What is a stock screener?
A stock screener (or stock scanner) is a filtering tool that automatically scans a universe of listed stocks and keeps only those that satisfy a set of criteria defined by the investor. The term comes from to screen — to sift, to filter.
Without a screener, analysing 300 stocks by hand would take weeks. With a well-configured screener, you get a shortlist of 5 to 20 stocks worth a deeper look in just seconds.
Screeners operate on two broad families of criteria:
- Fundamental criteria: financial data drawn from the balance sheet and income statement (P/E, EV/EBITDA, revenue growth, EBITDA margin, net debt…)
- Technical criteria: market data (price, volume, RSI, position relative to moving averages, relative performance…)
The most sophisticated screeners combine both families into a composite score — that is the approach taken by Screener Small Caps.
The essential fundamental criteria
Here are the indicators to include in any serious fundamental screener, grouped by analysis pillar:
These four equally weighted pillars (25 pts each) form the basis of the 0-100 composite score of Screener Small Caps. Academic research — notably the effectiveness of the P/E and the persistence of momentum — justifies never relying on a single criterion.
The 5 classic mistakes with a screener
1. The single filter — the value trap
Filtering on a P/E < 10 or an EV/EBITDA < 5 alone is the recipe for value traps. A low valuation can reflect structurally zero growth, unsustainable debt or a sector in decline. Always cross-reference valuation with growth.
2. Ignoring liquidity on small caps
A screener that surfaces a small cap with a perfect score but only €5,000 of daily trading is unusable. On Euronext Growth, some stocks go whole days without a single trade. The 20-day average volume should always be part of the criteria for investors with positions above €2,000.
3. Confusing TTM and annual data
Financial data can be expressed on a TTM basis (Trailing Twelve Months — the last 12 rolling months) or as the latest annual close. For small caps that report half-yearly, the gap can be significant, especially after an acquisition or disposal.
4. Failing to normalise by sector
An EBITDA margin of 8% is excellent for a food retailer but mediocre for a SaaS company. Without sector normalisation, a screener compares apples and oranges. Relative sector scoring is a key feature of Screener Small Caps.
5. Backtesting on survivorship data
If you test your screener criteria on the history of currently listed stocks, you mechanically exclude every bankruptcy and delisting — survivorship bias. The results always look excellent. An honest backtest reconstructs the universe as it existed at each historical date.
A 5-step workflow for using a screener
Define the universe
Choose your scope: Euronext Growth, a broad pan-European small-cap index, or a specific sector. The more focused the universe, the more relevant the screener.
Apply elimination filters
Rule out the obvious negatives first: market cap < €10M, insufficient daily volume, negative equity, net debt/EBITDA > 5x. These filters cut the universe by 40 to 60%.
Score the survivors
Compute a composite score on the remaining stocks. Weight valuation, growth, strength and momentum according to your style (value investor vs growth vs GARP).
Build a shortlist of 10-20 stocks
Take the top of the ranking. A shortlist of 10 to 20 stocks can be analysed in depth over a week. Beyond that, the quality of individual analysis drops.
Individual qualitative analysis
The screener opens the door — it does not replace reading the annual report, checking the business model, analysing the competition and understanding governance.
Comparison: screeners available for European stocks
- Powerful interface
- Many filters
- Free (limited)
- Incomplete European data
- No Euronext Growth
- No European sector normalisation
- Broad coverage
- Free
- Easy to use
- Limited filters
- No composite score
- Data sometimes lagged
- Excellent technical criteria
- Global coverage
- Real-time alerts
- Shallow fundamental data
- Paid for the full version
- No ML score
- 800+ stocks on Euronext Growth & Access
- ML score 0-100 (4 pillars)
- Daily updates
- Public track record
- European sector normalisation
- Universe limited to European small caps
Screener + Machine Learning: the qualitative leap
Traditional screeners apply fixed rules: "P/E < 15 AND growth > 10%". These thresholds are arbitrary and fail to capture the interactions between variables. A P/E of 14 with 8% growth may be better than a P/E of 12 with 5% growth — but no binary filter detects this.
Machine Learning models — notably ensemble methods such as XGBoost — learn these interactions from thousands of historical observations. Instead of fixed rules, they compute an expected-outperformance score that incorporates all the variables simultaneously.
| Approach | Logic | Advantage | Limitation |
|---|---|---|---|
| Fixed filters | Binary rules (yes/no) | Simple, transparent | Arbitrary thresholds, no interactions |
| Weighted composite score | Sum of normalised scores | Combines several criteria | Fixed, non-adaptive weights |
| Supervised ML | Learning from history | Interactions, adaptation | Overfitting risk, partial black box |
Screener Small Caps combines all three layers: elimination filters + rules-based composite score + XGBoost ML prediction. The public track record lets you judge the model's added value: every prediction logged before market close, results net of costs, with failures published alongside the successes.