Calculation rules

Every number has to earn its place on the page.

See how SIRJ defines ad spend, attributed revenue, ROAS, leads, cost per lead, campaign performance, form conversion, pipeline, and unsupported metrics.

Open live demo See data routing

Buyer answers

The calculation rules buyers need before they trust the report.

These answers explain when dashboard cards unlock, how ROAS and computed Ask Data answers are produced, and how client reports reuse the same source-backed calculation layer.

Public answer

How does SIRJ decide which numbers can appear?

Dashboard cards unlock from source evidence and provider category. Spend records unlock ad spend and campaign performance, revenue records unlock revenue, spend plus revenue unlock ROAS, lead records unlock lead source rows, form sources unlock form conversion, and CRM sources unlock pipeline.

The buyer sees the current proof state instead of a full-looking dashboard that hides missing sources. If the source cannot support the surface, the surface does not pretend to be complete.

Follow the proof path

Public answer

How is ROAS calculated?

ROAS is attributed revenue divided by total ad spend for the active client and date range. The value remains null when spend is zero or missing, because a divided number without spend evidence would be a confidence trick instead of a calculation.

ROAS only appears when both sides of the equation exist. That makes a missing ROAS card a useful setup signal, not a broken marketing page.

Follow the proof path

Public answer

How does Ask Data calculate computed answers?

Ask Data validates query specs against allowed fields before it calculates cost per lead, source performance, spend versus revenue, campaign budget recommendations, or uncontacted leads. Unsupported or insufficient-data questions return a constrained response instead of invented certainty.

Plain-English answers still have a calculation contract. The assistant can cite fields, rows, and filters when evidence exists, and it can refuse when the connected stack is not enough.

Follow the proof path

Public answer

How do reports reuse dashboard calculations?

Report snapshots extract dashboard totals, source rows, campaign rows, top performers, insights, and recommendations from the same dashboard configuration. Recommendations only appear when the supporting spend, lead, revenue, or attention evidence exists.

A client report is not separate sales copy. It freezes the same calculation layer the dashboard and Ask Data use, then shows gaps when source evidence is incomplete.

Follow the proof path

Evidence map

Each number names the source evidence it needs.

The definitions below connect SIRJ dashboard cards, Ask Data answers, campaign rows, form conversion, pipeline, and report snapshots to the source evidence required before a claim appears.

Ad spend Attributed revenue ROAS Leads by source Cost per lead

Calculation map

Definitions, required evidence, and user-visible surfaces.

A number becomes useful only when the buyer can see what it means, what source evidence it needs, and where the product is allowed to show it.

Calculation	Required evidence	Where users see it
Ad spend	At least one normalized spend event from an ad source, scoped to the active client and selected date range.The sum of spend event amount cents in the active account, client workspace, and date range, with clicks and impressions carried from numeric source fields when providers supply them.	Ad spend card, spend by source rows, campaign performance, source performance answers, and client reports.
Attributed revenue	Revenue-source records such as deals, invoices, payments, or booking outcomes plus attribution matches inside the same tenant scope.Revenue from deal or payment events that attribution matching connects back to supported demand paths, summarized for the active dashboard scope.	Revenue card, ROAS card, report summary, campaign rows, and cited Ask Data answers.
ROAS	Both spend evidence and revenue evidence in the same scoped dashboard context, for the active client and selected date range.Attributed revenue divided by total ad spend for the scoped period. The value remains null when spend is zero or missing.	ROAS card, ROAS by channel, campaign performance, source performance answers, and reports.
Leads by source	Normalized lead events from forms, call tracking, CRM, booking, or other lead-producing providers.Lead events grouped by normalized source fields such as utmSource, source, or provider id, then counted inside the selected scope.	Lead source tables, dashboard readiness, cost-per-lead calculations, and source-specific Ask Data answers.
Cost per lead	Spend records and lead records for the dashboard scope, plus matching source evidence for source-specific questions.Total spend divided by lead count. When the question names a source, Ask Data filters spend and lead events to that normalized source before dividing.	Ask Data cost-per-lead answers that cite spend, lead count, source filters, and the scoped fields used to divide spend by leads.
Campaign performance	Campaign spend rows, normalized campaign names, lead matches, and revenue matches when revenue performance is being shown.Spend events grouped by campaign, with leads from spend-to-lead matches, attributed revenue from matched downstream revenue, and ROAS from the supported spend and revenue fields.	Campaign performance table, budget recommendation answers, and report campaign rows.
Form conversion	Form lead submissions and measured form-start or view counts when a conversion rate is displayed.Form submissions divided by measured starts or views when those start fields exist. Without start evidence, the row stays in submissions-only status instead of implying a rate.	Form conversion table rows, submissions-only states, dashboard readiness, and setup guidance for form sources that lack measured starts.
Pipeline	CRM, booking, or deal-source events with status and amount fields available to the workspace.Deal events grouped by normalized stage or status, with count and value cents summed for each stage in the active scope.	Pipeline table, report metrics, closed-deal counts, closed-deal value, weekly insights, and client report summaries.
Uncontacted leads	Lead events and communication-source events in the same client workspace, with contact identity available for matching.Lead records without a later matching communication event for the same contact identity, returned as bounded actionable rows rather than a cached rollup total.	Ask Data follow-up answers and operational rows for teams reviewing lead leakage.

Validation workflow

From scoped records to a report-ready calculation.

The workflow keeps dashboard cards, Ask Data, and report exports on the same source-backed foundation.

Step	Calculation role	User-visible proof
Scope the records	Every calculation starts from the active account, client workspace, and date range. Connections, events, matches, notifications, and report snapshots must pass tenant-scope checks before they can influence a number.	Users see one client workspace and one selected period, not a blended account-wide total masquerading as client proof.
Unlock eligible cards	Dashboard generation checks normalized event types and provider categories before adding cards for spend, revenue, ROAS, lead sources, form conversion, campaign performance, or pipeline.	Missing cards communicate missing evidence, while available cards explain what the connected stack can support right now.
Compute from explicit fields	Each row uses named source fields, grouped events, attribution matches, safe division, and null states rather than prose-level assumptions about what a campaign probably produced.	Spend, revenue, ROAS, CPL, source performance, form conversion, and pipeline rows can be traced to a formula and source requirement.
Cite or refuse Ask Data answers	Ask Data uses validated query specs, allowed card fields, source filters, date ranges, and computed query handlers before returning cited metrics, rows, charts, or insufficient-data responses.	The answer names the fields it used or explains which evidence is missing instead of making a confident unsupported claim.
Freeze report snapshots	Reports extract the same dashboard metrics, source breakdowns, campaign rows, insights, and recommendation rules into a snapshot before export or weekly delivery.	A client-ready report can be audited back to the dashboard state and source rules that existed when the report was generated.

Proof standards

What prevents unsupported numbers from reaching the buyer.

These safeguards keep formulas, scopes, denominators, and AI-assisted answers from outrunning the evidence.

Proof standardCard eligibility

A dashboard can overclaim when it renders revenue, ROAS, pipeline, form conversion, or source rows before the required connector evidence exists.

Card generation reads normalized event types and provider categories, then leaves unsupported cards absent or incomplete instead of filling them with placeholder confidence.

Proof standardDate and tenant scope

A correct formula becomes wrong when it mixes clients, accounts, or date ranges that the user did not ask to evaluate.

Dashboard config and Ask Data query specs carry account id, client id, and date range through every calculation, cache key, answer, and report snapshot.

Proof standardSafe division and null states

ROAS, conversion rate, and cost per lead can become misleading when the denominator is zero, missing, or not measured by the source.

ROAS stays null when spend is absent, form conversion marks submissions-only when starts are missing, and Ask Data returns insufficient data when required evidence is unavailable.

Proof standardUnsupported-answer restraint

An AI-assisted answer can damage trust by sounding certain about a number the connected stack cannot prove.

Unsupported Ask Data questions return the supported areas and avoid reading dashboard data, caching a fake answer, or turning missing evidence into sales copy.

FAQ

Questions that come up when a client asks where the number came from.

The answers are short on purpose: missing source evidence should be easy to explain in a client meeting.

FAQ

Why are some numbers hidden?

Numbers stay hidden or null when the connected stack cannot support them yet. For example, ROAS requires both spend and revenue, while form conversion needs measured starts before SIRJ can show a rate.

FAQ

What does ROAS use as revenue?

ROAS uses attributed revenue from downstream deal or payment evidence that can be matched back to the demand path. It is not lead value, projected value, or a blended channel estimate.

FAQ

Can cost per lead be source-specific?

Yes. Ask Data can filter spend and lead events by a normalized source such as Google Ads or Facebook before dividing spend by lead count, when both sides exist for that source.

FAQ

What happens when a report has spend but no revenue?

The report should treat that as a setup or operational signal. It can recommend connecting or verifying a revenue source instead of pretending spend alone proves ROI.

Ready when the numbers need to hold up

Now inspect the calculation layer in context.

Open the live demo to see dashboard cards, Ask Data answers, and report-ready proof using the same evidence rules.

Open live demo See data routing

How does SIRJ decide which numbers can appear?

The buyer sees the current proof state instead of a full-looking dashboard that hides missing sources. If the source cannot support the surface, the surface does not pretend to be complete.

Follow the proof path

How is ROAS calculated?

ROAS only appears when both sides of the equation exist. That makes a missing ROAS card a useful setup signal, not a broken marketing page.

Follow the proof path

How does Ask Data calculate computed answers?

Plain-English answers still have a calculation contract. The assistant can cite fields, rows, and filters when evidence exists, and it can refuse when the connected stack is not enough.

Follow the proof path

How do reports reuse dashboard calculations?

A client report is not separate sales copy. It freezes the same calculation layer the dashboard and Ask Data use, then shows gaps when source evidence is incomplete.

Follow the proof path

Calculation

Required evidence

Where users see it

Ad spend

At least one normalized spend event from an ad source, scoped to the active client and selected date range.The sum of spend event amount cents in the active account, client workspace, and date range, with clicks and impressions carried from numeric source fields when providers supply them.

Ad spend card, spend by source rows, campaign performance, source performance answers, and client reports.

Attributed revenue

Revenue-source records such as deals, invoices, payments, or booking outcomes plus attribution matches inside the same tenant scope.Revenue from deal or payment events that attribution matching connects back to supported demand paths, summarized for the active dashboard scope.

Revenue card, ROAS card, report summary, campaign rows, and cited Ask Data answers.

ROAS

Both spend evidence and revenue evidence in the same scoped dashboard context, for the active client and selected date range.Attributed revenue divided by total ad spend for the scoped period. The value remains null when spend is zero or missing.

ROAS card, ROAS by channel, campaign performance, source performance answers, and reports.

Leads by source

Normalized lead events from forms, call tracking, CRM, booking, or other lead-producing providers.Lead events grouped by normalized source fields such as utmSource, source, or provider id, then counted inside the selected scope.

Lead source tables, dashboard readiness, cost-per-lead calculations, and source-specific Ask Data answers.

Cost per lead

Spend records and lead records for the dashboard scope, plus matching source evidence for source-specific questions.Total spend divided by lead count. When the question names a source, Ask Data filters spend and lead events to that normalized source before dividing.

Ask Data cost-per-lead answers that cite spend, lead count, source filters, and the scoped fields used to divide spend by leads.

Campaign performance

Campaign spend rows, normalized campaign names, lead matches, and revenue matches when revenue performance is being shown.Spend events grouped by campaign, with leads from spend-to-lead matches, attributed revenue from matched downstream revenue, and ROAS from the supported spend and revenue fields.

Campaign performance table, budget recommendation answers, and report campaign rows.

Form conversion

Form lead submissions and measured form-start or view counts when a conversion rate is displayed.Form submissions divided by measured starts or views when those start fields exist. Without start evidence, the row stays in submissions-only status instead of implying a rate.

Form conversion table rows, submissions-only states, dashboard readiness, and setup guidance for form sources that lack measured starts.

Pipeline

CRM, booking, or deal-source events with status and amount fields available to the workspace.Deal events grouped by normalized stage or status, with count and value cents summed for each stage in the active scope.

Pipeline table, report metrics, closed-deal counts, closed-deal value, weekly insights, and client report summaries.

Uncontacted leads

Lead events and communication-source events in the same client workspace, with contact identity available for matching.Lead records without a later matching communication event for the same contact identity, returned as bounded actionable rows rather than a cached rollup total.

Ask Data follow-up answers and operational rows for teams reviewing lead leakage.

Step

Calculation role

User-visible proof

Scope the records

Every calculation starts from the active account, client workspace, and date range. Connections, events, matches, notifications, and report snapshots must pass tenant-scope checks before they can influence a number.

Users see one client workspace and one selected period, not a blended account-wide total masquerading as client proof.

Unlock eligible cards

Dashboard generation checks normalized event types and provider categories before adding cards for spend, revenue, ROAS, lead sources, form conversion, campaign performance, or pipeline.

Missing cards communicate missing evidence, while available cards explain what the connected stack can support right now.

Compute from explicit fields

Each row uses named source fields, grouped events, attribution matches, safe division, and null states rather than prose-level assumptions about what a campaign probably produced.

Spend, revenue, ROAS, CPL, source performance, form conversion, and pipeline rows can be traced to a formula and source requirement.

Cite or refuse Ask Data answers

Ask Data uses validated query specs, allowed card fields, source filters, date ranges, and computed query handlers before returning cited metrics, rows, charts, or insufficient-data responses.

The answer names the fields it used or explains which evidence is missing instead of making a confident unsupported claim.

Freeze report snapshots

Reports extract the same dashboard metrics, source breakdowns, campaign rows, insights, and recommendation rules into a snapshot before export or weekly delivery.

A client-ready report can be audited back to the dashboard state and source rules that existed when the report was generated.

Every number has to earn its place on the page.

The calculation rules buyers need before they trust the report.

How does SIRJ decide which numbers can appear?

How is ROAS calculated?

How does Ask Data calculate computed answers?

How do reports reuse dashboard calculations?

Each number names the source evidence it needs.

Definitions, required evidence, and user-visible surfaces.

From scoped records to a report-ready calculation.

What prevents unsupported numbers from reaching the buyer.

A dashboard can overclaim when it renders revenue, ROAS, pipeline, form conversion, or source rows before the required connector evidence exists.

A correct formula becomes wrong when it mixes clients, accounts, or date ranges that the user did not ask to evaluate.

ROAS, conversion rate, and cost per lead can become misleading when the denominator is zero, missing, or not measured by the source.

An AI-assisted answer can damage trust by sounding certain about a number the connected stack cannot prove.

Questions that come up when a client asks where the number came from.

Why are some numbers hidden?

What does ROAS use as revenue?

Can cost per lead be source-specific?

What happens when a report has spend but no revenue?

Now inspect the calculation layer in context.

Preparing the source route.

Every number has to earn its place on the page.

The calculation rules buyers need before they trust the report.

How does SIRJ decide which numbers can appear?

How is ROAS calculated?

How does Ask Data calculate computed answers?

How do reports reuse dashboard calculations?

Each number names the source evidence it needs.

Definitions, required evidence, and user-visible surfaces.

From scoped records to a report-ready calculation.

What prevents unsupported numbers from reaching the buyer.

A dashboard can overclaim when it renders revenue, ROAS, pipeline, form conversion, or source rows before the required connector evidence exists.

A correct formula becomes wrong when it mixes clients, accounts, or date ranges that the user did not ask to evaluate.

ROAS, conversion rate, and cost per lead can become misleading when the denominator is zero, missing, or not measured by the source.

An AI-assisted answer can damage trust by sounding certain about a number the connected stack cannot prove.

Questions that come up when a client asks where the number came from.

Why are some numbers hidden?

What does ROAS use as revenue?

Can cost per lead be source-specific?

What happens when a report has spend but no revenue?

Now inspect the calculation layer in context.