Quick Navigation
I spent the first decade of my career as a pure discretionary investor. It didn't go as planned. My buys were solid, but my sells were emotional. I'd hold a falling stock because I 'believed in the story', then panic-sell at the bottom. After a particularly brutal year, I started tinkering with basic ranking models in Excel. That was my first taste of quantamental investing, and honestly, it felt like cheating at the time. Now I run a quantamental process for a living, and I'm convinced it's the only way to stay sane in today's market. Let me show you what this actually means and how to build it without a data science team.
What Is Quantamental Investing?
Quantamental investing sits at the intersection of quantitative analysis and fundamental investing. Instead of letting a model run the show on autopilot or relying entirely on gut feel, you encode fundamental insights into measurable factors and then let a systematic engine rank, score, and monitor opportunities.
The idea sounds simple, but the execution is nuanced. You might decide that 'value' means a company has a low EV/EBITDA relative to its sector. You also think 'quality' means high return on invested capital and stable margin trends. These are fundamental judgments, but you turn them into structured variables and score every stock in your universe.
I've seen many investors confuse quantamental with 'quant lite'. It's not just adding a few screens to your stock list. It's a feedback loop where your fundamental thesis is constantly tested against data. The best quantamental frameworks respect both sides: the numbers challenge your story, and your story gives context to the numbers.
Why Combine Quant and Fundamental?
Each approach has a fatal flaw. Pure quant funds often ignore narrative and context, which leads to buying good-looking stocks that collapse because the underlying business is deteriorating. Pure fundamental investors suffer from behavioral biases: confirmation bias, recency bias, and plain overconfidence.
Quantamental investing solves these problems in a specific way. For example, your model might flag a stock as 'attractive' based on cheap valuation and rising revisions. But without a human check, you might buy into a value trap like a company with shrinking cash flows and a broken balance sheet. That's where your fundamental intuition steps in. Conversely, your gut says 'this company is a great long-term compounder', but the model notes that momentum has been negative for six months and insiders are selling. That data check prevents you from catching a falling knife.
A key non-consensus point: the main value of quantamental doesn't come from generating alpha through the quant model alone. It comes from reducing your personal return drag. By forcing your fundamental ideas into a structured format, you automatically eliminate the worst emotional decisions. That's a hidden source of return.
How to Build a Quantamental Framework
There's no one-size-fits-all blueprint, but the steps below are the backbone of every serious quantamental process I've seen. I've also included the mistakes I made along the way, so you don't have to.
Step 1: Define Your Investment Universe
Start small. If you're a retail investor, don't try to scan 6,000 US stocks. Focus on a sector you genuinely understand, like technology or healthcare. My own universe is around 300 large-cap European names, which I know deeply from a business perspective. If you force yourself to only pick from that basket, the model's output becomes easier to sanity-check.
Step 2: Identify Data-Driven Factors
You need factors that reflect a fundamental edge. Classic examples are valuation (price-to-book, forward P/E), quality (ROIC, gross margin stability), and growth (revenue growth, earnings revisions). But don't blindly use the same popular factors everyone else uses. If you're using Bloomberg terminal factors, notice that they're highly crowded. Instead, tilt toward factors that match your own investing philosophy. For me, I love a factor that tracks free cash flow durability, which isn't in many standard kits.
Step 3: Build a Scoring Model
This is the hardest part, no matter how simple it sounds. Take each factor and rank your small universe from 1 to 100. Combine the ranks with weights that reflect your conviction. Let's say 40% value, 40% quality, 20% earnings revisions. The output is a composite score for each stock.
I can tell you from experience that the first version of your model will be terrible. Don't over-engineer it. Start with five factors max. If you can't explain in one sentence why a factor matters, drop it. My original model used twelve factors and the result was worse than a four-factor one. The value of simplicity is real.
Step 4: Backtest, But Don't Overfit
Yes, you need to test your model on historical data. But I beg you: don't try to squeeze the last basis point of performance. That's how you end up with a strategy that worked beautifully in backtest and fails immediately live. Use a long sample, like 20 years, and be brutal about transaction costs and short-selling constraints.
There's a moment when the backtest looks too good. That's a red flag. It means your model is likely recognizing noise or data mining. The best quantamental approaches have a modest but stable backtest, not a home-run chart.
Step 5: Add a Human Override
Finally, create a rule book for overrides. The model gives you a top 10 list every month. Your job is to veto two or three names based on fundamental context. But you need clear rules for vetoing, so you don't fall back into subjective chaos. My rule is simple: I can only veto a stock if I can cite a specific event (e.g., a regulatory probe or a broken acquisition) that the model can't capture. I can't veto based on a vague 'I don't like the CEO'. That discipline has saved me from my own biases more than once.
Data Sources Every Quantamental Investor Needs
If you're serious about this, you need reliable data. The raw numbers for trading and financial statements are non-negotiable. Beyond that, here's what I actually use:
- Financial statements and earnings calls: You can't build quality factors without clean accounting data. Services like S&P Capital IQ or Refinitiv are solid, but try to verify raw numbers against the 10-K filings occasionally.
- Analyst estimates and revisions: If earnings revisions are your factor, you need timestamped estimate data. Institutional Brokers' Estimate System (IBES) has been a standard for this. Even a feed from Zacks can work.
- Alternative data: Cautiously explore satellite imagery or web-scraped pricing trends. I've found these most useful as a confirming signal rather than a standalone factor.
One thing that surprised me early on was the data quality trap. Many cheap data feeds have survivorship bias or lag in updating delisted companies. If possible, use a dataset that includes point-in-time data. It will lower your backtest performance, but it'll be honest.
Pitfalls That Ruin Quantamental Strategies
Every quantamental investor hits the same five walls. Let me save you some pain.
Overfitting to the past: You change a factor weight so that the model catches the 2008 financial crisis correctly. Then it misses 2020. That's not science, that's history painting. Keep your weights stable for at least a year, and only change them if you can articulate a structural reason for the change.
Ignoring capacity and liquidity: A model might pick a tiny stock that you can't sell quickly. For a retail portfolio, that's fine, but if you're running a fund, you'll have to screen for liquidity. I once lost a full month of gains because I owned a microcap that the model loved and the market ignored. Nobody cared enough to buy it.
Disconnecting from fundamentals: The entire point of quantamental is to stay aware. When you simply copy your backtest and let it run, without reading any company updates, you've become a dumb quant fund. You'll miss fast-changing situations time and time again.
Misunderstanding factor cycles: Factors like 'value' and 'momentum' go through extended dry spells. If your model is value-heavy, it will underperform during growth bubbles. That's normal, but many people panic and change models at the worst time. You need a pre-defined benchmark and a long-term commitment.
Using the same screening threshold forever: Markets change. A PE of 10 might be cheap in one regime and expensive in another. Quantamental should adapt to current conditions. I recalibrate my factor z-scores every quarter to maintain consistency.
Quantamental vs. Traditional Fundamental and Pure Quant
The table below summarizes what I've learned after living through all three styles.
| Aspect | Fundamental | Pure Quant | Quantamental |
|---|---|---|---|
| Decision driver | Judgment, narrative | Mathematical models | Enriched judgment with data |
| Behavioral bias | High | Low | Medium-low |
| Adaptability | High | Low | Medium |
| Implementation burden | Moderate | High | Medium |
| Best for | Illiquid markets, deep value | Large liquid equities, diversified | Active managers who want discipline |
Notice that quantamental isn't the 'best' at anything, but it's the most practical for a human who wants to leave emotional pitfalls behind. It respects your experience while forcing you to be honest about your skill.
FAQ: Rapid Answers for Practitioners
I'm a solo investor with just Excel and public data. Can I run a quantamental framework?
Yes, and that's exactly how I started. You don't need Bloomberg or fancy Python code. Download 10 years of financial statements from free sources like SEC EDGAR, build a ranking formula in Excel, and run a basic backtest using a simple stock screener. A framework with 30 stocks and 4 factors already beats nothing. The discipline matters more than the tooling.
How do I avoid overfitting in my quantamental model?
Limit your degrees of freedom from day one. Set your factor weights based on your investing worldview, not historical optimization. If you must optimize, use only 5 years of data for testing and keep the other 15 years as a out-of-sample validation. When the out-of-sample results are poor, reject the strategy entirely. Trust me, a strategy that barely works but is robust will serve you far better than one that shines in-sample and collapses after you deploy.
What's your honest opinion on using alternative data in a quantamental process?
Alternative data is vastly overrated for most investors. It's expensive, noisy, and doesn't convert to reliable alpha for small players. I use it only to confirm or reject a hypothesis built from fundamental signals. For example, when satellite images showed a retailer's parking lots empty, that validated the negative earnings revision signal. That's a practical use, not a standalone factor.
How much time does a quantamental process actually take per month?
Plan for two solid days per month, not every day. Day one is for updating data and running the model. Day two is for reading the full financial reports of the top 10 names, then applying your human override. My own routine: run the model on the first weekend, then review the shortlist on Monday. This schedule is sustainable and doesn't devour your life.
Is quantamental investing suitable for options or futures trading?
Yes, but the concept shifts from equity analysis to its derivative overlays. You can apply quantamental logic to options by using implied volatility (quantitative) and earnings projections (fundamental) to pick option structures. For futures, the fundamental part might be macro supply-demand data, which gets blended with a momentum or carry factor. The framework works, but you need to adapt the factor definitions sharply.
Quantamental investing isn't a fad. It's the logical response to a market where your competitors are increasingly data-driven. If you can pair your underrated fundamental skills with a simple, robust scoring engine, you'll build a repeatable process and stop relying on luck. Start with a small universe, respect your own judgment, and let the numbers keep you honest.
