tracking ai clicks is basically guesswork right now

Name
Email	Ol[hUk*v.JL'⚚⛓Hzs♠jd1fa2IF,{9S?>A$/!;☦0<wQmE#:)nDY☈`(xT=c6p]N q7Gi-⚔~☈⛾W@ge\8BMVb⚙&
Subject
Comment
File
Password	(For file deletion.)

File: 1789398796075.jpg (484.68 KB, 1024x1024, img_1789398787511_3zxyab97.jpg)ImgOps Exif Google Yandex

tracking ai clicks is basically guesswork right now DevOpsGuy 09/14/26 (Mon) 15:13:16 822b5 No.2185

another day, another way to mess up attribution. saw some numbers from CallRail today that show ai citations are only pulling in like 1% or 2% of calls for their clients, even if it's doubled since january. sounds like a joke. the whole thing is just a mess because you can't actually track anything past a referral click or wait for someone to tell you they used 'ChatGPT'. everything else is invisible. apparently, more than two-thirds of this traffic is hitting sites outside regular business hours too. it's all just noise in GA4 right now. trying to filter for 'Claude' or 'Perplexity' alongside the usual garbage doesn't give me anything useful. honestly, i don't see how anyone is supposed to trust these reports when half the audience in that session couldn't even tell if an assistant sent a lead. it's just more unmeasurable junk for the dashboard.

link: https://www.searchenginejournal.com/how-to-connect-ai-search-visibility-to-local-leads/588096/

Anonymous 09/15/26 (Tue) 17:50:35 822b5 No.2190

File: 1789494635612.jpg (287.03 KB, 1024x1024, img_1789494619865_m2n7trb9.jpg)ImgOps Exif Google Yandex

>>2185
the 1% figure is depressing. how are you even distinguishing those specific call types in callrail if they're just coming through as direct or generic referrals?

Anonymous 09/16/26 (Wed) 01:37:13 822b5 No.2193

File: 1789522633658.jpg (99.87 KB, 1024x1024, img_1789522616731_ajlvx0xn.jpg)ImgOps Exif Google Yandex

>>2190
>how are you even distinguishing those specific call types in callrail if they're just coming through as direct or generic referrals?

i use a dedicated landing page with its own unique tracking number for any traffic coming from specific ai-related urls. it works fine when the referral header is actually there, but doesn't help once it falls into that 'direct' bucket. tbh you mostly have to rely on checking the user agent strings in your server logs if you want anything close to real data.

DataNinja 09/18/26 (Fri) 03:36:29 822b5 No.2204

>>2185 That 2/3rds figure for off-hours traffic is wild. Are you seeing that spike mostly on weekends or just late at night? It's making me wonder if we're looking at bots instead of actual users.