The attribution gap the CFO doesn't believe
Cobalt Ridge, a cybersecurity vendor, spent $310,000 on LinkedIn in Q1. LinkedIn's dashboard reports 1,240 conversions at a $250 cost per conversion. The CRM shows 480 leads with a LinkedIn source, 96 opportunities, and $4.1 million in pipeline. The CFO notes that the LinkedIn number implies a cost per opportunity of $3,229 and asks whether the channel is worth keeping. Three problems are known: a consent banner blocks the Insight Tag for about 35% of EU traffic; the tag also fires on careers and support pages; and sales reps manually choose lead source, with 40% of LinkedIn-originated demo requests logged as 'organic'. Leads are routed to reps with a three-day delay, and LinkedIn's view-through conversions are included in its dashboard. The team has a CRM, a tag manager, and engineering support for one sprint.
1. Reconcile the gap between LinkedIn's 1,240 conversions and the CRM's 480 leads. What explains it?
Read the worked answer
The numbers measure different things. LinkedIn's 1,240 is a count of conversion events, not deduplicated leads: one person can convert more than once, view-through conversions are included, and the tag fires on careers and support pages. The CRM's 480 is deduplicated leads with a manually selected source. Manual sourcing is the biggest visible error: if 40% of LinkedIn demo requests are logged as organic, the corrected LinkedIn lead count is 480 / 0.6 = 800. Add the three-day routing delay and the 35% EU consent gap-the tag undercounts, while the dashboard still counts some view-through events-and the two systems cannot be reconciled line by line. The correct comparison is CRM-defined leads, opportunities, and pipeline, with a documented source hierarchy.
2. Design the measurement fix. What would you implement in one engineering sprint?
Read the worked answer
One sprint: restrict the Insight Tag to marketing pages and implement consent-aware firing; add server-side Conversions API events for form submissions so EU consent gaps shrink; capture UTMs and li_fat_id in hidden CRM fields on every form; make lead source system-derived and read-only; remove the three-day routing delay; and import SQL and closed-won stages back to LinkedIn as offline conversions. Deduplicate by email and company domain. Add a weekly reconciliation report: LinkedIn spend, CRM leads, SQLs, opportunities, pipeline, and closed-won by source. Do not attempt incrementality in the same sprint; plan a geo holdout for next quarter once tracking is clean.
3. What should the team tell the CFO, and which metric should the program optimize next quarter?
Read the worked answer
Tell the CFO not to use cost per conversion as the channel metric. On the corrected view, LinkedIn likely produced about 800 leads, 96 opportunities, and $4.1M in pipeline on $310,000 spend-roughly 13x pipeline-to-spend and $3,229 per opportunity. At a 22% win rate, that is about $900,000 in revenue, or 2.9x spend, before lifetime value; the question is whether cost per opportunity falls as tracking improves. Commit to reporting cost per SQL, cost per opportunity, and pipeline-to-spend, with a target of ≤$2,500 per opportunity next quarter. Optimize campaigns to qualified conversions and offline stages, not raw conversions, and run a holdout before making a cut decision.