Summary
This page sets out how drinkYangu plans to test whether a pen-sized electrochemical probe, paired with a phone app and a database of genuine signatures, can tell genuine whisky from counterfeit in real Kenyan conditions. We have not yet run these experiments. This is our plan, grounded in published research. We'll update this page with results, good or bad.
Background
Counterfeit and illicit alcohol is a public-health and economic problem in Kenya. The Anti-Counterfeit Authority estimates fake alcohol at about 21% of the market by value. An industry-commissioned study claims around 60% of alcohol consumed is illicit, a broader category that includes unbranded brews, and public-health groups dispute that figure (Movendi International). Counterfeit and illicit brands (refilled bottles, cloned labels) are a distinct part of that market (The Star, 2026). They are the part a brand-matching tool can address.
Today, a drinker's only checks are the label, the seal and the price. We want to give them a check on the liquid itself.
Prior research
Whisky authentication is well studied (Okolo, Kilcawley & O'Connor, 2023). The studies most relevant to us:
| Study | Technique | What it showed |
|---|---|---|
| Novakowski, Bertotti & Paixão (2011) | Voltammetry, disposable copper and gold electrodes | Told expensive whiskies from cheap ones, and detected adulteration with a copper electrode alone |
| Wójcik et al. (2019) | Differential pulse voltammetry, iridium and platinum electrodes | Metal electrodes in a simple electrolyte distinguished wines and Scotch whiskies |
| Eluceda E-Sens (2021) | Electrochemical, custom catalytic electrodes | A device comparing samples to a database of authentic whisky, with results in minutes |
| Martins et al. (2017) | UV–Vis spectroscopy (optical) | Over 98.6% correct for genuine and 93.1% for counterfeit samples across seven blends |
| Macias et al. (2019) | Plasmonic "artificial tongue" (optical) | Distinguished whiskies, including by age and cask, with over 99% accuracy |
The optical studies set a high bar for accuracy, but they rely on lab optics. The electrochemical studies show that simple, cheap electrodes can carry enough information to discriminate. That trade-off is exactly what a consumer pen needs.
Our approach
- Probe. Low-cost electrodes in a replaceable pen tip, driven by a miniature potentiostat that runs a voltammetric sweep.
- App. The pen sends the voltammogram to a phone over Bluetooth. The app handles the product lookup (the user scans or selects the bottle) and shows the result.
- Signature database. Fingerprints of genuine product, measured from bottles sourced directly from brand owners or verified supply chains, across several batches of each product.
- Verdict. A statistical model compares the sample with that product's signature and returns match, no match or inconclusive, with a confidence level. "Inconclusive" is a valid answer. We'd rather say "we can't tell" than guess.
Target metrics
| Metric | Target |
|---|---|
| Time from dip to verdict | Under 30 seconds |
| Sample needed | A few drops |
| Genuine bottles wrongly flagged | Under 5% in the field pilot |
| Counterfeits wrongly passed as genuine | Under 5% in the field pilot |
| First database | The whisky products most commonly sold in Kenya |
Roadmap
- Lab feasibility. Benchtop potentiostat, commercial electrodes, genuine and deliberately adulterated samples. Question: are the fingerprints stable and separable?
- Signature database. Reference samples for the first whisky products, across batches, with partner brand owners.
- Pen prototype. Miniaturised electronics and replaceable tips, tested against the benchtop results.
- Field pilot. Supervised testing with retailers and bars, blind samples included.
- More categories. Other spirits, once whisky performance is proven.
Open questions
- How much do fingerprints drift with electrode wear, temperature and handling?
- How well can we catch genuine whisky that has been diluted or blended with a cheaper whisky?
- How many batches per product do we need to capture natural variation?
- What's the lowest-cost electrode that still separates products reliably?
How partners can help
- Brand owners: genuine reference samples and batch information.
- Universities and labs: instrument time and electrochemistry expertise.
- Regulators: testing protocols and access to confirmed counterfeit samples.