Three reports, three answers, one 30-day window

They are not measuring the same thing, so they should not agree

• Email insights counts activity on the message: opens (total, inflated by image pre-fetching and privacy-driven automatic loading), clicks (total, includes repeat clicks by the same person) and unique clicks (per recipient). The 1,860 versus 640 gap is normal - it is the same people clicking several links.
• Journey insights counts contacts moving through a journey, so it excludes anyone who received the email outside the journey and anyone whose journey instance ended before the window closed. 9,800 opens against 12,400 is a plausible difference in scope, not an error - but you must confirm it is scope and not a refresh lag before you say so.
• Web analytics counts sessions, not people, and only those that survived consent banners, cookie blockers, and redirects through the tracking domain. It will always undercount clicks. 38 demo requests from 1,200 sessions is a site-side conversion, not a journey goal.

What to quote

• Deliverability and creative quality: email insights (bounce, complaint, unique click rate).
• Journey effectiveness comparisons: journey insights, used relatively between journeys in the same window, never as an absolute count of reality.
• Business outcome: the CRM. Demos booked is the only number the CMO should fund against.

The 0% goal for four days

Before blaming the journey, check three things: whether the goal is bound to a form or action that changed four days ago; whether goal completions require the contact to complete the goal inside the journey's time window; and whether journey insights have simply not refreshed. A metric that goes to exactly zero is nearly always a definition or wiring change, not a behavioural collapse.

Step 1 - Remove the contamination before measuring anything.

With two journeys mailing overlapping segments, a demo cannot be attributed to a journey because the person was exposed to both. Either consolidate the overlapping journeys, or enforce one-journey-at-a-time for a contact and hold that rule for the whole measurement period. Any cost-per-demo figure produced before this is decorative.

Step 2 - Carry the source through to the outcome.

Put a journey identifier into the tracked link as a UTM or custom parameter, capture it on the demo booking form, and persist it to the lead/opportunity record. That turns the CRM into the attribution source of truth and gives you a field you can group on. Do the same for a journey-name field stamped on the contact at entry as a backstop for people who arrive without a parameter.

Step 3 - Define the attribution rule and write it down.

Last-touch on the booking, with a stated lookback window, is the defensible default for a small team. Publish the rule alongside the number; a changing rule is how marketing reports lose credibility.

Step 4 - Calculate against real cost.

Cost per demo = (media spend + platform and licence allocation + list/event cost) / demos attributed. Include the fixed licence allocation or the comparison between journeys is meaningless.

Step 5 - Prove causality, do not infer it.

Run a holdout: 10% of each journey's eligible audience suppressed, measured over the same window. The difference between the two groups is the journey's actual contribution, and it is the only number that survives a CFO's second question.

What I would tell her

Not yet a funding decision off the current numbers. What the data does support today:

• The three journeys are not separable in the reporting because two of them mail the same people. That is a measurement defect with a one-week fix, and it is the highest-value action available.
• Email-level creative quality is measurable now: unique click rate per message is the honest engagement signal, and it is the lever that moves every journey at once.
• The goal metric has been broken for four days, so any journey-level conversion comparison drawn from it is invalid until it is repaired and back-filled or the window is re-cut.

What I would refuse to conclude

• Which journey 'performed best', on the strength of overlapping audiences and a metric that hit zero.
• That cost per demo reflects the journey rather than the list. A journey mailed to the most engaged segment will win on every metric regardless of design quality; without a holdout you are ranking audiences and calling it journey performance.

Plan for the quarter

Week 1: fix the goal definition, verify refresh, de-overlap the journeys, instrument the source parameter through to CRM. Week 2-6: run holdouts. Week 7: present cost per demo with the attribution rule and holdout deltas attached. That gets the CMO a fundable answer in seven weeks instead of a confident wrong one this afternoon - and the deliverability and creative work can start immediately in parallel because it does not depend on the measurement fix.