FICTIONAL EXAMPLE · NOT CLIENT WORK · ALL COMPANIES, SYSTEMS, AND FINDINGS ARE INVENTED
Representative deliverable

Revenue Data Blueprint

A fictional example showing the level of clarity, prioritization, and technical detail included in a real engagement.

CompanyNorthstar Software (fictional)
Business modelB2B SaaS
Primary questionWhich channels create revenue?
Delivery windowFive business days
01 · Executive summary

The decision Northstar cannot make today

“Should we move budget from paid search to partner marketing next quarter?”

The company cannot answer this reliably because campaign cost, lead source, opportunity, and closed-revenue data use different identifiers and attribution rules. Three executive reports currently produce three versions of customer acquisition cost.

Primary risk No shared account key

Marketing activity is stored at contact level while revenue is stored at account and opportunity level.

Decision risk Conflicting CAC logic

Finance includes salaries and tools; marketing includes media spend only; the board sheet mixes both.

Opportunity No new warehouse required yet

The current BigQuery environment can support a reliable first version without a platform migration.

02 · Source and reporting inventory

What currently feeds the decision

The real Blueprint records ownership, refresh behavior, access, identifiers, and known quality issues for each source. This shortened example shows the decision-relevant subset.

SOURCEOWNERKEY DATAREFRESHOBSERVATION
HubSpotRevenue OperationsContacts, companies, opportunitiesLiveOriginal source is overwritten during imports
Google AdsDemand GenerationCampaign, cost, clicks, conversionsLiveUTM naming changed twice this year
Partner spreadsheetPartnershipsReferrals, partner name, submitted dateWeeklyNo stable contact or account identifier
StripeFinanceCustomer, invoice, recurring revenueDailyCustomer ID not written back to HubSpot
Board reporting sheetFinanceCAC, pipeline, new ARRMonthlyManual copy/paste with undocumented adjustments
03 · Metric conflict register

Why the dashboards disagree

METRICMARKETINGFINANCERECOMMENDED DEFINITIONRISK
Customer acquisition costMedia spend ÷ new customersSales + marketing cost ÷ new customersPublish both as Paid CAC and Fully Loaded CACHIGH
Sourced pipelineFirst-touch campaignOpportunity source fieldSeparate sourced, influenced, and self-reported viewsHIGH
New ARRDeal amountNormalized subscription valueUse finance-normalized annual recurring valueMEDIUM
LeadAny form submissionNot usedReport inquiry, MQL, and accepted lead separatelyMEDIUM
04 · Recommended architecture

The smallest reliable system

Northstar already has BigQuery. The recommendation is to repair identity and metric logic inside the current environment—not introduce Snowflake, a new BI tool, or an attribution vendor.

Business systemsHubSpotGoogle AdsPartner sheetStripe
BigQuery + dbtAccount identity mapCampaign taxonomyTested revenue modelsMetric definitions
Decision layerChannel-to-pipeline viewPaid and fully loaded CACExecutive scorecardData-quality alerts
Recommendation: keep the current tools, create one stable account identity layer, and publish governed revenue metrics before attempting algorithmic multi-touch attribution.
05 · Prioritized roadmap

What to do over the next 90 days

DAYS 01–30 · TRUST

Reconcile the foundation

  • Approve the metric dictionary
  • Create the account identity map
  • Preserve original source fields
  • Standardize campaign naming
DAYS 31–60 · VISIBILITY

Ship decision reporting

  • Build tested revenue models
  • Publish sourced and influenced pipeline
  • Separate paid and fully loaded CAC
  • Add freshness and quality alerts
DAYS 61–90 · LEARNING

Improve attribution

  • Capture self-reported attribution
  • Measure partner influence
  • Compare rule-based models
  • Document budget experiments
06 · Implementation boundary

What the proposed build would—and would not—include

Included

  • HubSpot, Google Ads, partner, and Stripe models
  • Account identity resolution rules
  • Governed metric definitions and tests
  • Executive and channel reporting
  • Documentation, training, and handoff

Not included

  • Campaign strategy or media management
  • Migration away from BigQuery
  • Machine-learning attribution
  • Historical data repair before the agreed date
  • Ongoing support unless separately requested

Your Blueprint will reflect your systems and decisions

This fictional sample is intentionally abbreviated. A real engagement includes the complete inventory, evidence behind each finding, implementation assumptions, and a fixed-price build proposal.

Discuss your data