@timsoulo@ahrefs I would guess this doesn’t include zero-click behaviour that shows up in Clouldflare requests but not in “traffic” where the LLM visits the site on behalf of the user but the user themselves never visits. Is this where the missing traffic has gone?
@GVG_Genetics@photofinishgame@AnthropicAI it basically guesses the answer (and will get it wrong). last paragraph is wrong - suggests a correlation - but you shouldn't correlate the overall numbers - you need to correlate at a race level and measure the coefficient at diferrent lengths (apols if that's what you did)
@GVG_Genetics@photofinishgame@AnthropicAI to get true correlations you need to analyse line by line data. the response doesn't have any actual correlation coefficients - which is weird. also need to account for field size, especially in smaller data sets. ask it to multiply two large numbers and check the answer vs excel
@brettbilina@ErikLambert1 Five is tough, but:
Harbaugh Bears>Colts, QBR 74>86
Gannon Vikings>Raiders 74>91
Testeverde Bucs>Browns 64>81
which proves your point...
You can see that a ChatGPT style bot tailored to your business would be incredibly transformative but don't have in house experience building such chatbots?
Talk to Kortical!
Reach out for a demo today 📤
#chatgpt#chatbot#ai