Three charts to inform and delight

Self-driving cars, online shopping, and the cloud wars

It’s Friday, so we’re going to keep it light. A tasteful chart potpourri for the weekend.

• Waymo’s mythical smile curve

• Temu is a marvel when you think about it

• What jobs postings say about the cloud wars

---

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Sticky self-driving cars

Slowly, but surely, Google’s self-driving taxi company, Waymo, is doing something pretty incredible:

Earnest

Waymo retains ~33% of its customers 3 years after their initial purchase, which is substantially higher than Uber or Lyft’s.

But that’s not the incredible part.

Not only does Waymo retain its customers, its retention rates improve over time. Waymo is demonstrating what

Alex Izydorczyk

called the mythical “smile” curve, wherein customers say “hey, actually, we stopped using it for a bit, but we missed this thing, and now want more of it.”

Alex happened to be talking about Only Fans and shows the ‘smile curve’ in action:

For better or for worse, porn customers come back for more.

But Waymo customers appear to demonstrate a similar kind of loyalty. Retention troughs after a year at about ~25%, and then builds back up to 33%.

Now, there could be other forces are work here, but the one Random Walk will choose to believe is that fully self-driving taxis are an amazing experience, and having done it once—perhaps on a trip to San Francisco—customers are eager to do it again the next time they are back in town.

More, please.

Self-driving cars are emerging as a “post-hype sleeper” for transformational technology.

My friend

Brad Hargreaves

has written some thoughtful pieces here and here, but generally speaking, “commute time” (and to a lesser-extent parking set-asides) are substantial limiting factors when it comes to where and how people can live. That’s why, for example, homeowners pay a substantial premium to live near a commuter rail—because they can commute to work without having to drive. It’s also why elected officials should be drawn and quartered for stifling the development of new rail lines.

But soon it may not matter.

A fully autonomous driving option functionally puts a commuter rail stop in everyone’s driveway (minus the traffic part, but I suspect that self-driving cars will manage that better, as well). That’s a huge unlock for sub- and ex-urban development, and will fundamentally change the way developers and homeowners think about location.

Temu rising

Random Walk has written variously about the trials and travails of ecommerce generally, but also Temu specifically.

It’s interesting for a number of reasons:

• the persistence of the ecommerce S-curve (and the implications for the world’s bellwether companies)

• the role ecommerce, specifically cheap ecommerce, plays in consumer spending trends (and the second-order beneficiaries of those trends); and

• the curious, but consistent, ability of China to build viral consumer apps (but few enterprise ones)

There are probably other reasons that it’s interesting, but those are enough for now.

Anyways, some data suggests that Temu is now the world’s second most-visited e-retailer (after Amazon) in just two years:

Similarweb/MarketplacePulse

Temu (apparently) gets ~700M monthly site visits.

That’s pretty remarkable growth, and of course, site visits doesn’t include mobile, where most of the shopping likely happens anyway. Temu is the third-ranked shopping app, by usage.

So, Temu has blitzscaled its way into the hearts and smartphones of many, many people, in a very short period of time. It did it with an unsustainable unit economics, and certainly benefited from post-pandemic consumer dynamics that favored both cheap and online, but it did it nonetheless, and honestly, had you asked me at the outset, I would have said “impossible.”

Now that Temu has done the impossible, it remains to be seen whether it can stick without the benefit of spending more than it makes, but I find it impressive, nonetheless.

AWS losing ground?

Three quick observations on the Cloud Wars from jobs data.

AWS AMZN 0.00%↑ is losing its dominant share in the hiring market:

Revelio

The share of job postings mentioning AWS has declined from 65% to about 40%.

The biggest beneficiary of AWS’s decline has been Microsoft’s MSFT 0.00%↑ Azure, which is now in ~35% of cloud postings.

The decline is driven by a decrease in AWS-only postings

AWS-only used to be slightly more than half of all postings, but is now just a shade under 30%.

The flipside is that Azure-only is now almost a quarter of all postings. The implication is that the overall decline in share reflects a real shift in customer preferences away from AWS, and not just greater specificity among recruiters.

AWS remains the preferred choice of first-adopters by far, i.e. the hiring shift away from AWS is driven by firms with cloud infrastructure already in-place:

AWS has 60% of new- and first-time cloud adopters, while only a very slim lead for firms making their second cloud purchase.

It’s hard to draw too many conclusions from any of this, but if you want to speculate about the long term success (or not) of AWS relative to the field, you certainly can.

Personally, what I found interesting is:

• (a) Azure does not see any meaningful bump following the release of ChatGPT. Whatever Azure’s gains, they seemed to have plateaued around 2021; and

• (b) AWS is either less sticky than I would have thought and/or firms are increasingly amenable to spreading out across multiple cloud offerings. Anecdotally, I’ve heard variations of the latter, but perhaps this is some validation in the data.¹

Have a nice weekend.

---

Previously, on Random Walk

Getting old is what we choose to do

Insurance and Private Credit, two peas in a pod

Banks reporting for duty (part deux)

Big banks reporting for duty

July 12, 2024 Weekly Recap & Good Reads

Insurance prices don't always make total sense, but it's not climate change

---

Random Walk is an idea company dedicated to the discovery of idea alpha. Find differentiated data, perspectives and people, and keep your information mix lively. A foolish consistency is the hobgoblin of small minds. Fight the Great Idea Stagnation. Join Random Walk. Follow me on twitter. Follow me on substack:

1. Another vague speculation, but the idea of using multiple cloud providers is somewhat consistent with the industries driving the shift:

Aerospace+Defense, and Information technology have changed their hiring preferences the most, and certainly in the former case, one could see very good reasons for diversifying (rather than churning) one’s cloud vendor.