Version: Latest

uam

R

library(ChannelAttributionPro)

token <- "yourtoken"

df_aggr <- read.csv("https://app.channelattribution.io/data/data_mmm.csv")

D_variables <- data.frame(
  variable = c(
    "direct_searches",
    "facebook_impressions",
    "facebook_cost",
    "google_impressions",
    "google_clicks",
    "google_cost"
  ),
  channel = c(
    "direct",
    "facebook_ads",
    "facebook_ads",
    "google_ads",
    "google_ads",
    "google_ads"
  ),
  measure = c(
    "direct_searches",
    "impressions",
    "cost",
    "impressions",
    "clicks",
    "cost"
  ),
  stringsAsFactors = FALSE
)

target <- "revenue"

Aggregated-data attribution without customer journey paths:

res <- uam(
  df_aggr = df_aggr,
  D_variables = D_variables,
  target = target,
  df_paths = NULL,
  baseline_model = "reward_model", # default; use "auto_mmm" to use the MMM baseline
  max_p = 12,
  nsim = 1000,
  seed = 1234567,
  verbose = 1,
  password = token
)

print(head(res))

With already aggregated customer journey paths:

df_paths <- read.csv("https://app.channelattribution.io/data/data_uam_paths_t.csv", sep = ";")

res <- uam(
  df_aggr = df_aggr,
  D_variables = D_variables,
  target = target,
  df_paths = df_paths,
  baseline_model = "reward_model", # or "auto_mmm"
  var_path = "path",
  var_conv = "total_conversions",
  var_null = "total_null",
  order = 1,
  sep = ">",
  verbose = 1,
  password = token
)

print(head(res))

With event-level customer journey paths:

df_paths <- read.csv("https://app.channelattribution.io/data/data_uam_paths.csv", sep = ";")

res <- uam(
  df_aggr = df_aggr,
  D_variables = D_variables,
  target = target,
  df_paths = df_paths,
  baseline_model = "reward_model", # or "auto_mmm"
  channel_conv = "((CONV))",
  order = 1,
  sep = ">",
  verbose = 1,
  password = token
)

print(head(res))

R​

R