Minimum Viable Operations: Why 70% Capacity is More Valuable Than Perfect Recovery



Imagine this: your most important IT system fails. The cloud provider has issues. Or ransomware. Or something unexpected.

The classic question is: "How quickly can we recover?"

The better question is: "How can we keep working while we recover?"

That is the difference between Disaster Recovery and Minimum Viable Operations.

The problem with perfect redundancy

Everyone wants 100% availability. Sure.

But let’s talk about costs.

To make a system truly 100% redundant, you need:

  • Duplicate infrastructure (second data center, second cloud provider)
  • Active real-time data replication
  • Automated failover mechanisms
  • Regular testing of failover scenarios
  • Staff who know both environments

For a mid-sized company with CHF 50M in revenue? That quickly costs CHF 200K-500K per year. Per critical system.

And here is the uncomfortable truth: most companies do not have one critical system, but five or ten. Perfect redundancy for everything is simply unaffordable.

The 70% rule

Here comes the pragmatic approach:

70% capacity within 30 minutes protects more revenue than 100% capacity after 48 hours.

Let that sink in for a moment.

If you need 48 hours during an outage to get back to 100%, you generated zero revenue for 48 hours. For an e-commerce company with CHF 180K daily revenue, that is CHF 360K in losses.

If instead you can switch to 70% capacity within 30 minutes, you may lose CHF 15K during the switch and then continue operating at reduced capacity.

That is MVO: not perfect, but functional.

What is Minimum Viable Operations?

MVO is the answer to the question: "With what reduced capacity can we continue working while we solve the actual problem?"

It is not about keeping everything running perfectly. It is about defining the minimum that keeps business operations going.

For each critical process, you define:

The beauty of it: MVO forces you to distinguish between nice-to-have and must-have. This clarity is also valuable in normal operations.

How to define MVO: 4 steps

Step 1: Identify critical processes

Do not start with IT. Start with the business.

Ask yourself:

  • Which processes generate revenue directly?
  • Which processes have regulatory consequences if they fail?
  • Which processes affect customers directly?

Typical candidates:

  • Order intake and processing
  • Production (for manufacturing companies)
  • Payment transactions
  • Customer service (for critical inquiries)
  • Supply chain coordination

If you have already done a Business Impact Analysis, use it as the basis.

Step 2: Define the minimum

For each critical process: what is the absolute minimum needed to keep working?

Specific questions:

  1. Minimum capacity: At what percentage of normal capacity can we operate without suffering existential damage?
  2. Nice-to-have vs. must-have: Which features are comfort, which are existential?
  3. Manual workarounds: If system X fails, can we continue manually? With which tools? For how long is that sustainable?
  4. Priority under resource constraints: If we have only 50% of staff, who works on what?

A practical example:

Production at an industrial company:

  • Normal: 1000 units/day with fully automated manufacturing
  • MVO: 700 units/day with reduced quality inspection (only level 1+2, not level 3) and manual packaging
  • What is dropped: Level-3 quality inspection, automated packaging
  • What remains: Core production, basic quality control
  • Acceptable duration: 5 days

Step 3: Develop fallback options

Now it gets concrete: what is Plan B if the primary method fails?

Fallback options template:

Questions for each fallback option:

  1. Alternative technology: If system A fails, which other platform can we use?
  2. Alternative suppliers: If supplier X cannot deliver, is there a supplier Y?
  3. Manual processes: Can we temporarily work manually?
  4. Reduced-feature mode: Can we continue with 70% functionality?
  5. Activation time: How quickly can we switch? Who decides?

Step 4: Test regularly

An MVO plan that is never tested is just hope.

Define test scenarios:

Important in testing:

  • Tests should be announced (no surprise), but with real time pressure
  • Document: what worked? What did not?
  • After each test: update the plan, improve fallback options
  • Also test: who makes decisions? Who communicates?

Our article Business Impact Analysis: Identifying critical business processes offers a deeper look.

Practical example: MVO in action

An e-commerce company with CHF 45M annual revenue (~CHF 180K daily revenue). The main shop runs on AWS.

The starting point:

  • Shop on AWS EU-Central-1
  • No fallback options defined
  • In case of AWS outage: total outage

After defining MVO:

Fallback strategy:

  • Option 1 (Activation: 30 min): Emergency landing page on Azure with "Order by phone now" + hotline expansion. Enables reduced order processing.
  • Option 2 (Activation: 2 hrs): Basic shop on Azure with reduced features (main products only, no customization). Enables ~70% of normal capacity.

Preparation costs:

  • Emergency landing page + Azure setup: CHF 15K one-time
  • Annual maintenance: CHF 5K
  • Hotline expansion agreement: CHF 3K/year

Result during the first outage:

  • AWS outage lasted 6 hours
  • After 30 min: hotline active, emergency page online
  • After 2 hrs: basic shop on Azure
  • Revenue loss: CHF 15K (instead of an estimated CHF 45K in a total outage)

ROI: The CHF 23K investment paid for itself threefold in the first incident.

The mental shift

MVO requires a rethink for many executives.

Old mindset: "In an outage, we must get back to 100% as quickly as possible."

New mindset: "In an outage, we must continue operating as quickly as possible—even if it is not perfect."

That feels uncomfortable at first. No one wants to tell customers, "Our shop is currently available with limitations." But it is better than "Our shop is not available at all."

Common objections (and why they are not true)

"Our customers expect 100% availability."

Do they? Or do they expect you to communicate transparently when problems occur and still offer a solution? Experience shows: customers are surprisingly tolerant of temporary limitations—if you communicate proactively and offer an alternative.

"Manual workarounds are too error-prone."

Yes, they are more error-prone than automated systems. But an error-prone process that runs is better than a perfect process that does not run. You are not planning MVO for permanent operations—but for bridging a disruption.

"This is too much effort to document."

A 5-page MVO plan tested quarterly beats any 50-page plan that is never tested. Keep it simple. One table per critical process is enough.

The short version

  • MVO answers: With what minimum capacity can we continue working?
  • The 70% rule: 70% immediately protects more revenue than 100% after 48 hours
  • Nice-to-have vs. must-have: MVO forces clarity
  • Fallback options: Define Plan B (and C) for every critical process
  • Test, test, test: An MVO plan without testing is just hope

What now?

Take one critical business process. Just one.

Answer these questions:

  1. What is normal capacity?
  2. What is the absolute minimum (MVO)?
  3. What can be dropped?
  4. What must remain?
  5. How long is MVO acceptable?

If you have thought this through for one process, you have understood the principle. Then do the same for the remaining critical processes.

(And if you realize you need support to work through this systematically for all critical processes—we are here for that.)

Further reading