Make the event definition explicit
Define page view, click, enquiry, qualified lead, opportunity and revenue in terms the business agrees on.
Marketing measurement should improve commercial decisions, not produce a contest between dashboards. We examine what can be observed, what cannot, and how much confidence the evidence really deserves.
The marketing meeting starts with a 40-slide report. Search, social and email all show conversions, but their totals exceed the number of customers. Sales mentions private referrals and WhatsApp conversations that never appeared in the dashboard.
The reporting system is mistaking visibility for complete influence. Last-click reports, platform attribution and private buyer behaviour are answering different questions.
The question worth askingWhat do we genuinely know about marketing contribution, and what are we assuming?
These are illustrative situations, not claims about a particular client. Select the closest one to explore an appropriate first response.
Web forms show success, but some enquiries never become usable records.
A sensible first moveTrace events, validation, delivery, deduplication and CRM hand-offs.
Discuss a situation like this →Attribution models point to different winners.
A sensible first moveSeparate observed touchpoints from causal contribution and explain model assumptions.
Discuss a situation like this →Acquisition grew while qualified progression weakened.
A sensible first moveBuild source-to-stage reporting and investigate changes in intent, quality and follow-up.
Discuss a situation like this →Define page view, click, enquiry, qualified lead, opportunity and revenue in terms the business agrees on.
Compare GA4, campaign platforms, consent-aware web measurement, CRM and sales evidence.
Use buyer feedback, assisted journeys and incrementality tests where feasible; avoid turning estimates into certainty.
No fixed package is presumed. The right mix depends on what we find when the problem is examined.
Consent-aware events, lead interactions, traffic sources and technical validation.
Multi-touch visibility, dark social, self-reported influence and the limitations of modelled credit.
Qualified stages, pipeline progression, deduplicated records and sales feedback.
Practical tests, confidence assessment and decision logs where data and budget permit.
These are possible deliverables, not a promise that every engagement needs all of them.
Not reliably. Attribution allocates observed credit according to rules; it is not automatically a causal estimate.
No. Visitors who decline optional analytics may not appear under a consent-first installation.
Buyer-reported sources, sales notes and carefully interpreted CRM evidence can fill some gaps without pretending the view is complete.
Bring the situation, even if the right solution is not yet clear. We can start by examining what deserves attention.