Reporting Stack Intake Audit (Onboarding)

Reporting Stack Intake Audit (Onboarding)

1. 01

   Time the current reporting cycle end to end before opening any vendor trial

   Have one account manager log every touch from data pull to client send for two consecutive cycles. Record hours per client and the number of people involved; that baseline is the only honest measure of what a platform can recover.

2. 02

   Inventory every data source the client actually cares about, ranked by decision weight

   List the ad platforms, analytics properties, CRM, call tracking, and commerce systems behind the numbers clients ask about first. A tool covering 85 sources is worthless if the three that drive the retainer conversation are missing.

3. 03

   Map each source to its owner, access method, and refresh cadence

   Note whether access runs through a client-owned login, an agency MCC, or a shared API key. Flag any source that refreshes weekly or slower, since a live dashboard will imply freshness the data cannot support.

4. 04

   Reconcile one month of numbers by hand against the client's own system of record

   Pull a single month from two sources (for example platform-reported spend versus the client's finance export) and document every variance above 2%. Reconciliation gaps surface in onboarding or they surface in a client review, never both.

5. 05

   Confirm white-label and access controls against the client contract

   Check whether the retainer promises branded portals, custom domains, or SSO, and whether the client's procurement team requires data residency or a DPA. Some platforms ship full white-labeling, others none, and the gap is a contract issue, not a design preference.

6. 06

   Define the three to five metrics that will anchor every recurring report

   Write the definitions down, including attribution window and currency handling. Reach is losing ground as a headline paid media KPI heading into 2027 budgets, so agree now on what replaces it rather than inheriting last year's template.

7. 07

   Run a parallel cycle where the platform output is checked against the manual report

   Deliver one report from each process to an internal reviewer and diff them line by line. Adoption proceeds only when the automated version matches or beats the manual one on accuracy and time.

8. 08

   Assign a named owner for source health and a review date 30 days out

   Connector breakage is the most common failure after go-live. Put a person and a calendar date on the first health check so drift is caught before a client notices a flat line.