The Scale of the Problem
Consider a single lasagne. It might contain 35 or more distinct components — pasta sheets, multiple cheeses, minced beef, tomatoes, onions, garlic, herbs, seasoning blends, béchamel sauce ingredients, and packaging layers. Each component has its own supply chain, agricultural practices, processing steps, and transport routes. Conducting a rigorous manual Life Cycle Assessment for this single product typically takes four to six weeks of expert time.
Now multiply that by a product portfolio of 500 or 1,000 SKUs, and the impossibility of manual LCA at scale becomes clear.
The Data Granularity Problem
Not all tomatoes are equal. Tomatoes grown in open fields in southern Spain have a fundamentally different environmental profile from those produced in heated greenhouses in northern Europe. The energy source for those greenhouses, the irrigation method, the fertiliser regime, and the transport distance to the processing facility all matter.
Manual LCA struggles with this granularity because:
- Sourcing varies seasonally — the same product may use ingredients from different origins at different times of year
- Supplier-level data is inconsistent — some suppliers provide detailed information, others provide almost none
- Proxy data choices introduce bias — the analyst must decide which database entry best represents each ingredient, and those choices compound across a complex recipe
Methodological Consistency
When multiple analysts work on different products in a portfolio, methodological drift is almost inevitable. Key areas where inconsistency creeps in include:
- Allocation decisions: How do you split the environmental burden of a dairy farm between milk and beef? Different analysts may apply economic allocation, mass allocation, or system expansion — each yielding different results.
- System boundary choices: Does the assessment include retail refrigeration? Consumer cooking? End-of-life packaging treatment? Inconsistent boundaries make cross-product comparisons meaningless.
- Database version mismatches: If one product was assessed using Ecoinvent 3.8 and another using 3.9, the underlying data differs even for identical processes.
Scalability and Modelling
Automation addresses these challenges by encoding methodological decisions once and applying them uniformly across every product in the portfolio. A well-designed automated system:
- Maps each ingredient to the most appropriate background dataset based on origin, production method, and processing
- Applies consistent allocation rules, system boundaries, and impact assessment methods
- Handles multi-level recipes (a sauce within a lasagne within a meal deal) without losing granularity
- Runs sensitivity analyses across the entire portfolio simultaneously
Sensitivity Analysis at Scale
One of the most powerful advantages of automation is the ability to run sensitivity analyses across hundreds of products in minutes rather than months. This reveals which assumptions matter most and where primary data collection would most improve result quality.
For example, automated sensitivity analysis might reveal that for 80% of your chilled ready meals, the single largest driver of uncertainty is the origin of the dairy ingredients — directing your data collection efforts precisely where they will have the greatest impact on result confidence.
Manual vs. Automated LCA: A Comparison
| Dimension | Manual LCA | Automated LCA |
|---|---|---|
| Time per product | 4-6 weeks | Minutes to hours |
| Methodological consistency | Analyst-dependent | Systematically enforced |
| Portfolio coverage | 10-50 products feasible | 1,000+ products feasible |
| Sensitivity analysis | Rarely performed | Standard practice |
| Update frequency | Annual at best | Continuous |
| Cost per product | High | Dramatically lower at scale |
Frequently Asked Questions
Does automation mean less accurate results?
No. Automation applies the same scientific methodology as manual LCA but does so consistently and at scale. The quality of results depends on the quality of the underlying data and the rigour of the methodological framework, not on whether a human or a system performs the calculations.
What about primary vs. secondary data?
Automated systems can incorporate primary data wherever it is available and fall back to high-quality secondary data where it is not. The key advantage is that as primary data becomes available from suppliers, it can be integrated across all affected products simultaneously.
How does automation help prevent greenwashing?
By enforcing consistent methodologies, maintaining full audit trails, and enabling portfolio-wide sensitivity analysis, automation makes it much harder to cherry-pick favourable assumptions or selectively report results. Every product is assessed on the same basis.