One of Australia’s fastest growing start-ups shuttered last week.
Unable to raise fresh funding from venture capital (VC) firms gone miserly in 2023, ultra-fast grocery delivery service Milkrun decided to wind down while it still had cash to pay staff and suppliers.
Milkrun’s rise was one of the fastest and attracted a lot of investor attention.
In early 2022, the company closed one of the largest early-stage VC capital raises in Australia’s history, raising $75 million.
Big players were involved. Tiger Global Management, AirTree Ventures, Skip Capital, and Grok Ventures.
But that $75 million wasn’t enough.
Milkrun’s fast growth came with high costs. The firm was losing $13 for every fulfilled order. Last Friday, it closed.
Founder Dany Milham told staff Milkrun suffered from deteriorating market conditions:
‘Since we announced our structural changes in February, economic and capital market conditions have continued to deteriorate, and while the business has continued to perform well, we feel strongly that this is the right decision in the current environment.’
But what does Milkrun’s closure tell us about investing and placing big bets?
AirTree’s Milkrun bet explained
Taking on the oligopoly of giant supermarkets like Coles and Woolworths is a steep challenge.
So why did Milkrun appeal to big VC firms? Weren’t the business risks clear?
Milkrun is to supermarkets what ghost or dark kitchens are to restaurants.
The company operated out of warehouses in select, dense inner suburbs of Melbourne and Sydney, stocking about 2,000 grocery products, which were delivered by a fleet of riders on e-bikes.
The warehouses permitted 10–15-minute deliveries for customers in the relevant suburbs.
Milkrun didn’t need to lease a huge store and recruit retail staff or bother with stocking fringe items.
It had appeal, going from zero to one million orders in 15 months.
In January last year, AirTree’s partner Jackie Vullinghs wrote a blog post explaining the company’s investment in Milkrun.
Vullinghs wrote that the grocery shopping experience is ripe for disruption:
‘Clearly, there’s an opportunity for a new company to improve the grocery shopping experience for a convenience-led world, delivering groceries in minutes at supermarket prices.
‘Grocery is a $122bn market in Australia, dominated by slow-moving incumbents with low NPS.
‘The population is urbanised, wealthy, and willing to pay for on-demand food delivery (Melbourne & Sydney are in the top 5 cities globally for Deliveroo).’
While admitting the grocery segment is ‘traditionally a low margin category’, AirTree thought an upstart like Milkrun could shake things up with ‘improved unit economics in multiple areas’:
- ‘Personalisation: With an app that uses data effectively, you can personalise the shopping experience for each consumer, increasing average basket size and improving frequency and retention.
- ‘Range localisation: A smaller retail footprint in lower-cost areas, combined with a solid data infrastructure allowing a granular understanding of which products sell in which suburbs, enables lower leases, wastage and inventory costs.
- ‘Vertical Integration: Having a direct to consumer relationship leads to constant customer feedback. With that feedback, you can create new products directly with producers and expand product margin.
- ‘Range Expansion: Over time, you build up a dense network of micro fulfilment centres throughout cities, allowing you to expand into higher-margin products and potentially partner with other brands that want to deliver a delightful consumer experience.’
Vullinghs said the extent of what Milkrun achieved up to that point in early 2022 gave AirTree a ‘glimpse of the extraordinary things they’ll make possible in the future’.
That glimpsed future proved to be illusory. But the curdled Milkrun investment has not dented AirTree’s confidence.
In correspondence obtained by the Australian Financial Review, AirTree told investors failure is an ‘inevitable part’ of venture capital:
‘Failure is an inevitable part of VC. With great upside potential comes risk, and if we don’t see some failures, we’re likely not adding enough risk into the portfolio for outliers to emerge and become fund returners. With that said, this wasn’t the outcome we hoped for.’
And herein lies the crux of the matter.
In investing, risk is everything. How to foresee it, how to manage it, how to come to grips with it, and how much to take on.
The risk you assume must be commensurate with the implied reward.
I don’t think firms like Tiger Global and AirTree underestimated the risk of Milkrun’s failure. The risk was clear to see in the numerous closures of rival grocery delivery start-ups last year.
In March 2022, Quicko collapsed. In April, Send did. In July, Delivr folded. And in November, Voly and Deliveroo both shut down.
VC firms surely understood the risk. They just thought the potential reward was worth it.
Milkrun, risk, and investment returns
How much risk are you willing to take on?
Depends on the upside.
You will accept more risk on a stock you think can rise 10-fold than on a stock you think will trade sideways.
If you were presented with a bet where you had a 50% chance of losing $50 and 50% chance of winning $100, you’d likely take the bet.
In the aggregate, you come out ahead as the expected value is positive: (0.5 x $100) + (0.5 x -$50) = $25.
The expected value may be positive but won’t soothe you if you wager once.
That’s why investors, and especially VC firms, spread their bets.
Investment manager Emanuel Datt made this point when commenting on Milkrun’s closure:
‘Venture capitalists are paid by their investors to take risks on companies that are deemed to be too immature for publicly listed markets. This could be by backing entrepreneurial individuals or investing at more aggressive valuations than listed market comparables etc.
‘It’s important to judge an investor’s performance in aggregate overtime rather than focusing on individual losses within a portfolio. All investors, without exception, experience losses on individual positions.’
What you could criticise VCs for is not necessarily the risk they accepted, but their assessment of the potential reward to warrant it.
Did investors inflate Milkrun’s upside?
In hindsight, one of the biggest hits to Milkrun’s long-term upside came from the very firms it sought to disrupt — the big supermarkets.
Seeing the positive response from Milkrun’s customers, Coles and Woolworths rolled out their own delivery services.
In a farewell dispatch on LinkedIn, Milkrun’s Dany Milham said scale was vital for success:
‘We always knew that achieving the scale needed to reach overall profitability would require significant initial and ongoing investment over a long period of time.’
Who has more scale than Woollies and Coles?
Faced with sterner competition from deep-pocketed rivals, Milkrun’s cash burn turned into an inferno.
Already losing money on each order, Milkrun now had to pay more to acquire new customers.
In the end, Milkrun’s upside was capped by its business model’s calculus.
As Datt elaborated in his Milkrun commentary:
‘You can’t eat revenue. Revenues can be remarkably easy to build quickly, but the true measure of success for a company is sustainable profits and market share over time. Clearly the traction to date had not been sufficient to justify further investment.’
ChatGPT takes on Wall Street
Before I go, a final note.
Earlier this month, I wrote about the rise of artificial intelligence technology popularised by OpenAI’s ChatGPT.
In that piece, I wondered whether the large language model tools underpinning ChatGPT will be taken up by investment companies hunting alpha.
And it does seem like generative AI research is heading for Wall Street.
Overnight, Bloomberg reported this:
‘Two new papers have been published this month that deployed the artificial intelligence chatbot in market-relevant tasks — one in deciphering whether Federal Reserve statements were hawkish or dovish, and one in determining whether headlines were good or bad for a stock.
‘ChatGPT aced both tests, suggesting a potentially major step forward in the use of technology to turn reams of text from news articles to tweets and speeches into trading signals.
‘“It’s one of the rare cases where the hype is real,” said Slavi Marinov, head of machine learning at Man AHL, which has been using the technology known as natural language processing to read texts like earnings transcripts and Reddit posts for years.’
The hype is real, it seems.
Regards,
 |
Kiryll Prakapenka,
Editor, Money Morning
