Pinduoduo and the continued growth sell-off

Dear investors,

Portfolio update

The fund returned +7.6% in September. Strong performances from Carvana, Pinduoduo and Pointsbet outweighed continued falls in US software and more popular ‘growth’ stocks, with many of them 30-40% below recent highs.

Liquidations like this, where high momentum stocks sharply reverse, generally involve extraordinarily large moves from the highs to the lows (see Q418), but the opportunity to make decisive investments is often fleeting.

This happens every year or two, sometimes in different sectors, sometimes across the whole market. Taking the other side of these moves is not for the fainthearted, as you have to buy when prices are moving 5% a day, but it’s usually rewarded. With markets at all time highs, we can safely say that every dip should have been bought.

The good news is that rallies at the back-end of these sell-offs can be smooth and consistent, with low volatility, as shorts are forced to buy back shares and hot money slowly buys back in at higher and higher prices.

I’ve summarized the moves across some of our companies below, which represent a significant majority of the portfolio. I’ve added CY18 EV/Sales which marks the strongly negative sentiment at the end of last year, which was quite extreme. As you can see, 2020 multiples are well below even those extremes.

The growth rates below give some indication of the disruption these companies are causing while taking meaningful market share from incumbents. I find this more helpful than TAM analysis. If a disruptive model is that good, customers should flock to it.

This combination of extraordinary growth, innovative business models, depressed valuations, and significant short interest, is one of the most favorable stock market setups for this kind of strategy I’ve seen in a long time.


For some reason I’ve never been able to interest people in Pinduoduo. Afterpay, Pointsbet, Carvana, sure. But few seem excited about this strange Chinese company that’s somehow taken a 10% market share bite out of Alibaba in two years.

I’m guessing this is because Pinduoduo has developed a genuinely unique business model, but it’s only obvious if you look closely. Here’s some of the reasons we think it’s been such a hit:

All buying is viral and social. Every item on the site has a list price, which you can buy at any time, and a lower price, which you can access by buying in a group, sometimes with only with one other person. Customer acquisition is viral and exponential. It’s no accident that Pinduoduo accumulated 500 million customers in four years. When you first go on the site, you’re offered steeply discounted goods – but only if you share the link and a friend commits to buying.

The model is consumer-to-manufacturer, and they mean it. Pinduoduo only takes a 0.25% transaction fee – astonishing strategic discipline. This means all the savings are shared between the customer and the manufacturer, which also means they have a sustainable competitive advantage over other platforms that do take a cut.

PDD optimizes for user time spent on their platform, rather than sales. Long before PDD overtook, customers were spending far more time on Pinduoduo’s apps. This offers several advantages:

1. Advertising is cheap, because they have so much screen real estate to sell, and

2. They have better data, and more opportunities to A/B test different prices, deals, rewards and lotteries, in each case carefully assessing ROI.

Users are encouraged to browse, not search. Pinduoduo’s search bar is almost hidden. When you go on the app you see one-off deals, games, rewards, lotteries and new offers. This simple change leads to a materially different user experience to Alibaba (or Google and Amazon, for that matter), while offering new opportunities for the merchants that use the platform.

The platform benefits small and mid-tier manufacturers. Search-driven e-commerce really only advantages the largest and most efficient manufacturers, who can afford to pay for ads, while sucking oxygen from anyone who can’t make the front page of search results. Ten years ago things may have been different, but in 2019 the market for adwords is highly efficient. In the West Amazon is not only search-driven, but also comes with a division that screens for fast-selling items and creates internal competitors. With Pinduoduo, the factory can retool for a single product and sell the lot in advance on a Pinduoduo flash sales.

Pinduoduo earns money through a high margin advertising. It costs little to offer ads on a website, and the sales process can all be automated. This amounts to about 3% of total transaction value.

The platform aggregates demand. Ben Thompson at Stratechery developed a robust framework for evaluating different tech ‘platforms’, arguing that those that aggregate demand have all the power and accrue all the value. This is a helpful lens, demanding an answer to the question: who is directing customer flow?

It’s one of the reasons we think Afterpay might go all the way, as unlike Paypal or Visa, Afterpay is building a direct relationship with customers and driving incremental traffic.

‘Aggregation theory’ highlights strategic errors too: Uber aggregates passenger demand, so could position themselves as a capital-lite open layer on top of any competing self-driving platforms. Instead, they’re burning capital in a highly crowded field against multiple well resourced and technologically sophisticated competitors.

Facebook and Google have been chillingly ruthless in using the demand aggregated on their sites to constrict independent journalism. The sweetest advertising model of all is when customers come to you to be directed to their next purchase. Pinduoduo has this in spades.

The fact that user experience is unique also means that Pinduoduo might carve out a sustainable corner of the market.

This was not an easy investment. We first bought around $24 in March, then dramatically increased our position from May to July when the stock fell below $20, the chart looked awful, and all kinds of bearish reports were coming out from leading investment banks and bomb-throwing short sellers. We wore a 20% hit, and the stock advanced over 120% in the months following, which is broadly why we’ve been OK through this growth sell-off. After a strong run, no doubt there will be further twists and turns.

I’ve added an updated version of my favourite PDD chart below. If you take 31% of the growth of a fast-growing industry, your market share approaches 31%. Needless to say, with a high margin advertising business, if PDD maintains this momentum their market value will be an order of magnitude higher than the ~$20b EV we were buying at.

Podcast update

I had a chat with Paul Brennan, who spent over 20 years at Citi, most recently as Chief Economist. Paul began his career at the Reserve Bank of Australia and the Treasury, and now consults a range of clients, and teaches university students in his spare time. We went deep on market history, Australian economics and the financial landscape in the 90s, when the conversation of the day revolved around inflation and high yields.

Peter Stevens also came on, and we talked about the growth sell-off, the opportunities we were seeing, and what we were buying.

Mario Emmanuel also joined us to talk about how to value companies that have no balance sheet assets and no profits, but nevertheless constitute some of the greatest investment opportunities of today.

Available on iTunes here or desktop here.


Growth stocks are down 30-40% in the US, and we’ve taken the opportunity to add to the most competitive, disruptive, fast-growing and undervalued companies. Pinduoduo and Carvana have performed quite well in this environment, perhaps because they were crowded shorts, rather than crowded longs.

Some of these stock market falls have been well deserved. The sharp moves in companies like Zillow, and Grubhub shouldn’t surprise anyone who ran the numbers or gave consideration to their competitive positioning.

Still, the WeWork debacle has put something of a cloud over technology. There’s a chance we are already past ‘peak WeWork’, but there’s also a chance we see further ~20% falls in the companies tabulated above. This would bring software multiples down to the lows of early 2016. Over the three years that followed, these same firms advanced 4x-6x or more.

We would actually welcome this scenario, as it would be the optimal path for long term


Best wishes


You may also like