HOME

ABOUT

CONTACT

Marketing Analytics · Data Science · West Palm Beach, FL · Remote

Most companies
don’t have a data
problem. They have a
thinking problem.

I build the measurement systems that replace intuition with evidence — so the right questions get asked, the right levers get pulled, and capital gets allocated with statistical confidence instead of instinct.

8+
Years
$50M+
Budget Guided
↑20%
CAC Reduction
16%
Conversion Lift
↗ Start a Conversation View Portfolio →
01 — What I Do

The Work

Senior-level analytics consulting across marketing measurement, causal experimentation, and data infrastructure. Everything I build is designed to end the guessing.

01
Media Mix Modeling
End-to-end MMM frameworks with saturation curve analysis, adstock modeling, and diminishing returns quantification. Weekly analysis cycles — model scenarios, Model vs Plan vs Actual, channel-level spend recommendations, and executive meeting facilitation. Built and owned at $50M+ annual budget scale.
MMMSaturation CurvesBudget OptimizationWeekly Cadence
02
Incrementality & Causal Testing
Geo-based holdout experiments that produce causally credible evidence of channel contribution — not correlation, not modeled attribution, actual incrementality. Designed and executed a $40K/week holdout experiment yielding a 16% conversion lift; methodology adopted as the organizational standard.
Geo HoldoutsCausal InferenceSignificance TestingA/B Design
03
CAC & LTV Prediction
Predicted CAC modeling, weekly variance analysis, and WBR-ready outputs built for Finance and Marketing leadership. LTV forecasting and cohort survival models that translate customer data into acquisition strategy and budget decisions — delivered as clean, stakeholder-ready slides every week.
Predicted CACLTV ModelingCohort AnalysisWBR Design
04
Attribution & Measurement Infrastructure
Full-stack marketing attribution: MMP platform migrations, Rockerbox integration, GA4 pipeline builds, and cross-channel data governance. Built unified analytics stacks that reduced reporting latency by 60% and replaced fragmented ad-hoc data with a single source of truth consumed across four departments.
RockerboxGA4BigQuerydbtTableau
05
Forecasting & Scenario Planning
Pull-forward scenario modeling, incremental spend vs. budget returns analysis, and channel-level ROI simulation for quarterly planning cycles. Built scenario dashboards consumed by VP Marketing and Finance leadership to make data-driven budget allocation decisions — replacing intuition-driven planning with statistical evidence.
Time-SeriesProphet / ARIMABudget ScenariosMarginal ROI
06
AI-Assisted Analytics Workflows
LLM-integrated analytics pipelines — prompt-engineered systems that automate insight summarization, anomaly narration, and recurring report generation from raw dashboard exports. Active practice in GPT-4 and Claude-assisted workflows that scale analytical output without scaling headcount.
LLM ToolingPrompt EngineeringAutomationClaude / GPT-4
02 — The Approach

How I Think

Most analytics engagements fail not because of bad data — but because the wrong questions were asked. I come in upstream of the analysis, interrogating the premise before building the model.

“The map is not the territory. The model of reality is not reality. I always distinguish between the representation and the thing itself.”

This means I will challenge the framing of your measurement problem before I build anything. I’ll ask what you’re actually trying to decide, what assumptions are baked into your current measurement stack, and where the data is lying to you.

Then I build the system that produces honest answers — statistically rigorous, executive-facing, and connected to actual capital allocation decisions.

↗ See the work in detail — portfolio and case studies

I
Question the Premise
Every measurement problem comes packaged with assumed frames. I find the frame before engaging the problem. The question behind the question is where insight lives.
II
Replace Intuition With Evidence
Budget decisions, channel bets, and growth investments should be governed by statistical models — not what worked last quarter or what the platform is reporting.
III
Translate Rigor Into Decisions
A model that can’t be explained in a VP meeting is a model that won’t change behavior. Every output is designed for the decision-maker who has to act on it.
IV
Build for Continuity
Infrastructure over one-off analysis. Weekly cadence over quarterly check-ins. Systems that outlast the engagement and make the team permanently more capable.
↑20%CAC Reduction
$50M+Budget Guided
16%Conversion Lift
Holdout Validated
60%Latency Reduction
1M+CRM Records
Modeled
4.2×LTV : CAC
Achieved
03 — Experience

Where I’ve Built

8+ years across SaaS, logistics, security tech, and Fortune 500 CPG — always as the person building the measurement infrastructure, not just reporting from it.

SimplePractice
Jun 2022 – Present
SaaSHealth & Wellness
Staff / Senior Marketing Analyst — Business Analytics
Owned the full measurement and analytics infrastructure for a SaaS platform serving health and wellness practitioners. Built end-to-end MMM framework informing $50M+ in annual budget allocation, designed and executed incrementality experiments (16% conversion lift), drove 20% CAC reduction, led full Rockerbox attribution migration reducing reporting latency by 60%, and built LTV and propensity models. Embedded IC partner to VPs across Marketing, Product, Customer Success, and Sales.
REEF Technology
Jan 2021 – Jun 2022
Logistics4,500+ Locations
Senior Business Intelligence Engineer
Owned BI and analytics infrastructure across operations, finance, and workforce planning. Built real-time Tableau dashboards integrating Samsara APIs (20% delivery time reduction), multivariate regression and time-series forecasting models for demand planning (10% labor cost reduction), and consolidated 12+ manual data sources into a single governed BI layer consumed by 20+ stakeholders.
ADT
Mar 2020 – Dec 2020
Security TecheCommerce
Customer Experience Insight Analyst
CX analytics across ecommerce, field services, and call center environments. Clickstream and funnel analysis informing UX changes that improved conversion rates. Optimized SMS pipeline (↑10% spam rate), reduced cost per lead 8% through predictive routing models, and synthesized VoC data influencing product roadmap.
Newell Brands
Aug 2017 – Mar 2020
Fortune 500CPG
Senior Analyst — Consumer Services
Python ML pipelines over 1M+ CRM records across 50+ product lines. Built retention and churn cohort models, demand forecasting systems, anomaly detection for quality escalations, and feature engineering pipelines for production compliance workflows. Reduced manual processing time 35%, improved agent utilization 10%.
04 — Contact

Let’s Work Together

If your marketing spend decisions are still governed by platform reporting, last-click attribution, or gut instinct — there’s a better way. Reach out and let’s talk about what measurement could actually look like for your organization.

@
Business Email
taj.anderson@divergentmcs.com
#
Phone
561-215-3548
↗
Portfolio & Case Studies
tajhebs-p0v.github.io/portfolio
Availability
Open to New Engagements
Mon–Fri · 9am–6pm EST · Contact for pricing