Who became a new customer?
Define new versus repeat customers, deduplicate records, and compare the same periods. A count of new contacts is not automatically customer growth.
REVENUE ATTRIBUTIONFollow the journey.
Follow the journey.
Explain the credit.
Connect identifiable marketing touchpoints to contact and deal records. Compare model assumptions and keep unattributed outcomes visible instead of forcing every sale into a channel.
INTERACTIVE EXPLAINER
One customer. Several marketing touchpoints.
Organic search
Discovers your businessSocial post
Tracked campaign clickWhatsApp enquiry
Identity confirmed with permissionCRM deal
Won value recordedAccounting payment
Planned reconciliation
Discovers your businessSocial post
Tracked campaign clickWhatsApp enquiry
Identity confirmed with permissionCRM deal
Won value recordedAccounting payment
Planned reconciliation
Illustrative S$1,200 won deal; synthetic data. The production engine supports five rules-based models. This demonstration is not a live customer report.
MEASURE THE RIGHT OUTCOMECUSTOMER GROWTH REVENUE SOURCE ATTRIBUTION CREDIT
Growth, revenue and attribution are different questions.
What amount was recorded?
Separate won deals, billed amounts and collected payments. Refunds, currency, taxes and recurring revenue need explicit treatment. Today’s engine uses recorded deal values.
Which observed touches receive credit?
Credit depends on identifiers, lookback windows, coverage and model choice. It does not establish causation; incremental lift needs a separate experiment.
IMPORTANT BOUNDARIES
No tool can see every marketing influence.
Consent restrictions, private DMs, anonymous search, offline referrals and missing CRM records leave gaps. Google Search Console data is aggregate; it does not identify an individual searcher. WhatsApp identity does not reveal an earlier campaign unless a reliable link is available.
Deterministic, permissioned identity matching is the intended default. Server-side tracking must respect user choices; it must not bypass consent or privacy protections. Markov and Shapley models are planned and would still allocate credit—not prove “true” contribution.
MAKE MARKETING ACCOUNTABLE
Stop guessing. Start connecting the evidence.
Bring one client, one campaign and a measurable business outcome. Start with the data you can verify—not another promise of perfect attribution.