Stelios Tzellos | Why Pharma Companies Keep Overcomplicating Commercial Analytics
Stelios Tzellos
Every few years, pharmaceutical companies convince themselves that the next analytics platform, AI model, or data integration system will finally solve their forecasting and strategy problems. Then the forecasts still miss, the launches still underperform, and leadership still wonders what went wrong.
The issue usually isn’t the technology. It’s the thinking.
Stelios Tzellos of the UK has worked across oncology analytics, forecasting, and commercial strategy long enough to see multiple waves of pharmaceutical analytics transformation projects. Most promised clarity. Many created more complexity.
More Data Doesn’t Automatically Create Better Decisions
The pharmaceutical industry now has access to enormous amounts of information: prescribing data, claims data, conference coverage, real-world evidence, biomarker prevalence, physician surveys, competitive intelligence feeds, and advanced analytics platforms capable of modelling almost anything.
Yet many commercial teams feel less certain about decisions than they did a decade ago.
Why? Because more information often increases the number of possible interpretations.
At GlobalData, Tzellos worked on oncology forecasting projects where the challenge was not finding enough data but identifying which data actually mattered. The strongest analyses were usually the simplest ones: clear assumptions, transparent logic, and focused decision-making frameworks.
Complexity Creates False Confidence
One of the risks in pharmaceutical analytics is that highly complex models can appear more credible simply because they look sophisticated.
A forecast with dozens of variables and intricate statistical assumptions may impress leadership visually while hiding fragile logic underneath. Simpler models sometimes perform better because decision makers actually understand the assumptions driving them.
During his time at IQVIA, Stelios Tzellos worked in oncology disease insights and the Analytics Center of Excellence, where pharmaceutical clients increasingly sought advanced analytical capabilities. The most effective client work often involved simplifying the problem rather than adding more variables.
Oncology Makes This Harder
Oncology markets are inherently complex. Biomarker-driven patient segmentation, evolving treatment algorithms, combination regimens, and rapid clinical innovation all create uncertainty.
That complexity is real. But adding analytical complexity on top of biological complexity doesn’t necessarily improve strategic clarity.
The organizations that perform best commercially are usually the ones that identify the few variables most likely to drive market behaviour and focus relentlessly on understanding them.
Analytics Should Support Decisions, Not Replace Them
At AstraZeneca, where Tzellos works across business insights, analytics, and oncology marketing, analytics functions are most valuable when they help leadership understand uncertainty rather than pretending uncertainty can be eliminated.
A good analytics team clarifies the strategic tradeoffs. It identifies the assumptions with the highest risk. It frames the decisions leadership actually needs to make.
What it should not do is overwhelm decision makers with unnecessary complexity disguised as precision.
The Human Element Still Matters
Stelios Tzellos built his career by combining scientific training with commercial analytics experience. His background in Molecular Biology at Imperial College London taught him rigorous thinking. His consulting and industry experience taught him something equally important: models are only useful if people can interpret and act on them effectively.
The future of pharmaceutical analytics will not belong to the companies with the most complicated systems. It will belong to the companies that use analytics to think more clearly than their competitors.