Building a channel mix scorecard: how to know which of your marketing channels are actually earning AI citations

Key takeaways {#key-takeaways} - AI citations are a channel output now, and most scorecards don't measure them at all. A channel that looks mediocre in your dashboard may be quietly feeding every ChatGPT answer in your category. - Score channels per engine, not in aggregate. Reddit earns heavy citations from ChatGPT but nearly nothing from AI Overviews; YouTube is the exact opposite. One blended number hides both wins and waste. - Audience size metrics are poor proxies for citation potential. Most cited Reddit posts have fewer than 10 upvotes, and roughly 80% of cited YouTube videos come from channels under 100K subscribers. - Attribution is fixable if you know its gaps. GA4's native AI Assistant channel misses Perplexity entirely, and an estimated 60-70% of AI-referred sessions land in Direct. A custom channel group closes most of that hole. - Weight citations by business value, not volume. AI-referred traffic converts at multiples of organic search in some analyses, so small citation counts can carry outsized revenue weight in your scorecard. --- Every marketing team has some version of the same spreadsheet. Channels down the left side, spend and ROAS across the top, a quarterly review where someone argues about whether LinkedIn deserves more budget. That spreadsheet is now lying to you in a specific way: it has no column for AI citations. When someone asks ChatGPT