Why mccloudCo

Data infrastructure should be boring. We make it boring on purpose.

mccloudCo started with a simple observation: a school district and a growing business have almost nothing in common — except both needed senior leadership they couldn't justify hiring full-time. Building one advisory model that actually works for both was never the easy part. It was just the part worth doing.

mccloudCo is two practices led by two principal consultants — one for schools and districts, one for growing businesses — and neither of us learned this business from the outside. We built the firm after watching what happens when "senior advisor" quietly becomes "whoever's available" once an engagement runs past the first few months, and after watching how often advice gets bent toward whatever platform or curriculum the advisor happens to sell. Neither happens here. It shows in how we work: the same consultant from the first call to the last, a named backup so nothing stalls if that consultant is ever out, and recommendations that are never steered by a commission — no curriculum, platform, or product endorsement sits behind anything we tell you. When an engagement ends, the rubrics, playbooks, and frameworks we used stay with your team; capability is the actual deliverable, not a report that gets filed away and forgotten.

We diagnose before we recommend anything — a short review, either standalone or built into the first month of working together, that tells us what the situation actually calls for rather than what we'd like to sell. From there, everything gets scoped and put in writing before any work begins, so there's never a gap between what was discussed on a call and what actually shows up in the engagement. The relationship is built to let you start small and grow only when the work demands it: plenty of our clients stay exactly where they started for years, and we treat that as a good outcome, not a missed upsell. Response times are published, not promised — thirty minutes for anything critical, a single business day for a routine request — and if the diagnostic turns up nothing that justifies bringing us in further, we say so directly instead of finding a reason to proceed anyway.

Northlane engineers reviewing pipeline incidents together in the Toronto office
Toronto studio — Tuesday incident review

Trust

How Northlane handles your data

The controls, retention windows and sub-processors behind every Northlane deployment. Last reviewed 14 March 2025.

Where your data lives

Northlane never copies your rows. The platform reads metadata — query history, information-schema snapshots, dbt manifests and BI catalogue exports — and stores the resulting lineage graph in the region you pick. Column values, row contents and query result sets are not retained at any point in the pipeline.

Every deployment is pinned to one region at provisioning time: eu-west-1 (Dublin), eu-central-1 (Frankfurt), us-east-1 (N. Virginia) or ap-southeast-2 (Sydney). Regions cannot be changed afterwards without a full workspace migration, which our support team runs inside a scheduled maintenance window.

Metadata we collect

  • Query text and execution statistics from warehouse history tables
  • Table, view and column names, types and comments
  • dbt manifest and run-results artefacts, if you connect a project
  • Dashboard, tile and report identifiers from connected BI tools
  • Workspace user identities: email address, display name and role

What we never collect

  • Row-level data from any table, view or materialisation
  • Query result sets, cached or otherwise
  • Warehouse credentials beyond the scoped read-only role you grant

Retention and deletion

Lineage snapshots are kept for 90 days on Team plans and 400 days on Enterprise plans. Audit events — logins, permission changes, connection edits and policy approvals — are kept for 24 months on every plan and cannot be shortened, because they form the evidence trail behind SOC 2 and ISO 27001 reporting.

When a workspace is deleted, its metadata is soft-deleted immediately, removed from primary storage within 30 days, and purged from encrypted backups within 90 days.

Deletion is irreversible once the 30-day grace period closes. Export anything you need for evidence requests before you remove a workspace.

Access controls

Access to production is granted through short-lived, role-scoped credentials issued by our identity provider. Every production session is recorded, and no engineer holds standing access — elevation requires an approved change ticket and expires after four hours.

  1. All access is authenticated with hardware security keys. Passwords alone are not accepted.
  2. Production changes require review from a second engineer on the owning team.
  3. Quarterly access reviews are run by the security team and signed off by the CTO.

Sub-processors

We publish the full sub-processor list, including the purpose, region and data category for each one. Customers on any plan can subscribe to change notifications and receive 30 days' notice before a new sub-processor is added.

Note

Customer-managed encryption keys are available on Enterprise plans only. Talk to your account team before enabling them — rotation cannot be reversed without a re-encryption window.

Important

Deleting a workspace purges its lineage history after a 30-day grace period. Export any evidence you need for audits before the workspace is removed.

Implementation

From read-only credential to live coverage in three weeks

Every account runs the same sequence with a solutions engineer. Most teams are enforcing their first freshness contract on day two.

  1. Step 01

    Connect the warehouse

    Grant a read-only role in Snowflake, BigQuery, Redshift or Postgres. Northlane reads query history and metadata — no agents to install, no VPC peering, no copies of your data.

    Day 1 · 20 minutes

  2. Step 02

    Mark the tables that matter

    Tag the 30 to 50 tables your revenue, product and finance reporting actually depends on. Column-level lineage is traced upstream from there, including through dbt models and reverse-ETL syncs.

    Week 1

  3. Step 03

    Write the first contracts

    Declare what good looks like per table — arrival window, row-count band, null and uniqueness thresholds — then set who is paged when the contract fails. Templates cover the common shapes.

    Week 2

  4. Step 04

    Go live with routing

    Breaches open in Slack and PagerDuty with the failing column, the upstream job and every affected dashboard already attached. Nothing else changes in your stack.

    Week 3 · Ongoing

What you can expect

How mccloudCo actually works

Not projections — the commitments built into every engagement, regardless of practice or tier.

  • 68 %

    Fewer pipeline incidents

    Measured against the 90 days before rollout.

    42 accounts · platform telemetry

  • $ 4.2 M

    Warehouse spend recovered

    Idle compute and duplicate models retired automatically.

    Median enterprise account, FY25

  • 99.98 %

    Freshness SLA held

    Across 12,400 monitored production tables.

    Rolling 12-month average

  • 9 ×

    Faster root-cause analysis

    Alert raised to owner-confirmed cause.

    Customer survey, n=310

Platform capabilities

Everything between the warehouse and the dashboard

Northlane connects to your warehouse in read-only mode and maps every table, column and job in under an hour. Lineage, monitoring and access policy then come from one graph instead of three tools.

The principal consultants

The people behind the platform

Forty-one people across nine countries, building the data operations layer we wanted back when we were the ones carrying the pager.

Marisol Okonkwo, VP of Data Platform at Northlane

Marisol Okonkwo

VP of Data Platform

Lisbon

Runs the platform group. Nine years on warehouse infrastructure before joining Northlane in 2021, most of it spent on the query layer nobody wanted to own.

Every incident we cannot explain is a bill somebody pays two quarters later.
Devan Ashworth, Head of Product at Northlane

Devan Ashworth

Head of Product

Toronto

Owns lineage and governance. Built the incident tooling that three of our design partners were already using before Northlane existed.

Governance only works when it is the fastest path through the system, not the slowest.

Customer stories

What teams shipped after the guesswork left the pipeline

Every story below is a production deployment with numbers the customer signed off on. No pilots, no anonymised logos.

Fintech 480 employees

Halden Systems

Four hand-run reconciliation jobs became one governed pipeline, and the books closed in three days instead of eight.

63% faster close

Healthcare 1,200 employees

Ardent Labs

A decade of untracked SQL moved onto column-level lineage without pausing a single downstream dashboard.

1,400 models migrated

Logistics 260 employees

Corvane

Freshness checks run themselves, so nobody triages “is this dashboard stale?” tickets any more.

6 hrs back each week

Get started

Senior academic and technology leadership, without the cost of a full-time hire

Book a no-cost consultation, and we'll talk through your organization, your goals, and your budget cycle — then recommend a starting point, or tell you honestly if now isn't the time.

No cost to talk. No obligation to sign. No pressure either way.

  • 5 Tiers in the shared advisory ladder, from a light diagnostic to embedded leadership
  • 2 Practices — Education and Strategic Advisory — one advisory model
  • 6 –12 mo Minimum term, then month to month, with no automatic renewal into another long commitment