Question: Would You Burn €250k to Build or Spend €25k to Test a Digital Transformation?

In today’s fast-paced business environment, the allure of digital transformation—especially AI implementation—makes many business leaders ready to make substantial investments. 

But is pouring resources into unproven ideas the wisest approach?

My bet is no, but let’s explore it further.

The 70% Failure Rate Problem with Digital Transformations

This year, we visited BCG Singapore to discuss digital transformations.

Apart from the cool office (kudos to the office managers and everyone involved in supporting it :)), we discussed what they see with their accelerator program.

According to BCG X’s digital transformation report, a staggering 70% of digital transformations fail. 1

This statistic isn’t just a number—it represents countless businesses that have invested significant capital into initiatives that didn’t deliver the expected returns.

At this point, I’m facepalming. This is huge capital going to waste. If I had a magic stick and only one achievement in my life – I would focus on fixing this leaky bucket.

Why does this happen? 

Often, it comes down to a mindset issue, particularly among successful business owners.

The Success Paradox

Recently, I spoke with a Singapore-based business owner running an 8-figure operation. 

Having reached such an impressive scale, there’s a natural tendency to believe “he has made it” and “knows it all.” 

Previous success creates dangerous blind spots that are paradoxical:

• Passion-blindness
• Confirmation bias
• Resistance to challenging established ideas

This executive was prepared to deploy €250,000 for the vaguely defined goal of “Let’s enhance our product with AI…” 

A classic example of what I call the worst decision ever. 

BCG is debating that to succeed, established business owners need:

1. A strategy with clear goals and business outcomes (the why, what, and how)
   
2. Commitment from the CEO to middle management for accountability (middle managers are often overlooked, but are we surprised?)
   
3. Deploying high-caliber talent and freeing up resources (no, don’t outsource to the lowest bidder…)
   
4. Addressing roadblocks quickly and adapting to contexts and missions (yes, lack of hands-on support can make or break things)
   
5. Monitoring progress toward outcomes with clear metrics

Usually, business owners don’t have that. Their innovation trigger is FOMO. They hire a freelancer instead of promoting internally. They don’t engage at all, underinvest and at the end ask, “Where’s the 10x return?”?

The mindset shift is to go from “Let’s achieve major outcomes with minor resources” to “Let’s secure major wins with major dedication”. 

Everyone wants to win big with less effort, but this approach positions initiatives as a form of gambling.

Civilizations are not built on gambling with their future. 

Imagine Julius Caesar employing third-grade generals for his conquests.

However, Julius Caesar didn’t live in the AI era, so we can examine how the United States Marine Corps handles AI.

Download the document and scroll to the appendix. Here are my two highlights:

What do you take from this military initiative? Do you think they outsourced it to a low-budget freelancer?

The Real Cost of AI Implementation

The true cost of AI isn’t primarily in development—it’s in identifying the right problem to solve. 

Without this crucial first step, even the most sophisticated AI solution becomes an expensive solution searching for a problem.

Digital transformation and AI implementation costs extend far beyond initial development:

• Initial implementation costs (hardware, software, integration)
  
• Development costs and specialized talent (which is scarce and expensive)
  
• Ongoing maintenance and model updates as data and company needs evolve
  
• Security and regulatory compliance costs, especially when handling personal data
  
• Organizational culture and structural adaptation costs
  
• Risk of failed implementation without clear strategy, resulting in wasted resources
  

A Better Approach: Test Before You Invest

Instead of committing €250k upfront, here’s the strategy I recommended:

1. Start with €25k to test AI capabilities and validate assumptions. 
2. Invest in team development to build internal expertise. Example: Shift deadlines and give everyone a day off to watch an AI webinar
3. Gather user feedback early in the process
4. Foster a collaborative culture across departments
5. Deeply understand market needs before committing to solutions
6. Prioritize features solving real problems, not just impressive tech
7. Plan for scalable growth (remember, anyone can build an AI demo now)
8. Use budgets to reduce risk, not create technical debt

The remaining budget should focus on improving business-as-usual operations—the foundation that supports all innovation.

I think digital transformations are really about psychology. 

Research shows four additional psychological factors driving resistance:

• Plan continuation bias: Persisting with flawed strategies despite evidence of failure
  
• Loss aversion: Employees perceive potential losses from change as more significant than gains, even when benefits outweigh risks.
  
• Status quo bias: 63% of employees prefer inefficient familiar processes over untested digital solutions.
  
• Authority bias: Overreliance on leadership decisions without critical evaluation of digital tools4.
  

We can incorporate behavioral nudges like pilot programs with quick wins to demonstrate value early, reducing perceived risk.

All of those biases lead to friction.

Companies commonly fall victim to four specific blind spots that undermine digital transformation

1. Strategic friction spots: Leaders often fail to see opportunities for collaboration, even with competitors, to create network effects
   
2. Cultural friction: Organizations focus on “how things are done” but neglect the rewards and consequences that reinforce desired behaviors.
   
3. Human capital friction: Having appropriate talent policies to attract and retain the right people is crucial, and compromising on hiring standards to fill positions quickly often leads to failure4.
   
4. Personal friction: Leaders must proactively understand the range of skills, behaviors, and digital tools required in the digital age and be willing to ask for help4.
   

But are you ready? Academic research emphasizes structural readiness:

• Digital maturity assessment: Only 12% of companies evaluate legacy system compatibility before transformation, leading to 68% integration failures[2][3].
  
• Talent fragmentation: 74% of failed projects lack collaboration between IT and operational teams3.
  
• Dynamic capability building: Successful firms reallocate 30% of transformation budgets to upskilling programs in AI literacy and data governance9.
  

Framework Suggestion: Remember the military? They use the ADKAR model (Awareness-Desire-Knowledge-Ability-Reinforcement), which increases adoption rates by 50%

The Benefits of the Test-First Approach

I think there’s a common favoritism around “big bang” success and the desire for home runs.

But more often than not, small wins are the things that move the needle.

Smart investments in business expansion will:

• Dramatically increase success probability.
  • You want “small bets” with a higher likelihood of a favorable outcome, not “big bets” with unclear results. I also love black swan events, but those can swing your business a lot, so it’s better to hedge.
• Enable effective innovation.
  • Do the right “it”. Nobody likes being lost in a stormy sea.
• Build team capabilities and confidence.
  • Actually this is not a sprint, it’s a marathon. After a successful 5 year digital transformation do you know what comes next? Another wave and congrats to those who like surfing.
• Provide clear ROI metrics
  • I prefer to aim to 10x things; it cannot happen if we do not measure.

The Bottom Line

Would you rather burn €250k on an unproven concept or spend €25k to test your hypothesis and make informed decisions about the remaining €225k?

I will always choose testing hypotheses first. What about you? Join the discussion on linkedin