Build a marketing reporting dashboard people can trust.

A marketing reporting dashboard should answer a defined set of commercial questions using governed metrics. For each KPI, document its business definition, calculation, source, time basis, inclusion rules, owner, refresh cadence and known limitation. Then show whether the data is complete and current.

Google explains that attribution assigns credit across touchpoints using a selected model, while Shopify documents why analytics tools can disagree. Use Google's attribution guidance and Shopify's discrepancy guidance when governing platform comparisons.

Start with decisions, not data availability

List the recurring decisions made in the weekly or monthly trading meeting: budget reallocation, campaign continuation, inventory support, retention action or creative testing. The report should make those decisions easier and more consistent.

If a measure does not change a decision or explain an outcome, remove it from the primary view. It can remain available for investigation without consuming leadership attention every week.

Create a KPI dictionary

A name is not a definition. ‘Revenue’, ‘new customer’ and ‘return on ad spend’ can differ by source, time zone, attribution setting, returns treatment and tax treatment. Finance and marketing should agree which definition answers which question. Clerq does not invent accounting or tax logic.

• KPI name and plain-English business question
• Calculation and inclusion or exclusion rules
• Currency, time zone and reporting window
• Source of record and source-system field
• Attribution or cohort basis where relevant
• Refresh cadence, data owner and approval owner
• Known limitations and change history

Build the reporting control layer

Before a report refreshes, check whether every required source arrived, whether the extraction is current and whether key totals are within approved reconciliation thresholds. Make failed checks visible to the reader.

• Freshness check for every source
• Completeness check for expected periods or entities
• Duplicate and schema-change detection
• Reconciliation to approved commercial or finance totals
• Versioned metric definitions and change approval
• Named owner for investigation and release

What can be automated?

Collection, standardisation, approved calculations, quality checks, pack assembly and first-draft variance commentary can be automated. The system should retain source references so a reviewer can inspect the figure.

People should validate causes, add commercial context, approve important definition changes and make allocation decisions. Generated commentary should not turn correlation into a causal claim.

A practical dashboard hierarchy

Use three levels. The executive view shows outcomes and decisions. The channel or journey view explains performance. The diagnostic view exposes granular data for investigation. This keeps leadership reporting concise without hiding the evidence analysts need.

• Outcome: revenue, contribution or another approved commercial result
• Demand: qualified traffic, leads or customer demand by governed definition
• Efficiency: cost per approved outcome and budget utilisation
• Customer: acquisition, repeat behaviour or cohort measures
• Delivery: campaign, channel and creative drivers
• Data quality: freshness, completeness and unresolved exceptions

Measure the reporting process itself

Track preparation time, late or failed sources, reconciliation differences, definition disputes, post-publication corrections and time from meeting question to supported answer. A beautiful dashboard can still be an expensive manual process.

Frequently asked questions