Carvana, Snowflake, Credit Acceptance, Kinsale, Mainfreight, CSI

Content Published Last Week

1. Carvana: Post-COVID Challenges

2. Snowflake & Databricks: Comparison, AI & Convergence of Offering

3. Credit Acceptance: CAPS Workflow Challenges

4. Kinsale Capital: Competitive Advantage

5. Mainfreight Culture and US Market Challenges

6. Constellation Software & VMS Competition

Carvana

Yesterday, CVNA agreed with noteholders to restructure its debt and raise at least $350m in equity. It exchanged its existing ~$5.7bn unsecured debt at 7-8% for over $4bn of ~13% PIK notes secured by ADESA assets. The new PIK stack reduces total debt outstanding and reduces cash interest payments in the next 2 years only to increase thereafter to a similar cash interest expense as 2022.

Although this deal with bondholders seems to kick cash interest down the road, CVNA has made progress reducing SG&A costs per unit. The company reported over $1bn in cost has been removed in the last 12 months. We published an interview with a Former leader of CVNA operational team exploring how CVNA's philosophy and operational strategy has evolved since COVID to reduce the conditioning unit costs:

Another significant aspect, which can be overlooked, is how the vehicle is staged and pulled. From an associate's perspective, who is physically doing all the work, they have to sift through 6000 units to pull one, stage it, and get it ready to be on a truck by noon. They don't start pulling those cars until it's funded. This is a change from six to nine months ago, and everything prior to that, when they would start staging a car once the customer had put in their information and showed commitment. - Former Director at Carvana

It’s also early days; CVNA isn’t fully benefiting from its ADESA assets just yet:

The timing of the ADESA purchase was unfortunate, especially considering the price paid…ADESA will be beneficial in the future. It will be a key factor in Carvana's growth without adding significant operating costs, assuming they can prove their profitability and regain the trust of everyone. Leveraging ADESA is how they will grow without adding a ton of operating cost because they already have it. It’s already incurred. - Former Director at Carvana

This interview is worth its read in entirety for anyone curious on how CVNA’s infrastructure and operational strategy has evolved since COVID and its potential moat vs KMX.

Snowflake vs Databricks AI/ML Offering

In a previous interview, a Databricks executive discussed how its strategy is converging with Snowflake’s as the company scales. This interview with a large customer of both SNOW and Databricks compares the two offerings and their efforts in AI. Databricks has led the way with its recent acquisition of Mosaic AI:

What is rapidly becoming essential is the ability to either develop your own LLMs or utilize Gen AI products. Both approaches have seen interesting, yet distinct, developments. Take Databricks, for instance. They recently acquired Mosaic AI for over a billion dollars. This move signifies their message to large corporations that they can build their own LLMs using their proprietary data, within their own infrastructure. There's no need to send your data to OpenAI. Databricks can manage your data and, depending on how they integrate with AI, you can use it to train and build your own analytics. For example, if you're JPMorgan Chase and you want to train your preferred model on your data, you don't need to send that data out to Azure or Anthropic. You can keep all that data in-house and use Databricks to help you build and train models. - CTO, Customer of SNOW

On the other hand, SNOW’s ML strategy isn’t as clear:

To put it bluntly, Snowflake's ML story is a hodgepodge. I wouldn't even know where to begin using Snowflake to build medium-sized or complex ML models. Perhaps I could do simple things by running Docker containers with Python scripts in Snowpark, but there isn't a robust end-to-end pipeline of products spanning Spark, which is massive, and MLflow. It's important to remember that Databricks also has something that Snowflake lacks in the ML use case - developer mindshare. If you and I were to start a machine learning company or hire engineers who know how to build models, they would likely be familiar with TensorFlow and Spark. They probably wouldn't know much about Snowflake. Therefore, Snowflake has a significant gap when it comes to developer mindshare. - CTO, Customer of SNOW

Credit Acceptance & CAPS

Since 2018, CACC’s unit loan volume has stagnated:

close

Zoom InZoom OutReset

This interview with a Finance Manager at an auto dealer and customer of CACC explores how dealers use the CAPS system and why it may be limiting CACC’s unit volume growth:

For example, when pushing a loan application through a third party platform like Dealertrack into CAPS, Credit Acceptance only proposes a single vehicle for the deal. Another main reason we don't use them is the qualification process it uses. When we input the customer's information into Dealertrack, it pushes to CACC. We open up the CACC portal, which is CAPS, and we conduct our deal through there. If there was a way where it could provide us with multiple vehicles to choose from instead of the one it thinks is the best, that would be better. - Finance Manager at indie dealer and CACC Customer

This lack of choice incentivizes dealers to use other financing partners who offer loans across multiple cars:

Westlake Financial gives us options to pick any car on our lot that we can enter into their portal. It will show us exactly what the deal will look like. It will even tell us exactly what we need to make the deal profitable. I know CACC has the same feature in their CAPS Portal, but again, it does not give us the option to enter more than one car and to work deals that way. Finance Manager at indie dealer and CACC Customer

There are also other limiting factors including how the dealer holdback is calculated, vehicle book values, and other points of friction throughout CACC’s loan origination process discussed throughout the interview.

Kinsale Capital Group

This interview with a Former SVP at Kinsale explores the company’s competitive advantage:

From the perspective of small businesses, which make up 50% to 60% of the premium gross rate, the 24-hour rule and the ability to get multiple quotes in a single day are crucial. A broker with a small piece of business doesn't want to send it to numerous insurance companies. If they've had success with getting a quote from Kinsale within 24 hours at a good price, they will keep coming back. This is one of the main reasons why these wholesalers don't want to spend much time on a small account that's only generating $10,000 in premium. They prefer to make one phone call, one submission, get a quote and send it out. They might get two or three quotes for comparison purposes. But when dealing with Kinsale, they know they will get a competitive price and the speed and function save the wholesaler time and money. - Former SVP at Kinsale Capital

Mainfreight US

The heart of Mainfreight's culture is the empowerment of decision making of those closest to the customer. Everyone goes through the same training program and decisions are taken by relying on the company's set of values/principles.

The secret to their success, as I see it, is that they want the person closest to the customer making the decision. For instance, at Owens, when we were trying to win a customer, we would usually involve a few people from the head office who didn't know much about our business or the client's business. They were there more as a support mechanism than anything else. At Mainfreight, however, you don't need to involve those people in daily sales - Former President at Mainfreight

However, MFT's culture internationally doesn't seem to align with its roots. This interview with a Former MFT President in the US explores how the culture and operations abroad compares to the success in its home market.

Constellation Software & VMS Competition

In this IP Podcast, we discuss our 6-month research project that led to our report on CSI's competitive risks. We discuss why it seems like everyone today is rolling up VMS and the potential risks to CSI, including:

1. How many CSI copycats we found across US / EU

2. How their capital allocation strategy differs to CSI

3. How increased competition may increased CSI ROIIC

4. How many and why M&A executives are leaving CSI

Listen to the episode on Spotify here. We also have many published many other related articles on CSI's moat and potential impact on future IRR's:

Enterprise Research: Constellation Software: A Competitive Analysis

Constellation Software, Allscripts, & CSI 2.0 - Analysis of the Allscripts acquisition and what it means for CSI’s FCF deployment.

Halma, Danaher, CSU & Serial Acquirer Org Structures - Analysis of different organizational structures for serial acquirers and what it means for scalability of M&A.

Constellation Software Competitive Risks with Former CSI M&A Executive

Constellation Software: Competition, Hurdle Rates & Deal Flow