Based on https://developers.google.com/analytics/devguides/reporting/mcf/dimsmets/ I download using RStudio (on Windows 7)
dimensions = "mcf:dcmCampaign, mcf:basicChannelGroupingPath, mcf:conversionGoalNumber, mcf:timeLagInDaysHistogram",
metrics = "mcf:totalConversions"
Then filter
conversionGoalNumber=="002" and
as.numeric(timeLagInDaysHistogram) < 15 (to get chains of last 14 days)
Then I model the Linear Attribution Model focussing on a certain channel (e.g. "Display" or "Paid Search").
Main Problem: sum of totalConversion (before modelling) is already smaller than the value shown in the Interface of Google Analytics Attribution.
Has anybody experience whether these low numbers could be due to filtering for conversionGoalNumber and/or timeLagInDaysHistogram? Does anyone know of better alternatives?
Has anyone reproduced the Linear Attribution Model in R?
Thanks a million for any help and advice!
