top of page

You Do Not Land the Plane to Upgrade the Wings

  • Feb 13
  • 4 min read

Imagine being told your aircraft needs new wings.

Not next year.

Not during scheduled maintenance.

Now, mid-flight.


You are cruising at 35,000 feet. The cabin is full. The engines are under load. There is weather ahead. You have passengers who expect to reach their destinations. Landing is not an option.

That is what enterprise data platform modernisation really feels like.


From the outside, transformation programmes look controlled and linear. A new platform is selected. Architecture is redesigned. Data is migrated. A cutover date is agreed. The old system is retired. It looks impressive in presentations.


In reality, the business does not pause while you modernise. Revenue still goes to the Board. Forecasts still guide investment decisions. Operational dashboards still drive daily execution. Regulators still expect consistency. Commercial teams still expect answers by 9am.


You do not land the plane. You replace the wings while flying.


The Real Risk Isn’t Downtime

When people think about modernisation risk, they imagine outages. Systems down. Dashboards blank. Headlines about disruption.


In practice, the real risk is divergence.


The legacy platform reports £102.4m in revenue. The new platform reports £101.9m.

Everything loads. Nothing crashes. But there is a £500k gap.


Which number is correct? Which definition changed? Which filters are different? Which one goes to the Board?


No alarms are sounding, yet trust is lost immediately.


Modernisation rarely fails because the technology cannot cope. It falters when confidence in the numbers begins to fracture.


In aviation, asymmetry destabilises the aircraft.

In data, asymmetry destabilises leadership confidence.


Once executives begin to doubt, adoption slows. Parallel reporting persists. The new platform is treated as experimental rather than authoritative. The legacy system lingers because it feels safer.


The real challenge is not migration. It is preserving trust while foundations shift.


Why the “Big Switch” Is a Myth

There is a comforting fiction in many transformation programmes: the big switch.

A carefully planned weekend. Final loads complete. Old system off. New system on. Monday morning, the future has arrived.


Enterprise data estates rarely allow that simplicity.


Over years, they accumulate embedded logic, local workarounds, historic edge cases and implicit definitions that have become institutional truths. Downstream feeds connect to operational systems far beyond analytics. Regulatory expectations rely on reproducibility.

A single-step cutover compresses risk rather than eliminating it. Testing windows shrink. Reconciliation becomes rushed. Stakeholders are asked to accept outcomes that have not yet earned credibility.


Modernisation under operational pressure requires something less dramatic and far more disciplined.


Dual Running Is Engineering, Not Waste

In aviation, you do not remove a structural component and hope for the best. You test under load. You monitor stress. You validate stability before committing fully.

The equivalent in data transformation is dual running.


Two platforms operating in parallel. Metrics compared side by side. Capabilities migrated incrementally, with legacy components retired only once confidence transfers.


Dual running is sometimes described as duplication or temporary inefficiency. In reality, it is a deliberate trust bridge.


It creates space to reconcile differences properly. It allows business domains to move end-to-end rather than fragment by fragment. It makes variance visible before it becomes political.


Yes, it introduces cost and complexity. But it buys stability. And stability, in the middle of transformation, is not optional.


Migrate What the Business Uses

Another subtle mistake is measuring progress by technical artefacts.


Tables migrated. Pipelines rebuilt. Objects converted. Percentages complete.


It creates the appearance of motion. But businesses do not consume tables. They consume outcomes: revenue reporting, customer insight, supply chain visibility, marketing performance.


A table-led migration often results in fragments of functionality scattered across environments. Technically present. Commercially incomplete.


A capability-led approach works differently. It defines the business outcome first. It moves the full end-to-end capability. It reconciles commercially. It secures stakeholder sign-off. Only then does it begin retiring the legacy components behind it.


This may appear slower at first glance. In practice, it aligns progress with value and makes trust transfer measurable rather than assumed.


Tooling Enables. Discipline Protects.

Modern cloud platforms are powerful. Compute scales. Storage expands. Automation accelerates build.


Tooling, however, does not protect confidence.


The critical differentiator in modernisation is delivery discipline. Clear ownership of metrics. Explicit reconciliation thresholds. Structured variance investigation. These are not administrative details. They are critical success factors.


Without discipline, divergence becomes anecdotal. Definitions drift. Decisions are implied rather than explicit. Confidence erodes quietly.


With discipline, variance is expected, investigated and resolved systematically. Sign-off becomes credible. Adoption becomes durable.


In complex systems, rigour is not bureaucracy. It is protection.


Keep Flying

Changing the wings of an aircraft mid-flight sounds reckless.


In reality, it would be procedural, controlled and incremental. Load would be monitored continuously. Stability would take precedence over speed. The passengers would barely notice.


That is the real benchmark for successful data modernisation.


No dramatic cutover.

No prolonged instability.

No erosion of trust in the instruments.


The business keeps flying. Revenue remains credible. Forecasts remain reliable. Leadership confidence remains intact.


Modernisation is not a rebuild. It is a controlled transition carried out under operational pressure.


And the organisations that do it well understand a simple truth:


You do not land the plane to upgrade the wings.

 

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page