← Back
Home › Systems › Sales, marketing and customer service

marketing-measurement

Bayesian marketing mix modelling, geo-lift and synthetic-control incrementality tests, multi-touch attribution, budget optimiser

Google Meridian, Measured, Rockerbox, Haus

SystemTechniques · 23Primitives · 56 marketing-measurement Generalized linear models (GLMs)machine-learning · medium: Linear model for counts, rates, positive amounts and proportions: a linear score pushed through a link, fitted by iteratively reweighted least squares.›Shapley values treeshapmachine-learning · medium: TreeSHAP and Shapley attributions for per-prediction explanations and reason codesHamiltonian monte carlo nutsnumerics · medium: HMC with leapfrog integration and the No-U-Turn sampler; step size and mass-matrix adaptationMarkov chain monte carlonumerics · medium: Metropolis-Hastings and Gibbs sampling: proposals, burn-in, thinning, convergenceVariational inferencenumerics · medium: ADVI and mean-field VI: optimize the ELBO with reparameterization gradientsQuadratic programmingoptimization · medium: Convex QP solvers (active-set, interior-point, ADMM) for constrained optimizationPrivate set intersectionsecurity · medium: Compute the overlap of two parties' sets without revealing the rest (ECDH-PSI, OPRF)Bootstrap confidence intervalsstatistics · low: Resampling-based confidence intervals for arbitrary statisticsSessionizationstorage · low: Group events into sessions by inactivity gapsAdstock saturation transformsstatistics · low: Carry-over (adstock) and diminishing-returns (saturation) transforms of spend before regression in mix modelsCUPED (variance reduction with pre-…statistics · low: Subtract the part of each unit's outcome that a pre-experiment covariate predicts; the effect estimate stays unbiased and its variance shrinks by 1 - rho^2.›Difference in differencesstatistics · low: Causal effect estimation from treated/control group pre/post changesSynthetic controlstatistics · low: Construct a weighted control from donor units to estimate an intervention effectLinear and logistic regressionmachine-learning · low: Fit a weighted sum of features to a number (least squares) or a probability (logistic); the baseline every model must beat.›LU and Cholesky factorizationnumerics · low: Factor A = PLU (any square matrix, partial pivoting) or A = LLᵀ (SPD) once in O(n³), then solve each right-hand side in O(n²).›Newton root findingnumerics · low: Linearize at the guess and jump to the tangent root; quadratic convergence near a simple root, made robust by brackets (1D) or line search.›Monte Carlo integrationnumerics · low: Estimate an expectation by averaging random samples; error falls as σ/√N in any dimension, and variance reduction (antithetic, control variates, importance sampling) shrinks σ.›Reverse-mode automatic differentiat…ml-training · low: Record the forward computation on a tape, then sweep it backwards once to get the gradient of a scalar loss for all inputs.›Constrained parameter transformsnumerics · low: Map constrained parameters to unconstrained space (log, logit, simplex, Cholesky) with Jacobian correctionsMcmc convergence diagnosticsnumerics · low: R-hat, effective sample size and divergence checks for MCMC runsPRNGs (PCG, xoshiro)numerics · low: Small, fast, statistically strong non-cryptographic generators: PCG32 and xoshiro256**, with seeding, parallel streams and unbiased ranges.›Verlet / leapfrognumerics · low: Half-kick, drift, half-kick: a second-order, time-reversible, symplectic integrator with one force evaluation per step whose energy error stays bounded for millions of steps.›AdamWml-training · low: Adam's per-parameter adaptive steps with weight decay applied directly to the weights; the default optimizer for transformers and LLM pretraining.›+ 16 more techniques+ 40 more primitives