If you're starting a stock portfolio in 2026 and you've already read 10 generic "how to invest" articles, here's something different: a sequenced 6-step playbook with specific position sizes, specific names to consider, and specific rebalance rules.
This isn't financial advice. It's a structured operating manual for building a value-investing portfolio from scratch, using the framework invest-like applies to every stock. Every step has a worked example.
Step 1: Decide your time horizon and stick to it
Before you buy anything, answer:
- Will I need this capital in the next 3 years? If yes → don't put it in individual stocks. Use a high-yield savings account or short-duration bonds.
- Will I need it in 3-7 years? Mostly index funds + a small individual-stock allocation (maybe 20-30% of the equity portion).
- Will I not touch it for 7+ years? Individual stocks make sense alongside index funds. The longer the horizon, the more you can absorb single-stock volatility.
Value investing requires a 7-10 year minimum horizon. If you can't commit to that, buy an index fund and stop reading. There's no shame in this - index funds beat 85%+ of actively-managed funds over 30-year periods.
Step 2: Allocate base index exposure first
Even if you're going to pick individual stocks, your first move should be a low-cost broad-market index fund. The math:
- 50-70% of equity allocation in a broad-market index fund (VOO, VTI, ACWI, or your local equivalent)
- 30-50% in individual stocks picked through the framework
This is the "core-satellite" approach. The index core gives you market exposure without single-stock risk. The individual-stock satellite is where you can outperform if your picks are good.
The reason: even Buffett has acknowledged that most retail investors should be primarily indexed. Berkshire's annual letter has repeated this point for 20+ years.
Step 3: Build a starting universe of 30 candidate stocks
Before picking, build a list of 30 candidate stocks you'd consider owning. The criteria:
- Pass at least 5 of 7 value-investing frameworks on invest-like
- Have business models you understand
- Are not in industries you have strong reservations about
A starter list might include (each passes 5+ of 7 frameworks):
- Payment networks: Visa, Mastercard, American Express
- Software franchises: Microsoft, Adobe, Oracle, ASML
- Credit-rating duopolies: Moody's, S&P Global
- Healthcare giants: Johnson & Johnson, Becton Dickinson, Stryker
- Consumer staples: Procter & Gamble, Coca-Cola, Costco, Walmart
- Industrial process oligopolies: Linde
- Diversified compounders: Berkshire Hathaway, Markel
- High-quality tech: Alphabet, Apple, Amazon
That's ~20 names. Add 10 names from sectors you understand specifically (your domain expertise).
Step 4: Start with a 5-stock conviction book
For the individual-stock satellite portion, start with just 5 names. Equal-weighted 20% each (or weighted to your conviction level). The rationale:
- 5 stocks is enough diversification to absorb single-stock blowups
- 5 stocks is few enough that you can actually research each one deeply
- Position sizing of 20% per stock is meaningful - small enough that no one stock destroys you, big enough that good picks materially compound
Build the 5-stock starter from the highest-conviction names on your candidate list. Run each through the Boardroom debate on invest-like to stress-test your thesis with the four legendary investor perspectives.
Example starter book (illustration only):
- 20% Visa (V) - payment network compounder
- 20% Johnson & Johnson (JNJ) - diversified healthcare
- 20% Microsoft (MSFT) - software franchise
- 20% Moody's (MCO) - credit-rating duopoly
- 20% Costco (COST) - membership economics
All five pass 5+ of 7 frameworks. All five have published 60+ year (or in the case of newer companies, 25+ year) track records of value creation. The portfolio would have generated returns close to or above the S&P 500 over the last 10 years.
Step 5: Add positions only as your conviction grows
Don't expand from 5 to 20 stocks overnight. The mistake retail investors make: they get bored holding only 5 names, so they add positions on weaker conviction.
The right pace: add one new position per quarter, only when:
- You've done deep research on the new name (read the 10-K, ran it through Buffett-Brain, opened the Boardroom)
- The framework consensus is genuinely 6-of-7 or 7-of-7
- You have free cash to deploy (don't sell existing positions to fund new ones unless the existing position has degraded)
Over 4 quarters → 9 stocks. Over 8 quarters → 13 stocks. Over 16 quarters → 20+ stocks. By the time you have 20 names, each was added with research, not impulse.
Step 6: Rebalance and review annually, not weekly
Once per year (pick a date - Jan 1, your birthday, whatever):
- Re-score every position. Open each /buffett/[ticker]/ page. Compare current framework consensus to when you bought.
- Identify positions where framework has degraded. Anything that has dropped 2+ tiers (e.g., from 7-of-7 to 4-of-7) needs investigation.
- Decide if degradation is structural or temporary. Structural → trim or exit. Temporary → hold.
- Identify opportunity-cost rotations. Are there 7-of-7 names you don't yet own that you could trade into?
- Rebalance to target weights. Trim positions that have grown to >25% of book; top up positions that have shrunk.
That's it. One annual sit, 2-3 hours, more rigorous than what most retail investors do.
The rest of the year, do not look at your portfolio more than once a month. Daily checking is the single biggest cause of poor returns. The market is noise on a daily timescale; signal on a 7-year timescale. Train yourself to ignore the daily.
Common mistakes to avoid
-
Overdiversification. 30+ positions = you're basically an index fund. If you want to be an index fund, just own an index fund (lower cost, lower work).
-
Underdiversification. < 5 positions = single-stock risk is too high. Even the best stock-picker has a 20%+ chance of catastrophic loss on any one name.
-
Chasing yield. The high-dividend names (T, VZ, MO) are usually high-yield because the dividend is at risk. See our dividend-trap explainer for the patterns.
-
Buying what's hot. NVDA at $148 in 2026 might be a great business; the framework consensus at 2-of-7 says the price is wrong. Don't buy "what everyone's talking about" without applying the framework.
-
Trading frequently. Each trade has a transaction cost + tax cost + emotional cost. Buffett's holding period is "forever" specifically because trading destroys returns.
How invest-like helps at each step
- Step 1 (horizon): not a tool decision, that's you and your bank account.
- Step 2 (index allocation): not on invest-like, but we recommend Vanguard ETFs for the index portion.
- Step 3 (candidate list): open /strategies/ - every framework lets you screen the universe by its own criteria. Build your candidate list in minutes.
- Step 4 (5-stock starter): open /buffett/[ticker]/ for each candidate. Compare 5-pillar scores. Run the Boardroom debate. Make your call.
- Step 5 (adding positions): subscribe to the weekly Buffett-Gift newsletter (we send 1 high-conviction pick per week with full reasoning).
- Step 6 (rebalance): each stock page shows the historical scoring versions, so you can see how the verdict has changed since you bought.
Disclosure
Educational tool. The specific stocks named (Visa, J&J, Microsoft, Moody's, Costco) are illustrative only, not recommendations. Past performance does not predict future returns. Position sizing, rebalance cadence, and time horizon decisions are personal and should be made in consultation with a qualified financial adviser.
Author: Zaid Ghazal, founder of invest-like, indie SaaS, Kiel, Germany. Not a registered investment adviser.