Service 05
MarTech audit & connector mapping
Stacks grow in silence until OAuth tokens expire, a vendor deprecates an API, and three teams each believe someone else owns the integration that quietly fed your attribution model. We inventory tools, flows, and licences so IT and marketing share one legible map.
Deliverables are deliberately boring: spreadsheets, diagrams, and written recommendations your CIO can forward without asking for a “strategy translation layer” first.
The audit lists every surface that touches customer data: tag containers, CDPs, email engines, enrichment vendors, and the ad platforms that receive hashed identifiers—each row tagged with billing owner, renewal date, and last successful sync timestamp where your contracts allow us to read it. We flag duplicate capabilities (two CDPs, three survey tools) with annualised cost and overlap risk so finance can see trade-offs in one table instead of three vendor demos.
Connector diagrams show direction of travel, batch versus streaming behaviour, and the PII classes that cross each boundary so security reviews start from facts instead from tribal memory. Diagrams are layered: executive summary, technical detail, and “break glass” incident view that shows which integrations must survive first when you cut over data centres or regions.
Access, credentials & ownership
We inventory service accounts, OAuth clients, and API keys with rotation cadence and named custodians—so nobody discovers at 02:00 that the only person who knew the Salesforce integration user left six months ago. Where vendors support SCIM or SSO, we note current posture and the smallest change set to reach your target state without forcing a Big Bang identity migration during quarter end.
- Webhook and batch job schedules with upstream/downstream failure impact.
- Schema drift history: where marketing renamed fields that broke warehouse loads.
- Data residency by vendor and subsystem, mapped to customer contracts you publish.
- Unused seats and sandbox environments still billing production SKUs.
Risks & recommendations
Recommendations are triaged into “switch now” items—broken consent propagation, silent sampling drift, duplicated user keys—and watchlist noise that should be monitored but does not justify a fire drill this quarter. Each “switch now” item lists blast radius, estimated downtime window, and dependencies on other programmes so programme managers can sequence work instead of competing for the same maintenance window.
Every proposed deletion or consolidation path includes rollback notes: which jobs to re-enable, which DNS or OAuth steps reverse first, and how long historical replay would take if you need to unwind a decision. We also capture vendor roadmap risk: APIs marked deprecated, SKUs being retired, and products your team relies on but never upgraded because the migration guide lived in a PDF from 2019.
Handover to IT & marketing
Deliverables are written for the engineers who will operate the stack Monday morning: escalation contacts per vendor, credential rotation cadence, and the monitoring hooks your observability platform already supports. Marketing receives a parallel glossary that maps campaign jargon to the same integration IDs so briefs and tickets stop referring to different names for the same pipe.
Optional follow-through includes RFP support for replacement tools: weighted scorecards, proof-of-concept scripts, and cutover checklists that reference the diagrams we already produced—so procurement does not restart from a blank spreadsheet.