Data nutrition, Feedback Loops and Building Value Creation Engines

Q&A with Simon Williams, Partner and Founder, WovenLight, a UK-based private equity co-sponsorship firm committed to unlocking value in firms by helping them navigate the challenges of AI transition. WovenLight is a partner of IPEM's Value Creation Summit, taking place in Paris on 25th September 2025. 

IPEM: What is the most misunderstood aspect of value creation in your area of expertise today?

SW: “Garbage in/garbage out” is often misunderstood in the wider context of operational value creation. People assume there’s nothing they can do and it becomes an excuse for inaction. Clearly having good data is better, but the reality of all firms is that “not all data is created equally”. It’s better to start with what you’ve got and iterate from there.  

How we gracefully cope with less than perfect data is how we make better decisions and improve performance. The ‘garbage in/garbage out’ mantra frames the challenge as a technical issue, rather than unlocking its true superpower, which is that it provides a feedback mechanism.

Eating junk food is not going to lead to good outcomes. Similarly, not all data is good for you. What’s important is understanding the quality of the data so that you can use it appropriately, in the same way that we think about the nutritional benefits of the food we consume. It comes down to being transparent on what’s good and what’s bad.

‘Data nutrition’ allows you to get started with what you have and build transparent feedback loops that enable you to learn faster; much like the way neural networks learn to improve through reinforced learning. While competitors chase model accuracy, winners build feedback loops that capture real-world outcomes and iterate continuously.

IPEM: Where are you seeing the biggest gap between boardroom ambition and execution on the ground?

SW: Boards often mandate ambitious targets but abdicate responsibility to the IT organisation. This misses the point. The biggest challenge is not technology but the firm’s culture, processes and ways of working.

Boards need to absorb the complexity of their business and transmit clarity on the firm’s North Star. In practice, this means being crystal clear on what the value creation levers are, how much they are worth, why they’re sequenced in that order. It means accepting that the talent profile has probably changed. And it means building the governance framework that embraces feedback loops to learn faster.

IPEM: What capability or mindset shift will define the next generation of value creation?

SW: I think the defining shift will be from ad‑hoc analytics pilots to systematic, compounding infrastructure of performance improvement.

Leaders will build codified, repeatable templates that scale across multiple deals, embedding learning loops in every capital deployment. This requires a mindset that views AI not as technology or research, but a strategic asset. This capability will require a different talent profile, with inter‑disciplinary teams combining investor discipline, domain knowledge, and data expertise, supported by agile governance frameworks enabling rapid iteration and impact measurement. Firms embracing this mindset shift - treating their data teams as value creation engines and feedback loops as performance multipliers - will pull ahead.

IPEM: If you could embed one question into every due diligence process, what would it be?

SW: “Can this thesis be operationalised through a repeatable, measurable template that contributes to compounding value across future assets?” 

This single query tests whether the data, talent and infrastructure are present or buildable, and whether the hypothesis can translate into measurable EBITDA uplift. It moves firms away from one‑off experiments toward designing value creation engines with multi‑deal repeatability; essential for consistently generating premium returns.

IPEM: How do you think the operational value creation playbook might evolve in the coming years?

SW: The value creation playbook will evolve from looking like bespoke consulting projects into scalable, replicable infrastructure with consistent outcomes.

At WovenLight this is powered by three engines; the first being configurable value-creation templates, the second being a compounding data platform that captures ‘outcome telemetry’ and the third, rapid diagnostics to radically reduce the cost and time of thesis development and diligence.