Frequently asked questions
For Directors and VPs at $50M+ DTC eCom evaluating whether their data problem is an accuracy problem or a trust problem between teams.
What is a Fractional Head of Analytics?
A Fractional Head of Analytics is an experienced analytics leader who runs your analytics function part-time, embedded in your team, on a multi-month engagement. Unlike a project consultant, a Fractional Head owns the foundation: definitions every department signs off on, dashboards the CMO trusts, production and dev environments aligned, and a roadmap that scales with the business. Unlike a full-time hire, a Fractional Head ramps in days, brings cross-industry pattern recognition, and exits cleanly when the team is ready to own the system. The right fit is typically a $50M+ DTC eCom that needs senior analytics leadership before it can justify a full-time VP of Analytics, or a company whose existing data team is technically strong but lacks the cross-functional governance work that turns data into trusted numbers.
Learn more →What is The First Loop?
The First Loop is a four-week diagnostic engagement designed for Directors and VPs of Analytics, Marketing Ops, or Growth at $50M+ DTC eCom. Week one is diagnosis: stack walkthrough, lineage map, and a definition audit that surfaces where the same metric is being calculated three ways. Week two defines the Horizon, a written success definition that stakeholders sign off on before any code gets written. Week three ships one trusted artifact end-to-end, on the foundation that's already there. Week four is stakeholder review and either a written proposal for an ongoing engagement or a clean exit with the artifact in hand. Fixed scope, fixed price, Director-budget-friendly. The First Loop is the smaller, cheaper, faster way to find out whether your data problem is an accuracy problem or a governance problem before you spend the budget on the next tool that promises clarity.
Learn more →What are DAR loops?
DAR stands for Diagnose, Act, Reflect. It is a two-week iteration cadence designed to replace rigid frameworks with iteration that has strategic intent. Diagnose: identify what is actually broken, with evidence from the lineage and the stakeholders, not from the dashboard alone. Act: ship the smallest unit that delivers value, end-to-end, on the shared foundation. Reflect: review with stakeholders, capture what worked and what did not, and adjust the next loop. Each loop is anchored to the Horizon, the success definition that does not move. Most engagements ship between three and twelve DAR loops over an embedded retainer.
What is the Horizon?
The Horizon is a written success definition that stakeholders sign off on before any code gets written. It names the artifact being shipped, the metric that matters, the decision the artifact will unblock, and the evidence that will indicate the engagement is on track. The Horizon is the difference between an engagement that drifts (everyone happy, nothing changes) and an engagement that produces a 9am dashboard the whole org trusts. Most failed analytics projects do not fail at the technical layer. They fail because the organisation never agreed, in writing, on what success looked like.
Why do most marketing data problems turn out to be governance problems, not accuracy problems?
Inside most $50M+ DTC eCom companies, the data is technically fine. The warehouse is full. The pipelines run. The dashboards refresh. What is broken is the agreement: marketing wants Gross Revenue when calculating ROAS, finance wants Net Revenue, operations wants something in between. Nobody has explicitly disagreed, but reports drift to align with individual preferences, not departmental ones. The same dimension gets calculated three ways across the warehouse. Engineers stop modifying tables and start building duplicates because they cannot see what depends on what. The result looks like a data-quality problem because the numbers do not match. The actual problem is governance: visible lineage, smaller iterations, and difficult conversations brought to a foundation everyone can see. Buying a new tool does not fix governance.
Read the case study →When is the right time to deploy AI on my marketing data?
AI does not fix data problems. It amplifies them. Garbage in, garbage out, on steroids. The right time to deploy AI on your data is after the foundation has been earned: a single source of truth across departments, definitions everyone has signed off on, lineage you can trace, and dashboards stakeholders trust at 9am. Deploying AI agents on top of a warehouse where the same metric is calculated three ways produces fluent nonsense at scale. The shortcut most vendors are pitching, agents on top of the existing mess, is the most expensive way to discover this. Adoption-first earns the right to deploy AI without amplifying the underlying mess.
How is a Fractional Head of Analytics different from hiring a consultancy?
Big consultancies sell six-month frameworks that look great in the deck and rarely survive contact with the warehouse. The work product is usually a strategy document and a slide library. A Fractional Head of Analytics works inside your stack, ships trusted artifacts on a two-week cadence, and is measured against a success definition stakeholders signed off on before the engagement started. The buyer does not pay for slides. The buyer pays for shipped artifacts the team owns after the engagement ends. Vendor-agnostic by default: no preferred partners, no reseller margins, no incentive to recommend the tool that pays the highest commission.
Do I need to buy a new platform to get a 9am dashboard everyone trusts?
Almost never. Most $50M+ DTC eCom companies already have everything they need: a warehouse like BigQuery or Snowflake, a transformation layer, and a BI tool. What is missing is governance: visible lineage, deduplicated calculations, definitions every department agreed to in writing, and production and development environments that stay aligned. Adding a new attribution platform, a new MMM tool, or a new AI agent layer on top of an ungoverned warehouse adds another layer of pipelines on top of the same soup underneath. The recent case study (a $100M+ DTC eCom getting to a 9am dashboard the whole org trusts) was delivered in two months on the existing BigQuery stack, with dbt added for visibility, and zero new tools or subscriptions.
Read the case study →Have a question that’s not here?
Book a free 30-minute call. No pitch, an honest conversation about whether the work I do is the work you need.
Book a free call