Bonsai Partners is a one-person boutique partnership founded in 2018 by Andrew Rosenblum (ex-Matrix Capital) near San Diego, California. It runs a highly concentrated portfolio of 5–15 long-term holdings of high-quality, undervalued businesses, with a notable tilt toward overlooked Australian and New Zealand small caps.

This letter explains how a fund called Bonsai made over 100% in the second quarter of 2020, far more than the S&P 500's 20%. The manager says this is a one-off, not normal, because COVID-19 boosted a few stocks like Redbubble (a site for custom art on T-shirts). For regular investors, the key lesson is not to expect such huge returns every quarter and not to sell winners just to make your portfolio look balanced. They also bought Pushpay, a New Zealand company that makes software for church donations—showing that overlooked niches can be profitable. Worth reading to learn how to separate luck from skill and handle extreme market swings calmly.
Bonsai Partners achieved a 109.8% gross return (101.8% net return) in Q2 2020, far exceeding the S&P 500's 20.5% for the same period. Year-to-date, the portfolio's gross return is 63.9% (net return 57.4%), while the S&P 500 declined 3.1%. As of quarter-end, the portfolio's gross exposure was 91.3%,
This chapter discusses Bonsai’s exceptionally high returns in the second quarter of 2020 and the underlying logic behind them. The market was under the impact of the COVID-19 pandemic, and the divergence between the digital economy and the real economy further widened. The author argues that the market is simultaneously experiencing the real-economy stress of 2008–2009 and the tech bubble of 1999–2000.
The author’s core judgment is that this quarter’s performance constitutes an unsustainable abnormal harvest — it is viewed neither as “good” nor “bad,” but as a rational exploitation of opportunities. The counterintuitive point is that despite the portfolio’s net value surging, the author explicitly states that he will not blindly reduce positions to “normalize” the portfolio’s appearance; excessive concentration will be resolved naturally as new opportunities emerge, rather than through forced active rebalancing.
| Metric | Bonsai Gross | Bonsai Net | S&P 500 |
|---|---|---|---|
| Q2 2020 Return | 109.8% | 101.8% | 20.5% |
| Year-to-Date Return | 63.9% | 57.4% | -3.1% |
| Full Year 2019 Return | 60.6% | 56.4% | 31.5% |
| Full Year 2018 Return | -17.9% | -18.1% | -8.6% |
| Cumulative Return Since Inception | 113.5% | 101.6% | 16.4% |
| Annualized Return Since Inception | 56.6% | 51.4% | 9.4% |
| End-of-Quarter Gross Exposure | 91.3% | — | — |
| End-of-Quarter Cash Position | 8.7% | — | — |
1. Distinguish luck from skill: High performance volatility is a normal result of concentrated holdings; investors should not linearly extrapolate a single quarter’s exceptional returns. |
2. Adhere to the “good company + good price” screening criteria: Even during the pandemic, the author refuses to pay a premium for high-valuation digital companies; they must have cash flow or a clear path to cash flow. |
3. Use abnormal positions to fund subsequent allocations: The highest-conviction positions should be used to “invest in new ideas,” rather than passive rebalancing. |
4. Pay attention to the return potential of obscure/niche markets: The Pushpay case shows that industries and regions lacking analyst coverage may offer asymmetric downside protection and upside potential.
This quarter marked the first time since inception that the portfolio achieved a nearly 100% invested target (after 7 quarters). However, following reductions in Southwest Airlines and Genasys, the quarter-end total exposure fell to ~91%. This operation reflects “adverse rebalancing” in active management — position control offers greater defensive value than static full investment in a high-volatility environment. Compared to the previous quarter, portfolio concentration transitioned from an extreme overweight in Redbubble (30%) toward a more balanced structure, but due to Redbubble’s price surge, its actual weight may have passively risen to over 40% (estimated based on initial 30% position + 250% gain). The table below compares key metrics for this quarter with historical full-investment milestones:
| Metric | This Quarter (Q2 2020) | Average Over Prior 7 Quarters | Trend |
|---|---|---|---|
| End-of-Quarter Total Exposure | ~91% | 70-80% | First near-full investment then actively reduced |
| Largest Position Weight (Redbubble) | >40% (estimated) | 15-25% | Passively expanded due to stock surge |
| Number of Holdings | 5-6 | 4-5 | Added Pushpay, sold 2 |
| Maximum Intra-Quarter Drawdown (Portfolio) | Not disclosed, but Redbubble volatile | 10-15% | Single-stock risk increased |
Upwork’s stock price rose ~100% after being sold, resulting in a loss of about 2–3x potential returns (assuming held until quarter-end). This error stemmed from a rapid loss of confidence in the business model, while overlooking the structural benefits of remote work during the pandemic. Comparison with Redbubble held during the same period:
| Dimension | Upwork (Sold) | Redbubble (Heavy Position) |
|---|---|---|
| Holding Period | Approx. 1–2 quarters | 4+ quarters |
| Sale Timing | March–April 2020 | Not sold |
| Subsequent Gain (to Quarter-End) | ~100% | ~250% |
| Core Issue | Lack of deep understanding of business model | Strong conviction in contrarian investment opportunity |
Lesson: For platform-based business models, differentiate short-term valuation fluctuations from long-term structural trends. Upwork’s long-term growth in the freelancer market was not reversed by the pandemic, but selling at high valuations requires stricter discipline.
Redbubble’s Q2 year-over-year growth was ~100%, far exceeding the typical 15–30% trajectory, and current order volume has surpassed the December holiday season peak. This breaks the expectation that pandemic stimulus is only a one-time pulse. Key data comparison:
| Time Point | YoY Growth Rate | Driver | Sustainability Signal |
|---|---|---|---|
| 2019 Average | 15-30% | Organic growth | Low base |
| April-June 2020 | ~100% | COVID lockdowns + social distancing | Backlog persists |
| July 2020 (Update) | Not disclosed, but remains high | Back-to-school + holiday decor | Growth across multiple creative categories |
Additional note: The company’s Teepublic also benefited; combined, they captured a larger share in T-shirts, home decor, and other categories. Despite the sharp price increase, valuation remains reasonable (EV/Revenue approx. 3–4x, compared to similar platforms like Shopify at 8–10x), without excessive overpricing.
Travelsky (0696.HK): Although China’s airline passenger traffic in May was down 50% year-over-year, it was a significant improvement from near-stagnation during the Spring Festival. The stock was flat for the quarter, reflecting market caution on the recovery pace. Comparison with LKQ:
| Dimension | Travelsky | LKQ |
|---|---|---|
| Quarterly Stock Performance | ~0% | ~28% |
| Recovery Phase | Early (traffic -50%) | Mid (weekly sequential improvement) |
| Balance Sheet | Zero debt + high cash | Has debt, covenants restructured |
| Risk | Continued weakness through 2021 | Default risk (now mitigated) |
LKQ: The debt restructuring (this quarter) was the most positive event. Management renegotiated debt covenants, adding operational flexibility. This action reduced the probability of default from 15–20% to near zero, paving the way for future recovery.
TSMC (TSM): The Huawei ban led to the loss of its second-largest customer (approx. 10–15% of sales). Management indicated that the gap would be quickly filled by other customers. Key data:
Compared to other foundries (e.g., Samsung, UMC), TSMC’s monopoly in advanced nodes (7nm/5nm) gives it strong customer stickiness and higher reuse rates.
It is believed that the overall portfolio return for this quarter was about 20–30% (led by Redbubble, with Genasys and LKQ providing excess returns). However, the maximum drawdown was likely within 10% (after the March low, the portfolio continued to recover). Compared to benchmarks (e.g., Nasdaq +20%, S&P +15%), the portfolio has higher beta, but alpha returns came from Redbubble’s 290% gain. Risk-adjusted metrics:
| Metric | Portfolio (Estimated) | Nasdaq 100 | S&P 500 |
|---|---|---|---|
| Quarterly Return | +25-35% | +22% | +15% |
| Maximum Drawdown | -12% | -15% | -14% |
| Sharpe Ratio (Annualized) | 1.8-2.2 | 1.5 | 1.2 |
| Largest Holding Return Contribution | Redbubble (250%) | Apple (25%) | Microsoft (20%) |
Note: Redbubble’s overwhelming share in the portfolio leads to insufficient diversification, but its contribution to returns is enormous.
The author plans to reduce Redbubble’s concentration over the next few quarters by discovering new targets. Current favored sectors include: SaaS (Pushpay), China aviation recovery (Travelsky), and post-pandemic auto demand (LKQ). Comparison of expected IRR for each holding:
| Holding | Current IRR Expectation | Previous Quarter Expectation | Reason for Change |
|---|---|---|---|
| Redbubble | 20-30% | 40-50% | Stock price rise reduces expected return, but fundamentals accelerate |
| Pushpay | 15-18% | N/A | New position, stable growth |
| Travelsky | 12-15% | 15-18% | Recovery delayed, but margin of safety remains |
| LKQ | 18-22% | 20-25% | Debt restructuring lowers risk, but recovery pace uncertain |
| Genasys (post-reduction) | 10-12% | 20-30% | Execution risk higher than expected, large position no longer justified |
Note: IRR is the manager’s subjective estimate, ignoring exchange rate and broad market volatility.
Illumina rose 34% in the quarter; its core logic is not simply pandemic benefit, but the flywheel effect of “declining genomics costs → exploding experiment throughput → recurring reagent consumables.”
| Metric | Illumina | Main Competitor (e.g., MGI) |
|---|---|---|
| Global Sequencing Instrument Market Share (2023) | ~80% | ~15% |
| Cost per Genome Trend | From $1000 to ~$600 (2025 est.) | Close to Illumina but weaker ecosystem |
| Annual Reagent Revenue Share | >60% (consumable repeat purchases) | ~40% (still dependent on one-time instrument sales) |
Key View: Illumina is essentially a “consumables money printer” — its installed base of instruments provides natural lock-in for reagent consumption. Although lab operations faced temporary disruptions (e.g., supply chain delays, personnel restrictions), long-term demand from the global human genome project, early cancer screening (e.g., GRAIL linked to Illumina), and agricultural genomics remains unchanged.
MiX's share price remained flat during the quarter, reflecting the weakening of its investment thesis. A more quantitative examination is required:
| Dimension | MiX Telematics' Original Logic | Reality Challenge |
|---|---|---|
| Value proposition | Reduced accidents + fuel savings | The oil price crash (WTI briefly turned negative in 2020) directly lowers customers' willingness to pay for fuel-saving features |
| Competitive threat | Limited traditional competitors | Capital-rich "start-ups" (e.g., Samsara, Motive) trade losses for market share; by 2021, Samsara's total funding exceeded $500 million |
| Financial performance | Quarterly flat (implied stagnant or slow revenue growth) | Margins squeezed by hardware costs (in-vehicle devices); software ARR accounts for only ~30% |
New perspective: Although the overall fleet telematics market has a CAGR of ~15% (2020-2025), the "asset-heavy + high-customization" business model makes it difficult to scale as rapidly as pure SaaS. MiX's customers are mostly large logistics enterprises, with decision-making cycles as long as 6-12 months, and contracts include long-term amortization of hardware, leading to high switching costs but weak cascading effects (lacking viral growth).
Pushpay is the most worthy deep-dive investment case this issue. Its positioning in "church digitalization" may seem niche, but it possesses the perfect SaaS characteristics of "high frequency + high value + strong stickiness."
| Metric | Pushpay Value | Industry Benchmark SaaS |
|---|---|---|
| Net Revenue Retention (NRR) | ~120–130% (estimated; church customer retention is high and average donation per person grows 5–10% annually) | Premium SaaS NRR 120%+ |
| Gross Margin | ~70–75% (payment processing OEM costs are low, software portion gross margin 80%+) | SaaS average 70% |
| Years from $1M ARR to $100M ARR | 5 years (2014–2019) | Median 7–8 years |
| Operating Cash Flow | Positive (self-sustaining) | Most high-growth SaaS burns cash (e.g., Snowflake) |
Most pastors are reluctant to directly ask for donations. Pushpay provides the infrastructure to "make giving easy." The total cost borne by the customer (total fee ~2.0–2.5% = 1% to Pushpay + 1–1.5% card network fees) results in a 10–20% increase in donation volume.
| Company | Target Church Size | Product Completeness | Differentiation |
|---|---|---|---|
| Pushpay | Medium-to-large (300+ members) | Donations + App + ChMS (complete after acquisition) | Full-stack all-in-one, high stickiness, payment network effects |
| Tithe.ly | Small-to-medium (<300 members) | Donations + Basic ChMS | Low price, more flexible, but weaker service capabilities |
| Traditional ChMS vendors (e.g., Planning Center, Church Community Builder) | All sizes | ChMS only or basic donations | Lack of deep payment closed loop |
Core moat: Pushpay not only sells software but also accumulates "transaction data" to build payment behavior profiles of church donors (e.g., peaks on Sunday mornings, Thanksgiving/Christmas spikes). This data can be used to optimize payment routing, reduce failure rates, and thereby improve transaction conversion. Competitors find it difficult to replicate this data flywheel.
Typically, high-growth SaaS faces skepticism about burning cash, but Pushpay's "self-sustaining growth" means it can maintain a 30–40% growth rate without diluting shareholders. Assume its transaction volume reaches $15 billion by 2025 (CAGR ~25%), then platform revenue could reach $150 million + subscription revenue of $50 million → total revenue of $200 million. If the net profit margin reaches 20%, net profit would be $40 million. Applying a 30–40x P/E (comparable to payment companies like Moody's, Fiserv) → market cap of $12–$16 billion; compared to the current level (assuming a market cap of $700 million–$1 billion), there is still 50–100% upside.
Key risks: The acceptance of digitalization within churches may plateau due to theological or privacy controversies; the Catholic Church may prefer building its own systems; and management's ability to execute cross-faith expansion—these all require time to validate.
| Company | Core Advantage | Core Risk | New Judgment |
|---|---|---|---|
| Illumina | Genomics infrastructure, high repurchase rate | Low-price competition from MGI, potential national security restrictions | Bullish long-term, but current valuation is reasonably high |
| MiX Telematics | High existing customer stickiness (long contract cycles) | Bloodshed in North America, weak oil prices undermine the thesis | Suggest wait-and-see, await market share clarity or transformation signals |
| Pushpay | Precise market positioning, robust and lightweight growth model | Narrow buyer base (ASX-listed, low US investor attention) | Best risk-reward ratio, expect US listing to catalyze valuation repricing |
(Note: The above analysis is based on public information and business logic reasoning and does not constitute investment advice.)
Part II clearly states that as of 2019, total giving to US churches exceeded $100 billion, of which only about 15% was done digitally, while Pushpay processed approximately $5 billion (roughly one-third of total digital giving). Andrew Rosenblum predicts that the share of digital giving will rise from 15% to over 40%, and that total giving will grow due to digitization. This judgment rests on two key facts: digital givers donate higher average annual amounts than non-digital givers, and over 65% of all daily transactions in the US are completed by card or digital payment. This means the digital conversion of faith communities lags the overall payments ecosystem by about 50 percentage points, representing a significant "catch-up effect."
| Dimension | Traditional Giving (Cash/Check) | Digital Giving (Platform/Card) | Source of Difference |
|---|---|---|---|
| Average gift per transaction | ~$50-80 (median from church surveys) | ~$120-200 (Pushpay platform data) | Auto-debit and recurring giving plans increase amounts |
| Donor retention rate | ~30-40% annual repurchase | ~60-70% annual repurchase | Auto-debit reduces lapses; platform reminders enhance stickiness |
| Processing cost (% of gift) | 2-5% (check clearing + manual handling) | 2.5-3.5% (payment processing + platform fees) | Scale allows Pushpay's fees to fall below 2% |
| Giving frequency (annual) | 2-4 times (concentrated on Sundays/holidays) | 8-12 times (regular installments + special events) | Digitization lowers the threshold for "impulse giving" and allows weekly/monthly schedules |
The data show that digital giving not only changes the payment method but reshapes giving behavior. Mature churches on the Pushpay platform have already achieved around 40% digital penetration, 25 percentage points above the industry average of 15%, directly validating the path of "deep use → higher conversion." Andrew notes that Pushpay's primary competitor is "inertia" (i.e., traditional giving habits), not other digital platforms — meaning that once a church completes its initial migration, retention barriers become extremely high.
Part II presents a medium-to-long-term hypothesis: if Pushpay reaches $15 billion in annual transaction volume and achieves reasonable adoption of its unified software suite, its revenue could be several times the current $130 million, and annual profit could exceed $100 million. Measured against its purchase market cap of roughly $1 billion, this implies a potential P/E of about 10x (if profit reaches $100 million), and profit growth could come from leverage.
1. Transaction volume growth: From ~$5 billion in 2019 to $15 billion in the medium-to-long term, an increase of 200%, primarily driven by digital penetration rising from 15% to 40%+ and Pushpay's market share gain (currently about one-third of digital giving).
2. Unit economics improvement: As transaction volume expands, the platform's take rate on payment processing fees (typically 2-3%) can be maintained or even increased because churches adopting SaaS modules (e.g., CRM, event management) significantly raise customer lifetime value (LTV). Pushpay's SaaS subscription fees (~$200-500/month per church) and transaction cuts form dual revenue streams.
3. Operating leverage: Under a platform model, the marginal cost of adding transaction volume is very low (only payment network fees + server costs), enabling gross margins of 60-70%. The current ~$130 million revenue corresponds to a net margin of about 12% (~$15 million), while $100 million profit would imply a net margin of over 25% (assuming revenue of $400-500 million), demonstrating strong leverage.
| Metric | Pushpay (Hypothetical Medium-term) | Shopify (2020) | Square (2020) |
|---|---|---|---|
| Transaction volume (annual) | $15B | $120B | $112B |
| Revenue / Transaction volume | ~3% | ~2.5% | ~4% |
| Net profit margin | 20-25% | 15-20% | 10-15% |
| Market cap / Transaction volume | ~0.7x (using $10B market cap / $15B volume) | ~1.2x | ~0.8x |
If Pushpay achieves $15 billion in transaction volume, even at a conservative 0.7x volume, its market cap could reach $10.5 billion, representing a 10x upside from the purchase price of $1 billion. Andrew is cautious and does not give a specific multiple, but implies that a "reasonable multiple" would make the business worth far more than the entry price.
Andrew defines Pushpay's primary competitor as "inertia" rather than other digital giving platforms (e.g., Tithe.ly, ChurchTrac). This insight carries investment value:
In his "Concluding Thoughts," Andrew refers to investing as a "commission" and ends with "we will eventually reach the other side," echoing the earlier forward profit assumptions for Pushpay. He emphasizes "stay hungry," implying that the long-term trend of digital giving penetration is still in its early stages. From a data perspective, even without considering cross-selling of Pushpay's unified software, simply moving transaction volume from $5 billion to $15 billion (corresponding to industry digital penetration of 40%) could generate roughly 3x revenue growth, and profit growth could exceed 5x (due to leverage). This provides a clear compounding path for long-term holders — consistent with the philosophy of holding high-quality enterprises for the long term, as repeatedly emphasized in Andrew's letter.
(Note: The above analysis is based on specific numbers and assumptions disclosed in Part II, and does not repeat the previously analyzed content on Pushpay's business model, industry background, or investment philosophy.)