Marketing Analytics Fundamentals

The core competency separating winners. Scope, stack, skill hierarchy, methods, SQL, dashboards vs analysis, and team building.

Why analytics is the core competency

Marketing has more data than any other corporate function. Programs that turn data into decisions outperform programs that drown in dashboards. The discipline isn't about more reports; it's about better questions, cleaner analysis, and faster decisions.

The mistake: hiring a junior analyst to maintain reports. The discipline: investing in analytics as a strategic capability that informs every major decision.

What marketing analytics covers

• Descriptive. What happened? Performance reporting, dashboards, KPIs.
• Diagnostic. Why did it happen? Drill-downs, cohort analysis, funnel analysis.
• Predictive. What will happen? Forecasting, LTV modeling, churn prediction.
• Prescriptive. What should we do? Budget optimization, recommendation systems, scenario planning.
• Causal. What did our marketing cause? Incrementality testing, MMM, quasi-experimental methods.

The modern stack

• Collection. Browser/app/server events; CRM data; ad platform data.
• Warehouse. Snowflake, BigQuery, Databricks, Redshift.
• ETL/ELT. Fivetran, Stitch, Funnel, Supermetrics.
• Transformation. dbt, SQLMesh, Dagster.
• BI / dashboards. Looker, Tableau, Power BI, Mode, Hex.
• Statistical / ML. Python (pandas, scikit-learn, statsmodels), R, Jupyter, BigQuery ML.
• Specialized. Amplitude / Mixpanel for product analytics; Recast / Haus for MMM; Statsig / Eppo for experimentation.

Skill hierarchy

Level	Skills
Reporter	Dashboard maintenance; data pulling; basic SQL
Analyst	Statistical analysis; SQL fluency; experimentation design; A/B test interpretation
Senior analyst	Causal inference; regression modeling; cohort analysis; stakeholder communication
Data scientist	ML modeling; predictive analytics; advanced statistics; productionization
Analytics engineer	Pipeline architecture; transformation layer; data modeling
Head of analytics	Team building; strategy; cross-functional partnership

Core methods every analyst needs

• Descriptive statistics and distribution analysis.
• Hypothesis testing (t-test, chi-square, proportion tests).
• A/B test design and interpretation.
• Confidence intervals and statistical significance.
• Cohort analysis and retention curves.
• Funnel analysis.
• Segmentation and clustering.
• Regression (linear, logistic).
• Time series analysis.
• Causal inference basics (difference-in-differences, regression discontinuity).

SQL and warehouse-native

SQL is the lingua franca of analytics. Modern analysts spend most of their time in SQL.

• Joins, aggregations, window functions, CTEs.
• Date/time manipulation.
• String manipulation and regex.
• JSON parsing (for event data).
• Performance optimization (partitioning, clustering).
• Warehouse-specific functions (BigQuery, Snowflake variations).

Beyond SQL: dbt for transformation; Python for statistical analysis; BI tools for visualization.

Dashboards vs analysis

• Dashboards: Recurring visibility. KPIs, trends, anomalies. Low-frequency new insight.
• Analysis: Specific questions answered deeply. Higher insight density.
• Common failure: Analytics team building dashboards only; never answering specific business questions.
• Healthy balance: ~30% dashboards / 70% analysis at mature programs.

Building the analytics team

• Early stage: One generalist analyst.
• Growth stage: 2–4: analyst + analytics engineer.
• Mature: 5–15: specialists in product analytics, marketing analytics, data science, analytics engineering.
• Enterprise: Separate teams; central platform + embedded analysts.

Advanced playbook

• Documented analytics charter. Purpose, scope, decision authority.
• Stakeholder request management. Intake process; prioritization; turnaround SLAs.
• Self-service BI investment. Train stakeholders to answer simple questions themselves.
• Analytics rituals. Weekly metric review; monthly analysis showcase; quarterly business review.
• Code review and version control. SQL and Python in git; PR review.
• Documentation discipline. Metric definitions, data sources, transformations documented.
• Cross-functional embedded analysts. Analysts attached to specific teams (growth, product, marketing).
• Annual capability audit. What can we do? What can't we yet? What should we invest in?
• External partnerships. Vendor specialists for MMM, advanced ML, etc. where in-house can't justify.
• Analyst training and growth paths. Senior IC and management tracks both available.

Common mistakes

• Analytics team as report factory; no strategic input.
• Dashboards proliferating without curation; nobody knows which to trust.
• Metric definitions drifting; same KPI reported differently across teams.
• No version control on SQL; changes break things silently.
• No stakeholder request triage; team buried in ad-hoc.
• Generalist analysts forever; never specialization.
• Tool sprawl; budget waste.
• Analytics disconnected from product; insights don't inform roadmap.
• SQL-only team without statistical skill; rigorous analysis impossible.
• External vendors used as crutch; in-house capability never builds.
• No documentation; institutional knowledge fragility.
• No prioritization framework; whoever asks loudest wins.

Operating checklist

• Analytics charter documented
• Stakeholder request intake and prioritization
• Metric definitions canonical; semantic layer enforced
• Version control for SQL and transformation code
• Dashboard library curated, not proliferating
• Self-service BI for stakeholders
• Analytics rituals (weekly, monthly, quarterly)
• Team skill mix balanced (analyst, engineer, scientist)
• Cross-functional embedded analysts where scale supports
• Annual capability audit
• Documentation: metric dictionary, runbooks, playbooks
• Career paths documented for IC and manager tracks

---

Part of the Marketing Analytics series.