Over the last 3 years, US retail media as a category has grown from $13bn to over $41bn. It’s the fastest growing digital advertising category in the US and Amazon has ~80% market share. Retail media includes any digital advertising offering on an online retailer’s property that leverages its data.
Retail media has grown rapidly because it enables advertisers to target audiences and understand purchase behaviors down to the SKU level. Retailers have more potent customer data than the brand itself. Media networks like Amazon offer closed-loop attribution which enables advertisers to directly link ad dollars spent to actual dollar sales. With the deprecation of cookies, this is crucial.
We believe there is a misconception that, like other marketplaces, Amazon’s advertising revenue / GMV is limited to ~5%. This interview explores why and how Amazon has the opportunity to defy this common base rate.
We estimate ~85-90% of Amazon’s revenue is from endemic advertisers. The company is just scratching the surface of leveraging its 1P purchasing data elsewhere on the internet:
Amazon is telling you, almost down to the user grain, this is the makeup of interests and affinity groups for your brand. That gives you a ton of signals on how you message your brand, but there is also classic reach. They are header bidding against DV360 on the world's top com score 2,500 sites for open web. They have the richest data when cookies go away and are matching hashed users against other logged in portals and properties. If I am logged in on Reuters and Amazon, there is hash matching in an API which is a valuable signal to target against. There is scale and value there which they are seeing great performance from, both endemic and non-endemic. That puts Google slightly on the ropes because of the signals they don't have. They know what you are searching for but they don't know what you bought. - EVP at Skai (Kenshoo)
There are many insights throughout this interview with a Former Director of Technology at Capital One who ran the Snowflake implementation and has experience with AWS Redshift and Azure’s Synapse data warehouse offering.
SNOW gained traction due to the fundamentally different warehouse architecture relative to incumbent offerings:
Synapse is built on legacy architecture, and the decoupling and clustering of data warehouses is not meant for the cloud. They were built using on-premise SQL Server and enhances from there. Snowflake was built directly on the cloud for the cloud, which makes a huge difference when it comes to scaling. Synapse is struggling in the industry to make that move simply because of the legacy architecture it was built on, which doesn't help them act quickly enough in that space. It has multiple problems including compute resources, scalability, administration and cost. - Former Director of Technology at Capital One
But SNOW’s biggest problem seems to be its cost:
Snowflake will continue to lead because they are niche in that area. If I had to do a split in five years using Fortune 500 companies as a sample, Snowflake will have less market share than today. Snowflake's biggest problem is cost, which is killing many companies who are on Snowflake. Even though they have a larger market share now, it will eventually shrink if they don't focus on cost, because people are tired. Snowflake is like a V12 engine in that it provides amazing performance but uses too much fuel…BigQuery performance and cost is 30% to 40% cheaper than Snowflake for the same workload. - Former Director of Technology at Capital One
There has been much hype about Bing and ChatGPT. How much of it is truly warranted given the demo's mistakes is unclear. But there is no doubt that Nadella wants to make GOOG ‘dance’.
Arguably, GOOG’s monopoly is facing the biggest challenge in its recent history from multiple angles:
1. Bing and ChatGPT trying to offer something new in search
2. DoJ suing GOOG over DoubleClick for Publishers monopoly
3. Deprecation of cookies
4. GCP unlikely to ever catch Azure and AWS
In short, as always, Buffett wins. And Apple. GOOG pays Apple $20bn to be the default search on iOS. Pretty sure the TAC is only going up from here regardless of how effective Bing is!
To your words about Google, a user can be lukewarm, but I don't think Bing has created anything innovative over the last few years. The reality is that Bing, at best or in most cases, can provide a similar, if not identical, experience to Google. Google is already so integrated into the system. A key point that we need to mention is the platforms that Google has to drive you to search. The friction for me, as a user, to take my iPhone and instead of tapping directly and going into Google Search – Google pays $20 billion to Apple for that, by the way – I go to Bing or download an app. All that stuff is too much friction for a very low marginal value. - Former Product Lead at Google Search
Alcon is the global leader in eye surgery equipment, a global oligopoly.
"Yes, I think, for margins and Covid, it's hard to equate that into margins. Sales dollars, yes. Everybody had a recovery period. I think more of their margins are due to competitive price pressures. For example, in the interocular lens business right now, if I could guess, Alcon has about a 47% share of that business, followed by Johnson & Johnson with probably a 35% share, but now here comes Zeiss. Zeiss is now entering the market, challenging them on everything from surgical microscopes to phaco machines to interocular lenses, and they are having to defend their share by price concessions." - Former Director at Alcon AG
It seems that after the spin-off, a potential serious new entrant in eye surgery and consolidation trends in the Vision Care segment may have impacted management's guidance of increased margins.
Let’s say Charlotte is a $2 million market, you multiply that by an approval rating of 50%; then another 50% of the engineers like and will accept us; then how many of those do we put a bid in on; and how many of those do we actually win? All of a sudden, that number goes down significantly. When you look at those four components, you can say in Charlotte, we’re okay with approvals, so we don’t have to work to DOT, but maybe our acceptance rate among the design engineering firms is horrific, so how do we fix that? We fix that with hiring engineers to go into those engineering firms and tell them how much they can save using us versus reinforced concrete pipe, because there are a lot of savings from putting plastic in versus reinforced concrete pipes, you can put a lot more feed in per day. Then your win-rate is just about the sales guys getting out there and doing their jobs. You could look at each metro area and see in this area, we need a lot of help with the design engineering firms on acceptance. Honestly, I would let sales guys go and hire civil engineers instead.
The Flow acquisition was not for the customers; it was a strategic M&A. They were aware that Shopify and Flow were working on the white label solution, and this was a way to position Global-e as the full top to bottom partner for international e-commerce with Flow. It was a tough time but it was the right long-term decision. What can Flow achieve as a standalone solution within Global-e? There will be knowledge sharing between the two solutions. If something works well in the shipping or fraud processes, I think they will copy it. Very smart people are working on the Flow solution. What can Flow's partnership with Shopify achieve? Everything we talked about. There will be a lot of opportunity for smaller brands and those who don't want to sign a contract with Global-e, which is similar to their contract with Shopify. We spoke about the Borderfree M&A. I don't know if you saw in my bio, but I actually started in the company that eventually became Borderfree.
Where the DSP component of Tremor has the ability to understand the supply that's coming from the supply-side platform sister company, it can be more successful from a delivery perspective. It also allows both sides to broker direct deals with the two ends of the existing ecosystem, which is the DSP side being able to broker greater deals with the publisher community and the SSP side being able to broker improved deals with the actual agency community. So having both middleware technologies and those connections with the demand and the supply allows for greater deal generation. Of course, the third part is part of the business practices, and there is a margin component where the company collects a margin either from the supply side or the demand side from the transaction, which improves its profitability.
This document may not be reproduced, distributed, or transmitted in any form or by any means including resale of any part, unauthorised distribution to a third party or other electronic methods, without the prior written permission of IP 1 Ltd.
IP 1 Ltd, trading as In Practise (herein referred to as "IP") is a company registered in England and Wales and is not a registered investment advisor or broker-dealer, and is not licensed nor qualified to provide investment advice.
In Practise reserves all copyright, intellectual and other property rights in the Content. The information published in this transcript (“Content”) is for information purposes only and should not be used as the sole basis for making any investment decision. Information provided by IP is to be used as an educational tool and nothing in this Content shall be construed as an offer, recommendation or solicitation regarding any financial product, service or management of investments or securities.
© 2023 IP 1 Ltd. All rights reserved.
Subscribe to access hundreds of interviews and primary research