Twenty-two guides in, you have the individual lenses. Valuation multiples. Rule of 40. Free Cash Flow Growth. Debt-to-Growth. Volatility Survival. Hidden Gems vs Value Traps. Institutional 13-F Tracking. Each one answers a specific question about a specific stock. This guide is different. It answers the meta-question that all the others feed into: how do you turn 22 lenses into an actual portfolio?

The framework in this guide is deliberately conservative because non-dividend investing rewards patience and punishes over-sizing. Get the shape right and the individual stock picks matter far less than the standard retail wisdom suggests.

Step 1 — The total allocation ceiling

Before picking a single ticker, decide how much of your total portfolio can defensibly sit in the non-dividend universe. The math is straightforward:

Max non-div allocation = your drawdown tolerance % ÷ 2

If a 30% total-portfolio drawdown is the emotional pain threshold you can genuinely sit through without capitulating, the non-dividend basket ceiling is 60% of the portfolio — because that basket could realistically draw down 50-60% in a bad macro year. Younger investors (20+ year horizon) can flex this upward; investors within 5 years of retirement should flex it downward because sequence-of-returns risk is now working against them.

Step 2 — The 70 / 20 / 10 core-satellite split

Within the non-dividend allocation, split into three buckets by conviction and thesis risk:

Step 3 — Position count discipline

Between 8 and 20 individual non-dividend positions works for most retail portfolios. The floor of 8 keeps you diversified enough that a single blow-up does not sink the ship. The ceiling of 20 acknowledges that you realistically cannot actively track more than that number of individual theses at retail scale. Beyond 20, move the marginal exposure into a broad growth ETF (QQQ, VUG, MGK) where the diversification happens automatically and you stop pretending to have conviction on names you do not actually follow.

Step 4 — Per-position sizing

The critical rule: size each position such that a 60% drawdown on that position is emotionally survivable without panic-selling. The math:

Max position size = per-position pain threshold ÷ 60%

A retail investor with $250,000 in the portfolio who can emotionally tolerate a $10,000 loss on a single position can size any single non-dividend name up to roughly $16,700 — about 6.7% of the total portfolio at entry. Core compounders can push toward that ceiling; satellites should sit noticeably below; speculative bets should be sized so a full write-down is not portfolio- threatening. See the Volatility Survival guide for the full drawdown-tolerance framework.

Step 5 — Correlation, not ticker count

The most common retail construction mistake is confusing ticker count with diversification. Owning Meta, Netflix, Amazon, Nvidia, and Tesla is five tickers but effectively one factor bet on megacap growth. Real non-dividend diversification comes from mixing categories:

• Platform megacaps (AMZN, GOOGL, META, MSFT if you exclude the dividend): mature moats, cash-flow machines.
• Elite Rule-of-40 SaaS (PLTR, CRWD, ADBE non-div-adjacent): efficiency + growth composite.
• Capital-intensive growth (RIVN, TSLA, LCID early): capex-heavy ramps where debt-to-growth is the primary risk.
• Consumer / delivery platforms (UBER, DASH): unit-economics-turning stories.
• AI / semiconductor cycle (NVDA, AVGO): concentrated exposure to a single capex cycle.
• Specialty pharma / clinical biotech (VKTX, BEAM, CRSP): binary outcomes with cash-runway protection.

A five-position portfolio drawn from five different categories is meaningfully more diversified than a ten-position portfolio drawn from a single category. The correlation matrix is what saves the portfolio during a factor drawdown.

A worked example

Imagine a $250,000 portfolio for a 35-year-old investor with a 30% drawdown tolerance and a 25-year compounding horizon. The construction sequence:

1. Total non-div ceiling. 30% pain tolerance × 2 = 60% ceiling → up to $150,000 in non-dividend stocks.
2. Split. 70/20/10 → $105K core, $30K satellites, $15K speculative.
3. Position sizing. $10K per-position pain tolerance ÷ 60% drawdown assumption = $16.7K position cap. So core positions run $10-16K each ($105K ÷ ~8 core = ~$13K average). Satellites run smaller ($4-7K each). Speculatives run smallest ($1-3K each).
4. Category distribution. Aim for at least one position each in platform megacap, Rule-of-40 SaaS, capital-intensive growth, consumer platform, and (optionally) semiconductor cycle. Reject any candidate that would push a category above 25% of the non-dividend allocation.
5. Filter using Capital Reinvestment Score. For each candidate ticker, verify the score is in the right band for the intended bucket. A “core” candidate that reads WATCH is either downgraded to the satellite bucket or dropped entirely.

Example core positions using live scores

Example satellite candidates

Examples of positions that should not be in the portfolio

Rebalancing discipline

A non-dividend portfolio drifts far faster than a dividend portfolio because the individual positions move so much more. A position that starts at 6% of portfolio can easily become 12% after a 100% return (which is not uncommon in this universe). At that point, the position is running larger than the sizing framework called for and needs to be trimmed. Three rebalance rules that work in practice:

1. Position-cap trim. When any single position exceeds 1.5x its intended sizing, trim back to 1x. Sell the excess into a broad ETF or into rebalancing the underweight positions.
2. Verdict downgrade trim. Any previously-BUY-rated position that migrates to WATCH gets downgraded to the satellite bucket regardless of entry conviction. Any WATCH-to-AVOID migration is a full exit.
3. Correlation trim. Recompute the category distribution quarterly. Any category exceeding 25% of the non-dividend allocation gets trimmed. This is how you enforce diversification against factor drift — it does not happen automatically.

Common pitfalls

• The concentration trap. Loading up on five megacap names because they feel safest, then discovering during a factor drawdown that “safest” did not mean uncorrelated.
• The story trap. Owning a AVOID-rated name because the story is compelling. The composite score exists specifically to override narrative conviction when the numbers disagree.
• The averaging-down trap. Doubling down on positions that have moved against you without first checking whether the forensic-safety picture has changed. See the Hidden Gems vs Value Traps guide for the check.
• The over-diversification trap. Owning 30+ individual positions and mistaking scattered exposure for research-backed conviction. Consolidate or move to an ETF.

FAQ

Is a non-dividend portfolio riskier than a dividend portfolio?

Structurally, yes. Non-dividend stocks are priced primarily on future growth, which means they carry more discount-rate risk and typically deeper drawdowns in bear markets. In 2022 the marquee non-dividend names drew down 55-84% peak-to-trough (META -76%, NFLX -76%, PLTR -84%, TSLA -74%, NVDA -66%, AMZN -56%). The tradeoff is that the recovery-plus-growth over the following two years produced returns that decisively beat dividend indices. Non-dividend portfolios reward patience and tolerate volatility, and punish over-sizing. The construction rules in this guide are designed to make the drawdowns survivable.

What percentage of my portfolio should be in non-dividend stocks?

A defensible framework: your non-dividend allocation should not exceed your genuine drawdown tolerance times two. If you can emotionally sustain a 30% total-portfolio drawdown without panic-selling, and you assume the non-dividend basket could draw down 60% in a bad year (a reasonable historical assumption), then non-dividend stocks should not exceed 50% of the portfolio. Younger investors with 20+ years to compound can safely run higher allocations because the recovery cycle has time to play out. Older investors approaching retirement should meaningfully lower non-dividend exposure regardless of conviction, because sequence-of-returns risk is now a real headwind against high-drawdown assets.

How many non-dividend positions should I hold?

Between 8 and 20 in most retail portfolios. Below 8, you are running concentrated risk — a single blow-up wipes out a meaningful percentage of the entire non-dividend book. Above 20, you have not enough conviction in each individual name to justify separate ownership rather than an ETF. Institutional non-div funds typically run 25-50 positions but they have research teams that can keep active track of that many; retail investors realistically cannot. If you find yourself owning 30 non-dividend positions, either consolidate to your highest-conviction 15 or move a portion to a broad growth ETF (QQQ, MGK, VUG) to let the diversification happen on autopilot.

How should I split between core compounders and speculative bets?

A useful shortcut: 70/20/10. Roughly 70% of the non-dividend allocation in high-quality core compounders (BUY verdict on the Capital Reinvestment Score, Altman Safe Zone, Beneish clean, positive Operating Momentum). Roughly 20% in growth-with-flags names (WATCH verdict, one signal amber, high-quality thesis). Roughly 10% in speculative pre-revenue or turnaround stories where you accept a full write-down probability of 30-50%. The exact split flexes with age and drawdown tolerance, but the shape holds: most of the money is in businesses that will still exist and compound in a decade, small amounts in bets that could double or die.

How do I use the Capital Reinvestment Score to pick core positions?

The Capital Reinvestment Score ranks every non-dividend name in the universe on a 1-100 scale using five weighted pillars (Free Cash Flow, Balance Sheet, Efficiency, Growth, Forensic Safety). A rank of 85+ with a BUY verdict pill puts the ticker in the top ~15% of the non-dividend universe. That is where core compounders live. Uber currently reads 96.1, Palantir reads 96.7, Nvidia reads high enough to be actively BUY-rated. WATCH-rated names in the 60-85 band are for satellite positions where you accept more thesis risk in exchange for a possibly higher entry-point return. AVOID-rated names (below ~40) should not be in the portfolio at all unless you have a very specific catalyst thesis and a small position size to match.

What is the biggest portfolio-construction mistake retail investors make with non-dividend stocks?

Owning too many correlated positions. Every retail investor who owned Meta, Netflix, Nvidia, Tesla, and Amazon in early 2022 discovered in real time that these five names moved together during the growth-stock re-rating. A "diversified" portfolio of five megacap growth names is really one concentrated bet on the megacap-growth factor. Real diversification within the non-dividend universe means owning across categories: platform megacaps, high-Rule-of-40 SaaS, capital-intensive growth (RIVN, Tesla-earlier-years), specialty pharma, and one or two idiosyncratic single-story names. The correlation matrix is what saves the portfolio during a factor drawdown — not the ticker count.

Try it

Open the Dashboard and open the Capital Analytics tab on 5-10 non-dividend names you currently own or are considering. Note the score and verdict for each. Sort them mentally into the three buckets (core / satellite / speculative). Compare the actual allocation you already run against the 70/20/10 target. Most retail non-dividend portfolios end up heavier in satellites than they should be — because satellite theses feel more exciting than boring compounders. The rebalance is usually the highest-value adjustment you can make in a single session.

For the underlying signals this construction framework uses, see the Hidden Gems vs Value Traps and Volatility Survival guides.

This guide is educational. Portfolio construction rules are frameworks, not personalised advice. Real portfolios should account for tax situation, retirement horizon, liquidity needs, and existing holdings that this guide does not address.