mmm_attribution

mmm_attribution estimates attribution from aggregated marketing signals. The function requires a D_variables table with variable, channel, measure, and prior_weight, so signals measured on different scales can be compared more realistically. prior_weight: business weight assigned to each raw signal. A common way to estimate it is conversions / touchpoints or conversions / cost, but it can also be set from domain knowledge or historical benchmarks. Use the parameters below to select the model, lag depth, diagnostics, and runtime options.

PARAMETER	TYPE	DEFAULT	DESCRIPTION
Data	data.frame / table	required	Aggregated dataset containing a timestamp column, the target column, and the marketing or business signal columns used for attribution.
D_variables	data.frame / table	required	Mapping table with one row per input signal. It must contain variable, channel, measure, and prior_weight. prior_weight is the business weight assigned to each raw signal. A common way to estimate it is conversions / touchpoints or conversions / cost, but it can also be set from domain knowledge or historical benchmarks.
target	string	required	Name of the target column in Data. The target must exist in Data and must not be included in D_variables.
model	string	"linear"	Attribution engine. Allowed values are "linear", "reward", and "copula".
max_p	integer	12	Maximum lag depth considered when evaluating the relationship between input signals and the target.
nsim	integer	1000	Simulation parameter used by model components that require simulation, kept for API consistency.
seed	integer	1234567	Random seed used for reproducible results where random or simulation-based components are involved.
return_diagnostics	bool	False	If True, returns additional diagnostic tables useful to inspect signal weights, scoring, saturation, and attribution shares.
verbose	integer / bool	1	Controls runtime logging. Use 0 or False to reduce printed output.
ncore	integer	1	Number of cores used by supported computation steps.
password	string	optional	ChannelAttributionPro license token.

Example structure for D_variables:

variable	channel	measure	prior_weight
google_clicks	google_ads	clicks	0.25
google_impressions	google_ads	impressions	0.015
facebook_impressions	facebook_ads	impressions	0.015
email_clicks	email	clicks	0.30
direct_searches	direct	searches	0.50