How I run an enterprise onboarding

Most onboardings I lead take a sports club, league or federation onto a Customer Data Platform (CDP): their ticketing, e-commerce, CRM and app data, merged into one profile per fan. I have worked on more than forty of them across 13 countries in Europe and the US, usually with a few running at once.

Each phase below shows what I own, who I bring in, what it produces, and what usually goes wrong there, with how I catch it before go-live.

  1. Phase 1: Kickoff and requirements

    What I own
    The scope, the milestone plan, and the full list of systems that will feed the platform.
    Who I bring in
    The customer's commercial lead, their IT or data contact, and our account CSM.
    What it produces
    A scoped plan: every data source, every integration, the go-live date, and who owns each piece.
    What goes wrong, and how I catch it
    The goal arrives as a feature ("we want segments") instead of a result, and the system list is missing the one nobody mentions. I ask what they will do with the data on day one, and ask for the system list in writing, including the systems "nobody uses anymore".

  2. Phase 2: Data mapping and integration design

    What I own
    The technical meetings with each provider, the integration method, and the data schema.
    Who I bring in
    Ticketing, e-commerce and CRM providers, the customer's IT, and our developers when something new has to be built.
    What it produces
    A specification per source: endpoints, fields, schema, the logic that derives anything the source does not send, and calls I have already tested.
    What goes wrong, and how I catch it
    Documentation is missing or wrong, or an endpoint does not return what the commercial goal needs. I test the endpoints myself before a developer touches them, and when data is missing I add logic to the schema to derive it rather than wait for the provider.

    See this step worked through on three invented records

  3. Phase 3: Build and configuration

    What I own
    Standard integrations from the library, the per-customer data warehouse configuration, and the identity resolution design.
    Who I bring in
    Developers, only for integrations that do not exist yet.
    What it produces
    Data flowing on schedule and contacts resolving into single customer profiles.
    What goes wrong, and how I catch it
    Credentials and access arrive late, which is the most common reason a date slips, and identity resolution merges too much or too little. I track credentials as a dependency from the kickoff, and check profile counts on the first load against each source's own numbers.

  4. Phase 4: Testing and go-live

    What I own
    Validation against the source systems and the go-live call.
    Who I bring in
    The customer's data owner, who signs off the numbers.
    What it produces
    Reconciled counts, a signed-off setup, and a live account.
    What goes wrong, and how I catch it
    The first dashboard disagrees with a report the customer already trusts, and confidence goes with it. I reconcile against their own reports before they see anything, and explain the differences, such as duplicates merged, before they have to ask.

  5. Phase 5: Handover and adoption

    What I own
    Analytics dashboards, segment training, and the handover to whoever runs the account day to day.
    Who I bring in
    The customer's marketing and CRM team.
    What it produces
    Working dashboards, a team that builds its own target segments, and a documented setup.
    What goes wrong, and how I catch it
    The platform is live and nobody uses it. Training is built on their own goals, so the first segments they create in the session are ones they will send.

What carries over to the next one

My onboarding playbook started as a step-by-step framework in Jira, so nothing was forgotten, and grew into standard analytics setups, integration methods and mapping schemas. With the integrations library that followed, a known integration now takes hours to set up instead of the four to six weeks an onboarding took when I started.

The latest layer is an AI toolkit I built on a knowledge library of how the platform is configured. It migrated customer accounts to a new segmentation engine through the portal's API, and it now builds dashboards and analytics tables from a customer's raw data sources. The case studies show the method on real projects.

Get in touch Read the full profile