Content Published Last Week

Amazon Retail: 1P vs 3P Economics

This interview with a Former Amazon VP who ran multiple different retail product categories is insightful to parse out 1P vs 3P economics.

The quote below best encapsulates AMZN's philosophy in operating 1P and 3P retail: AMZN would rather 3P sellers fight over low margin categories, but not if the sellers can't compete with the price of Walmart or other retailers.

Anything where Amazon is not necessarily going to make a large amount of profit, they'll willingly let 3P have the business. The times where they won't willingly do it, isn't just tied to profitability; it's more tied to customer perception. If this item is offered at Walmart, via 1P and Walmart discounts this item all the time, or if this item is offered at Macy's and Macy's discounts this item all the time, and we have it on our 3P, but we can't control price, then we are going to be perceived by our customer as not being the best option. - Former VP at Amazon

AMZN purposely loses money on 1P products to drive customer loyalty and LTV.

Diapers is an example of where Amazon chooses to lose cash. In some instances, for every diaper it sells, it may lose cash and it's okay with that. Some of the reasons might be competitive pressure, and they want to make sure that they're at the most competitive price. But some of the issue maybe that they know that they generate a long-term value on a customer that may not have been on Amazon before but because they now have a baby, and they have a subscription service with diapers, that same customer may buy everything from Amazon. They are thinking about the lifetime value of that customer. - Former VP at Amazon

But this doesn’t mean 1P isn’t profitable in some categories. Take Levi’s for example:

let’s take the Levi 501 as an example. I can guarantee you Amazon buys every single color and size of the Levi 501. I know their relationship with Levi's and the way Amazon would work on the 501. Internally, Amazon will view it as – I forget what it sells for on the site, but let's just say it's $40 – we likely pay Levi's $26. It's not quite the same normal apparel margins; Levi's has a little bit more leverage, let's just say, especially on its iconic jeans, like the 501 or the 550 or 505. Amazon may make only 35 to 40 points on that particular item. That being said, it's a highly replenishable, highly predictable item. That means the amount of excess that Amazon has in that item is not high. Once you start having thousands upon thousands of units sold, my statistical sample size of my demand forecast, especially if I launched it back in late 2008, means we have years and years of knowing how it sells, even in a recession. How it sells in boom times; how it sells seasonally. - Former VP at Amazon

Some categories like beauty or apparel are 30-50% gross margins where AMZN can run 1P profitably. Over time, even for lower margin 1P categories, AMZN can better forecast demand and optimise inventory to maximise 1P ROIC.

But AMZN isn’t optimising for profitability per category: it’s designed to offer the customer the best price, regardless of who sells the product. It's a game of scale: AMZN need as much volume as possible shipped through its fixed fulfilment cost base. The launch of Buy with Prime also aligns with this philosophy.

It's difficult to parse out the true unit economics of 1P and 3P. And it's not clear this is even the correct way to look at the retail business. Does AMZN sum the net loss of 1P and count this as CAC for the overall retail business? Would the LTV uplift outweigh the loss in 1P margin? It's hard to tell. It's all one big ecosystem. Similar to how it's difficult to grasp the net impact of AMZN losing a Redshift customer to Snowflake: AMZN loses data warehouse revenue but may earn greater EC2 and storage revenue via the customer operating SNOW on AWS.

However, we can parse out and estimate the overall retail gross margin. Our simple estimates below show how the retail product gross margin has declined from ~23% to 20.5% over the last 7 years. After netting out Whole Foods' gross profit and adding Prime subscription and ad revenue, the online retail gross profit margin has increased from ~29% to 37.4%. This is likely due to shift in mix to higher-margin 3P revenue.

After deducting fulfilment expenses, the online retail contribution profit margin has increased ~200bps since 2015.

Source: AMZN Filings, IP Estimates
Source: AMZN Filings, IP Estimates

These are very rough estimates given the reported numbers so please reach out if we're missing anything. We plan to publish more research on AMZN Retail margins and ROIC in the coming weeks.

Burford Capital: IRR's & Potential Regulation

Burford Capital, the leading litigation finance company, reports a curious data point that almost defies the gravity of capital markets itself:

As BUR's investment size has increased...

Source: BUR 2021 CMD Pres
Source: BUR 2021 CMD Pres

The case return has increased...

Source: BUR 2021 CMD Presentation
Source: BUR 2021 CMD Presentation

We've long questioned how this could be true. A Former Investment Manager at Augusta, a competitor to BUR, suggests there is a simple answer:

The higher return is also due to the value of cases. A £20 million case no longer involves straightforward contractual litigation. An investment duty claim or construction case has a higher risk towards the end and you could lose £20 million and not reap any rewards. Previously, you were only losing £2 million to £4 million but you had another 10 cases in the pipe which could be lucrative if they win. With £20 million cases, you invest in less claims and diversify your investments in a smaller pod of more expensive cases, but they are high-risk high-reward. You could potentially win £80 million at the end of it. As a result, the market prices in a higher return on investment, and your pricing gets higher as you put more money into these cases. - Former Investment Manager at Augusta Ventures

At the outset, litigation is a binary risk: you either win damages or lose your full investment. Although nearly two-thirds of cases settle, the binary nature of the asset demands a higher required return.

The executive suggests that the higher the absolute dollar amount invested in a binary asset, the higher the required return. The high-value cases are also more complex and therefore less competitive. BUR's experience and expertise potentially gives the company pricing power at the top end of the litigation market.

Financing litigation cases is a relatively new industry. And new financial services are typically at risk of being regulated.

Some clients argue funders take too much of the damages and the return should be capped. Others believe law firms increase their fees for funded cases. Law firms know funded cases have a willing backer and thus jack their fees higher. These higher fees can even increase the absolute $ return to the funder given the return calculation includes the amount invested in the case.

Success fees also need to be regulated in complex cases as funding becomes more commonplace. A funder shouldn't be able to charge huge amounts more on one case than another, if they are quite similar in nature, which does happen. It is often dependent on what they think the client will accept. That could definitely be regulated going forward, because it's not common for a marketplace to have such huge variables. - Former Investment Manager at Augusta Ventures

As the market grows, regulation could limit the potential case returns. But then surely this wouldn't compensate funders for the binary risk they are taking? Regulation is bound to happen, how it's structured is the question.

Datadog's Differentiation & Open Telemetry

This interview with a Former AppDynamics VP, who is now running Devops at one of the largest financial services company in the US, explores why DDOG is unique.

It’s tough because I think they’re all good; they’ve all got tools in their tool belt to make them the winner of everything. But Datadog move quicker and I think that’s the differentiator...the company embraces new technology and moves in that space and starts to get a product rolling out even if it’s an early like data product. That sets them apart. - Former VP at AppDynamics

DDOG is pulling ahead of competitors in AI/ML offerings too:

For the longest time I was very much biased towards AppD because I thought we had the best product, we moved fairly quickly. But companies get big and they start to slow down a bit and that’s normal. But Datadog has a got a Watchdog AI engine and it’s got fully automated root cause analysis which they all kind of do, but you have to look at how easy is this to use, how accurate will it give me my readouts; what kind of algorithms are they baking into it. I think Datadog is doing pretty well in that regard. I think AppD still does pretty well but I would say, if I was going to go top three, New Relic, AppD, Datadog, with Datadog kind of above the other two. - Former VP at AppDynamics

One risk that threatens the growth of all proprietary devops vendors is Open Telemetry: an open-source collection of tools and API's to collect and export telemetry data.

If you look really deeply at it, every monitoring and observability application has monitors. They all talk to their own systems and they can paint the picture of what you’re looking for. Open telemetry is like open-source monitors, open-source agents or agent lists. It can talk to AppDynamics, it can talk to Datadog, it can talk to Grafana, Splunk whatever the case maybe and its open source. You can have your entire fleet running these monitors, agents and you can use any third-party vendor you want – or open-source vendor – to graph your data, to show you real time analytics, to show you real metrics, so on and so forth. Basically, that opens the door to changing vendors much easier. - Former VP at AppDynamics

Salesforce: MuleSoft Strategy

This interview with former SVP of Mulesoft on CRM's acquisition strategy and Mulesoft's growing pains is interesting. Some believe CRM didn't truly understand what they were buying; it was a solution to the integration problem customers consistently explained to CRM:

I'm not even sure that Salesforce really at the heart of it, got what MuleSoft did. That's not a knock on Salesforce. I think it's just more that Salesforce had a big problem at the time and they were very public about this. In every customer survey that they did, and every customer roadshow, integration was consistently floated to the top of concerns for their customers. But integration is a very broad word, so with their customers, I think a lot of what they were talking about was, hey, Salesforce, you've acquired a bunch of stuff, marketing Cloud, commerce, whatever and a lot of that stuff is not integrated with itself. Not to mention integration with other stuff. The number one problem was, how did we get our stuff to integrate internally on the platform? - Former SVP, Mulesoft

A lot has changed since CRM acquired Mulesoft nearly 5 years ago. Notably, the building blocks of cloud computing architecture are much more integrated. SaaS vendors like DDOG are building their own integration solutions across the stack to drive stickiness of their own core solution.

But Mulesoft argues this has always been the problem: companies creating more point-to-point integrations in the stack makes it more complex and fragile. Increased integration tools by the hyperscalers and SaaS vendors could prove to be a headwind for Mulesoft.

I think at a tactical level – and this has been, frankly, the problem that MuleSoft has faced from the beginning – they have to convince a customer to take a longer-term view. They have to convince the customer, think about this stuff more strategically. Think about integrations, not as a tactical thing to glue point A to point B. But think about it more as a strategy for driving innovation, for driving new applications at a higher velocity. You're right that it's a competitive force that will always be there, either through custom code or through these embedded things. The easier those things make it to do that, the easier it is for customers to get themselves into trouble. I think the good news is a lot of customers at scale, figure that out over time and they come back, which we've seen in many examples in the past, but it is a headwind for the company, for sure. - Former SVP, Mulesoft

Ulta Beauty

As department stores have declined, competition for the high-margin beauty category has increased from all angles:

The primary justification is that beauty competition has gotten much tougher, specifically, over the last few years. It's a high-margin category, and as the department stores demised, you started to see a whole host of retailers more aggressively go after beauty. Target is a steep competitor there; most people go into Target way more often than they go to Ulta beauty. They may go to Ulta once a month for big purchases, but they're probably in Target once a week, a few times a week. So they're getting that unplanned traffic of I may not be going to Target for Ulta, but I'm there. While I'm there, I'm going to look, and while I'm looking, I'm probably going to purchase because I’ve got a whole host of things in my cart already, so what's one or two more? - Former VP at Ulta Beauty

Spirax Sarco's Decentralisation

Spirax is an amalgamation of 8 manufacturing sites and 8 manufacturing sites globally:

One key point about the supply chain is it's decentralized. If you look first at the manufacturing sites, there were eight manufacturing sites, four in the Americas, three in EMEA and one in Asia-Pac, back in 2014. They all reported locally, to a regional divisional director or a country GM where the manufacturing site was. It was decentralized and some great things come with a decentralized organisation but there are also certain weaknesses. Yes, they were largely autonomous, having their own P&L, but we had duplication of products made in those eight manufacturing sites. Some of this duplication was positive, it provided resilience and risk avoidance. Some of it was great, because it gave you that proximity to a customer, accommodated different and local standards, and very importantly gave you a lead-time and availability advantage. - Former Global Supply Chain Director at SPX.LN