DefaultSearchMediaModule: Model paid and/or organic search.

Description Usage Arguments Value

Description

Simulate the behavior of a paid and/or organic search, including observable variables (e.g., query volume, paid clicks, spend) and the effect on consumer mindset.

Usage

1
2
3
4
5
6
7
DefaultSearchMediaModule(data.dt, budget.index, budget,
  spend.cap.fn = function(time, budget, budget.indices) {     Inf },
  bid.fn = function(time, per.capita.budget, budget.indices) {     Inf },
  kwl.fn = function(time, per.capita.budget, budget.indices) {     1 },
  audience.membership = list(), query.rate = 1, cpc.min = 0,
  cpc.max = 1, ctr = list(), relative.effectiveness = c(0, 0, 1),
  transition.matrices = list())

Arguments

data.dt

data.table with rows corresponding to population segments and columns corresponding to specific variables

budget.index

vector specifying budget period each time point belongs to. For example, rep(1:4, each = 52) would correspond to 4 years of yearly budget periods.

budget

vector specifying the target spend for each budget period. For example, given the example budget.index from above, budget = rep(1e6, 4) would specify a budget of 1 million for each year.

spend.cap.fn

function mapping the current time, the budget, and the budget period to a spend cap for the current week. By default this is set to Inf, representing uncapped spend.

bid.fn

function mapping the current time, the per-capita budget over the population, and the budget period to a bid for the current week. By default this is set to Inf, so that the advertiser wins all auctions and will pay the maximum CPC.

kwl.fn

function mapping the current time, the per-capita budget over the population, and the budget period to the proportion of queries. that match the keyword list. By default this is the maximum value of 1. To specify the proportion of matching queries by population segment, have kwl.fn return a vector with entries for each segment.

audience.membership

list of multipliers used to calculate probability of audience membership. Each element of the list corresponds to a specific dimension of population segmentation. Multipliers corresponding to each dimension are multiplied to derive audience membership probability for each segment. A named list with members 'activity', 'favorability', 'loyalty', and 'availability' is expected. Each member is a numeric vector containing the multipliers to use for each state in the dimension. For example, if member "activity" is c(1, 0.5, 0.7), a multiplier of 0.7 should be used for all segments with activity state "purchase." By default, any missing members will have no effect.

query.rate

nonnegative numeric, or vector. Each member of the audience makes matching queries according to a Poisson process with this rate. A vector rate specifies the query rate at each time. Note that rate is the expected number of queries per person in the audience. Defaults to 1. Vector repeats as necessary, so that repeating patterns can be specified more simply.

cpc.min

minimum CPC, defaults to 1. Must be nonnegative. vector values are interpreted as the vector of minimum CPC's over time.

cpc.max

maximum CPC. Must be at least as large as cpc.min. vector values are interpreted as the vector of maximum CPC's over time.

ctr

list of multipliers for each dimension with an effect on the clickthrough rate (ctr). Values in each state are multiplied to derive the ctr for each population segment. A named list with members 'activity', 'favorability', 'loyalty', and 'availability' is expected. Each member is a numeric vector of the values for each state in that dimension. By default, any missing members will have no effect.

relative.effectiveness

effectiveness, relative to the maximum effectiveness specified by the transition matrices, by volume type: organic only, paid impressions w/o paid click (click on organic result included), and paid clicks. Default to maximum (1) effectiveness for paid clicks, and no effect otherwise.

transition.matrices

list of transition matrices for each dimension of population segmentation that may be affected by marketing interventions. A named list with members 'activity', 'favorability', 'loyalty', and 'availability' is expected. By default, any missing members will have no effect.

Value

invisible(NULL). data.dt updated by reference.


google/amss documentation built on May 20, 2019, 5:05 p.m.