Published Last Week

Union Pacific, BNSF, & Class I US Railroads Competition

This interview with a Former VP of Union Pacific, who spent over three decades of experience running Class I railroads across the US, explores the history and evolution of competition between major rail players:

When I started with Union Pacific in 1988, it was the start of the merger era. We went from 15 or 20 Class I railroads down to mainly two in the East, two in the West, two in Canada, and two in Mexico. To get merger applications through, you had to maintain competition in markets if you were absorbing another railroad. To address those concerns, the merging railroad would grant access to competitors into those marketplaces or to specific customers. This could take the form of trackage rights, where they are allowed to run their trains over your railroad, or haulage, where we haul their commodities for a set price. BNSF operates on an average of probably 6,000 miles of our territory, with access rights either overhead or serving rights where they stop and serve at the customer level. It's very complicated with all kinds of legal restrictions. - Former VP at Union Pacific

Union Pacific has the largest percentage of captive markets:

I once mapped out head-to-head competition using a system called the Serving Carrier Reciprocal Switch System. This industry-maintained system shows every rail-served customer in North America, who serves them, and what type of access they have. Union Pacific has the largest percentage of captive markets due to the purchase of Southern Pacific. We evolved differently, focusing more on industrial and production railroads. In contrast, Burlington Northern merged with merged with Santa Fe, who was very much into intermodal, making them the premier intermodal provider. - Former VP at Union Pacific

The interview goes on to explore how BNSF wins share from Union Pacific and the competitive dynamics of winning new customers.

ASML & TSMC: Customer Relationship & Growth Outlook

This interview with a Former Director at ASML, who spent over 17 years at the company, explores how the company works with customers like TSMC:

About three years before the ramp-up of a new node, ASML and, for instance, TSMC will discuss the requirements for this specific node, which is set to ramp up in three years. TSMC defines the total cost of technology for the given node, driven by the manufacturing process, corresponding capex equipment investments, and inputs from their customers, like Nvidia. They dissect this cost of technology into different parts and allocate certain budgets, focusing on wafer cost. The total wafer cost is split into the lithography cost per wafer, and even for each wavelength, there is a calculated bucket. They start discussions, saying, "For these layers, we need more volume of the product. Please give us a good offering. For this other one, we need a better overlay. What is your offering here? For that one, we need a better cost. - Former Director at ASML

The interview also explores the potential impact of multi-beam or nano-imprint technological improvements over the next decade:

Even if the cost is two or three times cheaper, it must also be less than two or three times slower to be viable. Multi-electron beam lithography might find a place in small to medium-sized factories that can afford a $50 million tool but not a $200 million one. These factories might not need to produce 40,000 wafers per month, but only 5,000. There is potential here, as EUV is not accessible to small and mid-sized factories. There is interest from sectors like defense and aerospace for these factories to use alternative technologies in leading-edge spaces, maintaining local sovereignty. - Former Director at ASML

Amazon FBA Seller: Impact of Tariffs

This interview explores how a large FBA seller is navigating tariff volatility when sourcing product from China and pricing on Amazon marketplace:

Everyone is increasing prices. We monitor competitors using Helium 10 and other software. For example, if we sell a filter for $12.97 and rank number one, but someone else sells it for $10 and their ranking improves, we might offer a deal or lower our price to maintain our spot. However, everyone is raising prices. - FBA Selller

However, in the face of higher costs, FBA sellers are scared to cut advertising spend:

Many sellers are scared to cut ads because they want to maintain their market placement and not lose market share. It also depends on the type of ads. The reason EP can cut their ads is that most of their customers are already subscribed to our coffee, tea, and filters. We have about 60,000 to 70,000 subscribers, so they recur every month. Lowering our ads means we're not acquiring as much of the new market, but we're keeping our existing market because subscribers don't leave unless the product quality declines or changes significantly. - FBA Selller

In our previous research, we explored how Amazon's endemic advertising business works. Endemic ads are ads purchased by 1P or 3P merchants advertising products sold on Amazon. In Q4 2023, we estimated ~80% of Amazon ad revenue was from endemic properties.

Amazon Endemic Ad Business, IP Research

There is little doubt that ad revenue is a critical driver for Amazon's retail profitability. In 2023, as the chart below shows, Amazon's contribution margin (e-commerce margin after paying all variable costs when shipping a product) jumps from ~4% to ~16% when including advertising.

Amazon Retail Contribution Margins, IP Research

Not only is advertising a 'must-have' for marketplace sellers, but we also explored how ad revenue may decouple from GMV growth due to growth in non-endemic advertising (ad spend from companies that don't sell products on Amazon.com leveraging AMZN 1P data to display ads on Amazon O&O and 3P-websites).

Amazon Advertising Revenue / GMV, IP Research

This interview with a large FBA seller can be read alongside much of our prior work on AMZNs retail and advertising business:

Formula 1: Hospitality, F1 Experiences & Race Promotion

This interview with a Former Formula 1 Executive explores how the hospitality business works on race day:

Many years ago, Bernie Ecclestone granted a license to Paddy McNally and All Sport for a fee. This license allowed them to manage two main areas, the sponsorship, so all of the trackside advertising and the partnerships in Formula 1, and the hospitality. Paddy McNally had those two things under license for many years and he ran them very successfully. On the other hand, Bernie sat in London, and all he needed to worry about was the big ticket stuff like TV rights and promoter revenue. That was where the division was split. - Former Formula 1 Executive

Teams are the biggest buyers of ticket for the Paddock Club, which they use to entertain sponsors:

We put the whole Paddock Club together. It has its own kitchen, so DO & CO build their own kitchens on site. We don't use on-site kitchen facilities. We kit it out with carpets, tables, and chairs, and have movable dividers to create suites for the teams. The rule was that a team had to buy from us; they got nothing for free. A team would come to us and say, "I've got five sponsors who want to bring guests, so I'm going to need 100 tickets." If they needed 100 tickets, I would say that qualifies them for me to put up the dividers to create a private suite. They could then decorate it internally themselves with team decoration. They bought 100 tickets to qualify for the divided space. If they had less than 100, they would sit in the general lounge area. Our biggest clients were the teams, buying on behalf of their sponsors. If I was a team, I would have done a sponsorship deal with a major sponsor. In that deal, the sponsor would give me money, and I would give them branding on the car and a certain number of hospitality tickets per year.- Former Formula 1 Executive

Prices for the Paddock Club have risen significantly in the last 7-8 years:

Back in 2017, a three-day ticket for Friday, Saturday, and Sunday with full hospitality in the Paddock Club was approximately $4,800. If you wanted to purchase that today, the price would generally be double, although it varies depending on the location and region. In the last few years, prices have essentially doubled, if not more. It's an expensive proposition, and the question is how far you can push that. - Former Formula 1 Executive

Farfetch vs Cettire: Luxury Fashion Marketplaces

In late 2023, Farfetch was acquired before bankruptcy by South Korean online retailer Coupang. In Q2 23, when Farfetch last reported earnings, the company had ~$4 billion in annual GMV and ~$800 million in annualized operating losses.

While there are a number of reasons why Farfetch might have failed, we explore two fundamental differences between Farfetch and Australian luxury marketplace Cettire:

  1. Sourcing Supply and Brand. Relationships
  2. Product Pricing

In its early years, Farfetch relied exclusively on inventory from small European multi-brand boutiques. To expand its inventory beyond boutiques, Farfetch signed brand partnerships to access inventory directly. By 2023, Farfetch sourced ~15% of supply from brands, 20-25% from 1P sources and 60-65% from boutiques. One area of exploration in our research covers how brand partnerships negatively affected Farfetch’s business model for 3 main reasons:

  1. Price competitiveness due to geopricing
  2. Non-differentiated brand inventory
  3. Loss of unique boutique inventory

We explore the challenges with brand partnerships and how it impacted Farfetch's competitiveness relative to Cettire's business model today. This research can also be read in parallel with our work on the Luxury Gray Market which covers:

Luxury Gray Market Distribution, IP Research

Kneat: Global Customer Perspective

An executive who has been a Kneat customer at three different companies quantifies the time savings from paperless validation:

The time savings are significant, though it's difficult to quantify exactly. I can give you an example. In projects, delays were often attributed to C&Q, but in Stein and Visp, we realized it wasn't C&Q causing delays; it was engineering. We were faster at qualifying and commissioning equipment than they were at implementing new equipment. The big advantage is the speed in this direction. Looking at the data I have, I calculated the time gains for a normal life cycle. By going paperless and doing it right the first time, we achieved a savings and agility gain of 0.57. You're at least twice as fast using a paperless system. (...) Even with conservative estimates, the savings were substantial. For the Lynx project, we had a target of $50 million but ended up investing only $6 million for C&Q. - Kneat Customer

Yet, siloed Commissioning & Qualification programs make it hard to convince multiple sites to change. This explains why greenfields are much easier pickings:

"The big problem Lonza had was that every site had its own C&Q program. Because I was part of the global project bringing in new facilities, they had to follow my rules. The C&Q program I implemented was initially used for greenfields. My goal was to eventually roll it out for brownfields as well, for lifecycle management, so they would also have to change.- Kneat Customer