Skip to main content


Warning: hta_model is in Beta phase and indicated only for experimental purposes.

HTA model (Hidden Touch Attribution Model) is a new attribution model alternative to Media-Mix Model. It uses traffic data (impressions and clicks for digital channels, estimated number of users reached by Tv or press ads, ecc..) to split attribution between digital and traditional channels. Attribution with HTA model can be performed with function hta_model.


Datadata.framedata.frame containing time-series of aggregated traffic data from different channels, aggregated conversions and/or conversion value
var_timestrname of the column containing time dimension
var_convstrname of the column containing aggregated conversions
channelslis(str)list of names of the columns containing aggregated traffic data for channels considered
var_valuestrNonename of the column containing aggregated conversion value
glob_conv_ratedoubleNoneglobal conversion rate for the use case considered. If None, it will be internally estimated
L_conv_ratelist(double)Noneconversion rate for each channel considered. If None, then the global conversion rate will be used for each channel considered
L_weightdouble/list(double)1.0weight for each channel considered. It is a quantity between 0 and 1 and it is used to weigh differently different types of touches (e.g. clicks and impressions)
mean_cj_lengthdouble1.0It indicates how many units of time are long the customer journeys on average for the use case considered. E.g., if input data are weekly and the parameter is set to 1.5, it means that customer journeys are long 1 week and a half on average for the use case considered
L_ads_ratedouble/list(double)0.0list of adstock rates (between 0 and 1) for the channels considered
nsimint10number of simulations to be performed
seedint1234567random seed. Giving this parameter the same value over different runs guarantees that results will not vary
serverstr""address of the server where password will be checked to authorize the execution of the function


attributiondata.framequote of conversions and/or conversion value attributed to each channel considered at each time point