumented when every number in the plan has a defined source, an owner, and a report it appears on without manual assembly. That is the whole bar. Not a dashboard, not a warehouse — a defined path from a question to a dataset.
The test is blunt: pick any target in the plan and ask who produces the number, from where, and how long it takes. If the answer involves a person exporting several files and joining them by hand, the metric will be produced twice and then quietly abandoned.
The four questions instrumentation has to answer
Most marketplace motions need the same small set, and they are worth naming explicitly because each one resolves to different data.
How much did we transact, and through which channel? Revenue by marketplace, separated from direct. This is the headline number and usually the only one teams set up in advance.
What is in flight? Offers created, sent, and accepted — the leading indicator that tells you in February whether the quarter is going to land, rather than in April.
Who is actually entitled to what? Entitlement state per agreement. This is the number that catches provisioning gaps and renewal risk before a customer reports them.
What did partners influence? Referral records linked to the opportunities they touched. Without this you can report marketplace revenue but not answer whether the partner motion is working, which is the question the investment actually rests on.
The datasets behind them
Each question maps to a dataset, and naming that mapping in advance is most of the work. In Suger, billing_revenue_record carries revenue across billing channels, marketplace_offer covers public and private offers so the in-flight question is answerable, marketplace_entitlement tracks usage rights and access levels per customer, and cosell_referral carries referral and co-selling activity.
Those are joinable against identity_buyer, which is what turns a marketplace figure into a customer-level view. The mapping matters more than the tooling: once each target names its dataset, whether you read it in a dashboard, a warehouse or a CSV becomes an implementation detail rather than a blocker. Exporting marketplace data to your warehouse covers that side.
Usage needs its own treatment
If you sell usage-based, instrument metering separately and earlier. Usage is the one place where the reported number and the billed number can diverge without anyone noticing for months.
The reason is structural: usage records are validated, aggregated, and reported to each provider in that provider’s own pricing-model format, so a dimension that was defined loosely at launch produces records that reconcile awkwardly forever. Set up the comparison between metered usage and billed revenue as a standing report, not as an investigation you run when a number looks wrong. Designing metering dimensions covers the upstream half of this.
Where instrumentation usually breaks
Three failure modes account for most of it.
No baseline. A target set without last year’s comparable number cannot be assessed mid-year — you can see the value but not whether it is good. Capture the baseline in December, while the year is still assembled.
Attribution decided late. If nobody defines partner-sourced versus partner-influenced before the year begins, the definition gets argued about in Q3 with real money attached. Partner-sourced vs partner-influenced is worth settling in writing first.
Reports with no reader. A metric nobody consumes monthly stops being maintained, and then it is wrong exactly when someone finally opens it. Give every standing report a named reader and a cadence, and delete the ones that fail that test.
Frequently asked questions
What does it mean to instrument a marketplace motion? Every number in the plan has a defined data source, a named owner, and a report it appears on without manual assembly. If producing a metric requires joining exports by hand, it is not instrumented.
What should I instrument first? Revenue by channel, offers in flight, entitlement state, and partner-influenced pipeline. Those four answer most operating questions, and each resolves to a different dataset.
Why instrument before the year starts? Because the baseline and the definitions are only cheap to establish while the prior year is still assembled. Reconstructing them mid-year costs days and produces numbers people distrust.
What is different about usage-based deals? Metered usage and billed revenue can diverge silently. Set up a standing comparison between them rather than investigating when a figure looks wrong, and define metering dimensions carefully upfront.
Why do reports stop being accurate? Because nobody reads them. A metric without a monthly reader stops being maintained, and it is discovered to be wrong at the moment it finally matters. Give each report a named reader or retire it.
Takeaways
- Instrumented means every target names its dataset, its owner, and the report it appears on — no manual assembly.
- Four questions cover most motions: revenue by channel, offers in flight, entitlement state, and partner influence.
- Map each question to a dataset in advance; the tooling you read it through is then an implementation detail.
- Treat usage separately and earlier — metered and billed figures diverge silently without a standing comparison.
- Capture the baseline and settle attribution definitions in December, before real money is attached to the argument.
Put the datasets behind your targets in one place: see Suger’s reporting and billing and metering, or book a demo.
Sources
Primary sources for the platform rules cited above. Last verified August 27, 2026. Cloud providers change fees, eligibility, and program terms without notice — check the source before relying on a figure.
- Suger docs: Available datasets — The billing_revenue_record, marketplace_offer, marketplace_entitlement and cosell_referral datasets named in this post.
- Suger docs: Metering — How usage records are validated and aggregated before they reach billing — the basis for reconciling usage against revenue.
Stay Updated
Get the latest Cloud GTM insights, product updates, and marketplace strategies delivered to your inbox.