Commercial analytics and revenue operations for AI companies
Marketing may count submissions, sales may count opportunities, product may count active users and finance may count invoices. Each can be correct in its own system while management still lacks a reliable view of how customers are acquired and served.
Discuss your AI-company project
Webvelopment helps define the relationships between those records and implement the agreed reporting and workflows. This is a scoped data and operations service, not an automatic promise to replace your finance function or run every revenue process.
Define the customer before calculating the cost
A person, seat, account, organisation and subscription are different units. Decide which unit represents a customer in the report, and maintain that definition through acquisition, activation and billing.
Then define the stages that matter. A demo request may become a booking, an attended meeting, an accepted opportunity and a customer. A trial may become an activated account and then a paid subscription. A developer may create an API key long before an organisation uses the product in production.
The system should preserve those distinctions instead of adding every event into a single “conversion” number.
What the work can include
Lifecycle and metric definitions
Create a dictionary with the event, denominator, owner, timestamp and evidence source for each reported stage. Define what is observed, estimated, excluded or unknown.
CRM and commercial workflow configuration
Implement agreed assignment, follow-up, pipeline and handover rules. Data quality and user adoption matter alongside the configuration itself.
Product and billing reconciliation
Connect approved account-level events to commercial records. Keep recurring subscription definitions separate from usage, invoices, collections and accounting revenue where they differ.
Acquisition and cohort reporting
Report media and other acquisition costs with the appropriate denominator. Follow cohorts long enough to observe later outcomes, rather than crediting every new payment to the most recent campaign.
Dashboards and operational reviews
Build views that support specific decisions, with visible definitions and reconciliation checks. A dashboard should show missing data and exceptions rather than hide them behind a polished total.
Revenue is not a single interchangeable number
Committed monthly subscription revenue, variable usage charges, invoiced amounts and collected cash answer different questions. They should be labelled accordingly.
For example, Stripe’s published MRR convention excludes variable metered billing. That is a provider-specific reporting rule, not a reason to assume every business uses the same definition. Your report needs an explicit policy that finance and the operating team understand. Stripe’s MRR explanation
Forecast usage revenue should not be presented as contracted recurring revenue merely because it appears in the same subscription account.
AI delivery costs belong in the commercial conversation
Acquisition can look efficient until inference, infrastructure, support and payment costs are considered. Those costs do not all belong in customer acquisition cost, but they affect the contribution available to recover it.
We distinguish the acquisition cost numerator from the monthly delivery-cost calculation. The result is a more useful conversation about payback and customer value than dividing advertising spend by revenue.
The AI SaaS acquisition calculator illustrates that distinction. Its examples are scenarios, not benchmarks, accounting advice or a forecast of your business.
Design for decisions, not just reports
A useful weekly review might ask whether enquiry quality changed, whether sales response slowed, whether activation fell after an onboarding change or whether high-usage customers contribute less than expected.
Each question needs an owner and a possible action. The report should not imply that marketing can fix a product defect or that an integration can resolve an unclear pricing model by itself.
Data minimisation is part of the design
Commercial reporting usually needs event and account identifiers, timestamps, stages and approved values. It does not need unrestricted access to confidential prompts, model outputs or customer documents.
Define data purpose, access, retention and export before connecting systems. Limit analytics payloads to approved fields. A hashing step does not by itself make an inappropriate data transfer acceptable.
What we need from your team
Access to current definitions and safe sample records is a useful start. The sales, product and finance owners need to approve what the report means. Relevant API access and platform permissions must be available for implementation.
Where the current data cannot support a metric, the honest output is a gap and a collection plan. It is not a fabricated historical result or a silently substituted denominator.
How the project is accepted
Acceptance should include approved definitions, documented mappings, reconciliation against source systems, tested edge cases, access checks and operator training. Reports should distinguish missing values from zero and estimates from observed records.
Ongoing work may include data-quality checks, workflow improvements, new reports and periodic reviews. Larger enterprise process redesign or platform-specialist work requires explicit scope and appropriate expertise.
A worked example: ten seats, one customer
A company buys ten seats. In the product database that’s ten users; in the CRM it might be ten contacts; in billing it’s one subscription. If the acquisition report counts users, CAC looks ten times better than it is. The same happens when forecast usage quietly becomes “committed” MRR. RevOps work fixes the unit first (organisation, account or seat) and keeps it consistent from the first campaign click to the invoice.
What you receive
A metric dictionary with each stage’s denominator, timestamp, owner and source; an account and event identity map; lifecycle and assignment configuration within the agreed scope; source-to-report lineage and reconciliation rules; cohort-aware acquisition reporting; separate views of recurring, usage, invoiced and collected revenue; and a dashboard with an exception queue, training and a review routine.
Why we take measurement personally
We run our own billing, CRM and project management on Clienthub, a system we’ve built and refined in-house for 13 years, so we know the daily cost of a definition that drifts. We’ve also built reporting products for others, such as the Suzuki secret shopper platform, with weighted scoring, branch comparisons and PDF reports. Different data, same discipline: agreed definitions, controlled access and numbers you can trace back to their source.
Questions about revenue operations
A few useful answers before you decide.
Do we need to replace our CRM?
Can you calculate lifetime value?
Is a dashboard enough to improve acquisition?
Discuss a measurement foundation · Read the acquisition-cost guide
Keep the right things connected.
Web development and integrations around your AI product
Connect the business systems around the product without confusing them with the product itself.
Paid acquisition for AI and SaaS companies
Use paid media to test a commercial hypothesis, not just buy activity.
AI SaaS acquisition and contribution calculator
Explore sales-led, product-led and developer-led acquisition assumptions, with separate CAC definitions, delivery costs and retained-cohort recovery.
AI SaaS customer acquisition cost: the formula is only the beginning
Calculate media, marketing and fully loaded CAC for AI SaaS, then separate recurring revenue, usage, delivery costs and cohort payback.
Start a conversation
Tell us what you are working towards. We’ll help you find a practical next step.