By Marco G. Pistrin, Founder and Managing Director, AquaOrganica SRL, Italy
Let’s be clear: Microsoft Excel is arguably the greatest tool ever gifted to the aquaculture industry. It has built more farms, balanced more budgets and tracked more biomass than any “AI-powered” startup could dream of. A well-constructed macro is a work of art, a digital tapestry that brings order to the chaotic slurry of a working fish farm. But there is a growing, dangerous trend in our sector: the rise of the “Digital-First Farmer.”
We’ve all seen the pitch decks. An “Aquaculture Engineer” presents a 50-tab workbook that predicts harvest yields to the third decimal point for a site that hasn’t even broken ground. In the sterile vacuum of a spreadsheet, oxygen is almost always hovering around 9.0 mg/L, FCR is a beautifully optimistic and solid 1.1, and mortality is a polite, scheduled event. It is a world where the water is always crystal clear, the pumps never lose prime, no uninvited catastrophic events are considered and mass balances always seem to add up. We won’t even mention the omni-present “Aquaculture Engineer” who seemingly just cuts and pastes other engineers’ unverified and potentially incongruent data and claims it for their own.
The Seductive Certainty of the Cell
The problem isn’t the software; it’s the hubris of the inputs. Excel allows us to build a “Baseline of Truth,” which is obviously a great starting point and is vital for securing capital, predicting all expenditures and planning logistics. However, when we start layering multiple data sources, often pulled from different latitudes, species strains or outdated white papers, we run the risk of creating a “Frankenstein Model” of sorts. I have seen with my own eyes; melding feed growth trial data for Atlantic salmon, regurgitated from a feed company, paired it with the energy costs from a Mediterranean grid and applied it to an Arctic charr land-based facility in the desert in the Middle east. Imagine the potential problems with this? Yikes!
But hey, the maths checks out. The cells turn a reassuring shade of forest green. The Internal Rate of Return (IRR) looks “bankable.” But in the real world, biology doesn’t follow a linear VLOOKUP. Biology is messy, stubborn and remarkably indifferent to your pivot tables. When an “Aquaculture Engineer” purports to have the definitive solution in hand, we must be critical and look deeper into the data and the surrounding methodology. Otherwise, we may run the risk of relying too much on hyper-polished data, hypothetical fluff, conjecture and exaggerated and/or ideal culture parameters.
The Table of Truth
To bridge the gap between the investors and the farm we need a reality check. Here is how the “Standardised Projection” usually meets the “Biological Reality”:
| Parameter | Spreadsheet Assumption (Theoretical) | Empirical Biological Reality |
| Growth Period | Constant Thermal Growth Coefficient (TGC) resulting in a fixed harvest date. | Variable growth due to metabolic depression, social hierarchy and photoperiod shifts. |
| FCR (Feed Conversion Ratio) | Fixed at 1.1 based on laboratory-grade feed trials. | Real-world FCR of 1.25+ due to feed fines, settling velocity issues and erratic appetite. |
| Mortality Distribution | Linear 0.5 percent monthly attrition rate for financial amortisation. | Stochastic “spike” events; 90 percent of losses often occur in 5 percent of the production cycle. |
| Oxygen Demand | Static BOD/NOD values based on average biomass weight. | Diurnal spikes; oxygen demand peaks post-prandial (after feeding) and during temperature shifts. |
| System Reliability | 98 percent runtime based on manufacturer mean time between failures (MTBF). | 80–85 percent effective runtime due to biofouling, scaling, and sensor calibration drift. |
The “Standardised” Trap
The danger of these standardised projections is that they strip away the nuance of the local environment. Every farm has its own set of unique variables: the specific mineral profile of the bore-hole water, the local haulier’s reliability or the way the wind whips across the site in February. Yet the Excel Architect often treats these as rounding errors.
When we present these exaggerated models to investors, we create a “Success Gap.” When the real-world results inevitably deviate from the “Perfect Spreadsheet,” the technology gets blamed, the farm manager gets sacked and the investment capital dries up. We are setting ourselves up for failure by refusing to incorporate a foundational layer of realism.
Closing the Laptop
Realism is not the enemy of innovation; it is its safeguard. To overcome the folly of “Excel Farming,” we don’t need to delete the software; we need to stress-test the soul of the model. We need to move away from the “Best-Case Scenario” and start embracing stochastic modelling, calculating for the “What Ifs” and the “God Forbids.”
We need to build pessimistic buffers into our formulas. We need to stop treating biology as a sub-component of mechanical engineering. It isn’t! Most importantly, we need to ensure that the person building the spreadsheet has spent enough time on a farm to know that a 0.05 percent change in any one parameter can be the difference between a profitable harvest and a pile of compost.
Excel is an excellent place to start, but it’s a terrible place to live. If your farm only works on a screen, it doesn’t work at all. It’s time to stop worshipping the spreadsheet and start listening to the fish.
