WhatsApp's inflection, LLM network effects, and another Meta win
12 November 2023 | Issue #15 - Mentions $META, OpenAI, $AMZN
Welcome to the fifteenth edition of Tech takes from the cheap seats. This will be my public journal, where I aim to write weekly on tech and consumer news and trends that I thought were interesting.
Let’s dig in.
WhatsApp monetisation
Despite its immense popularity outside the United States, WhatsApp has flown under the radar for most American consumers due to the dominance of iMessage. With over 2.7 billion monthly active users globally, WhatsApp is the most widely used messaging app in the world. However, until recently, WhatsApp's service was deeply under-monetized and considered an expensive acquisition by investors. Meta (then known as Facebook) acquired WhatsApp in 2014 for $19 billion when it had about 450 million users and was generating less than $30 million in revenue. Compare this to Meta's acquisition of Instagram, for which they paid $1 billion. Instagram is now estimated to bring in over $40 billion in revenue for Meta, versus less than $1 billion for WhatsApp just a year ago.
Monetisation has been a challenge for WhatsApp for almost a decade, which can be traced back to its original founders’ philosophy of eschewing ads. To make matters worse, the company decided to ditch its $1 annual subscription fee for consumers in 2016. Instead, it opted to focus on monetising its service through businesses. This new strategy involves charging companies a monthly subscription fee for added features like the ability to build a WhatsApp website, as well additional fees for messaging campaigns sent to WhatsApp users and on a per-conversation basis.
WhatsApp's revenue has been reported under "Other revenue" in Meta's financials, as part of the broader "Family of Apps" segment. This segment "consists of net fees we receive from developers using our Payments infrastructure and revenue from WhatsApp Business Platform and various other sources." Despite touching over 2.7 billion monthly active consumers globally and 200 million businesses, WhatsApp's revenue feels underwhelming - generating only $800 million in 2022.
The core issue is that WhatsApp's parent company, Meta, is an advertising juggernaut. Meta has been unable to realize synergies between WhatsApp and its ad business because an encrypted messaging platform is not conducive to advertising. WhatsApp's founders were adamant about user privacy and encryption from the beginning, making targeted ads impossible. This philosophy conflicts with Meta's ad-driven business model and makes it challenging to directly monetize WhatsApp's massive user base.
While WhatsApp has tried to boost revenue through business tools and features, this B2B approach has not yet proven nearly as lucrative as Meta's ad targeting capabilities. With over 90% of Meta's total revenue coming from advertising, WhatsApp's contributions remain relatively miniscule in comparison. Unlocking the full monetization potential of a platform used by over a quarter of the world's population remains an ongoing struggle.
Enter click-to-message ads. In 2017, Meta introduced click-to-message ads to help businesses drive engagement on WhatsApp and other messaging platforms like Messenger and Instagram. These ads allow consumers to move seamlessly from viewing an ad to having a direct conversation with the business via messaging.
At Meta's Q4 2022 earnings results, the company shared that revenue from click-to-message ads across WhatsApp, Messenger and Instagram was already at a $10 billion annual run rate. By Q3 2023, click-to-WhatsApp ads alone were generating revenue at a multibillion-dollar annual run rate, and “continues to grow very quickly”.
Are ads within WhatsApp off the table though? There was speculation from the Financial Times that Meta was considering showing ads in WhatsApp’s contact conversation lists on the chat screen. However, Will Cathcart, President of WhatsApp, quickly denied this on social media, calling the story “false”. In a recent interview with Brazil’s Folha De S.Paulo, Cathcart gave further context on WhatsApp’s stance.
(Translated from Portugese)
PC: A recent report in the Financial Times said that Meta was discussing placing ads on WhatsApp. WC: The article claimed that we were going to place ads in the inbox. We are not doing this, nor discussing it. We don’t think this is the right model. When people open their inbox, they don’t want to see advertising PC: So will WhatsApp continue to be free without ads? WC: Yes, in your inbox and in your messaging experience. The reason I qualified the answer is that there could be ads in other places - channels or statuses. For example, channels might charge people to subscribe, they might be exclusive to paid members, or the owners might want to promote the channel. But, no, we won’t put ads in your inbox.
Channels is a new feature that was introduced on WhatsApp that enables one-to-many communication, as opposed to the app's traditional one-to-one or group messaging. With Channels, businesses and organizations can broadcast updates to an unlimited number of subscribers. These subscribers can sign up to receive updates from public Channels run by verified accounts. Unlike private group chats, Channels allow for unidirectional communication from the Channel creator to subscribers. Users cannot reply or post their own content back to the Channel.
Its grown very popular with news publishers, and paves the way for further monetisation of the app.
At the end of September, the Financial Times launched its first WhatsApp Channel, for global markets news. The Meta-owned WhatsApp describes Channels, launched in beta over the summer and rolled out globally in September, as “a one-way broadcast tool for admins to send text, photos, videos, stickers, and polls.” The FT’s global markets channel, which sends out one free story a day, grew quickly, and on October 9, the FT launched a second channel, dedicated to news of the Israel-Hamas conflict.
“It took a year for us to grow an audience of 35,000 on Telegram, first around the pandemic and then the war in Ukraine,” said Rachel Banning-Lover, the FT’s head of social media and development. “Comparatively, we [grew] a similar-sized following around markets on WhatsApp in two weeks.”
The FT is one of a growing number of news publishers testing WhatsApp Channels. “We know that the way people share news stories is changing. They’re not sharing stories on Facebook or Twitter as much as they used to; they’re texting them to their friends and their group chats,” said Nisha Chittal, Vox.com‘s managing editor. “We wanted to find a way to tap into that audience behavior.”
Another feature to be monetized will be AI chatbots. Meta introduced their AI agents in September, which I wrote about previously. Cathcart shared in his interview that these agents will allow users to search the internet and provide recent information (monetized via search ads), as well as act as customer service agents for businesses to help solve initial (or complete) customer queries.
It’s been a long decade for WhatsApp, but the business seems to be reaching an inflection point with these new monetization features. WhatsApp's $19 billion acquisition price may turn out to be cheap after all, if it can successfully unlock revenue from its enormous user base while maintaining its core values around privacy.
Related: Google and prominent telecom groups call on Brussels to act over Apple’s iMessage
LLM network effects
This week, OpenAI held its inaugural developer conference, which included numerous announcements about new and improved products. Rather than rehashing the same content you’ve probably already read online, I’ll link to some pieces that I thought were worth reading.
Everything announced at OpenAI’s first developer event - TechCrunch
How OpenAI is building a path toward AI agents - Platformer
The OpenAI Keynote - Stratechery
Last week, I wrote about the costs of operating various models, as well as the risk of applications built on top of foundational models getting vertically integrated by OpenAI. The announcements from OpenAI's developer conference confirm part of this thesis and highlight the benefits of being an early mover in the battle for large language model dominance. Custom GPTs look to capture the demand that AI wrappers like Character.AI have created. Meanwhile, OpenAI wants to become the control point for user workflows by reducing the need to leave ChatGPT for additional functionality. As the most popular AI chatbot, ChatGPT needs to retain users within its own interface.
It’s also managed to reduce the cost of GPT-4 with an updated model, GPT-4 Turbo.
We released the first version of GPT-4 in March and made GPT-4 generally available to all developers in July. Today we’re launching a preview of the next generation of this model, GPT-4 Turbo.
GPT-4 Turbo is more capable and has knowledge of world events up to April 2023. It has a 128k context window so it can fit the equivalent of more than 300 pages of text in a single prompt. We also optimized its performance so we are able to offer GPT-4 Turbo at a 3x cheaper price for input tokens and a 2x cheaper price for output tokens compared to GPT-4.
GPT-4 Turbo is available for all paying developers to try by passing
gpt-4-1106-previewin the API and we plan to release the stable production-ready model in the coming weeks.
I found this question posed on Reddit super interesting.
The costs for running AI models seem to be rapidly decreasing, creating a flywheel effect that benefits OpenAI. Scaled infrastructure is reducing OpenAI's costs, allowing them to pass on savings to customers through lower prices. This makes OpenAI more competitive in gaining market share, which gets more users onto their platform, generating more data to improve their models. With seemingly unlimited capital available for AI companies, offering the lowest price to grab share appears strategically sound for OpenAI, as long as this enables them to become the dominant player long-term where the economics would be favorable. There are still question marks on the latter, though.
Related: Google in talks to invest in AI startup Character.AI, Google Announces Expansion of AI Partnership with Anthropic
Another win for Meta (and Amazon)
The last couple years have been transformative for Meta. Apple’s App Tracking Transparency (ATT) feature, released in iOS 14.5 in 2021, significantly handicapped Meta's ability to track users and target/measure ads effectively. ATT requires apps on iOS devices to ask for user permission before tracking their data across apps and websites. If users opt out, then apps can only use data gathered from their own apps to serve ads. This led to decreased ad targeting efficiency and ROI for Meta due to the loss of cross-app signals and data, estimated to cost them $10 billion in revenue.
Fast forward to today, Meta has invested heavily in AI and machine learning to improve its ad systems within a more privacy-focused landscape. Much of the targeting and measurement efficiency has now returned to pre-ATT levels.
This week, Amazon and Meta announced a partnership that allows shoppers to purchase Amazon products directly from ads on Instagram and Facebook without leaving the apps.
Two of tech’s largest players, Amazon.com Inc. and Facebook parent Meta Platforms Inc., are testing a feature that lets shoppers buy Amazon products directly from ads on Instagram and Facebook.
The initiative, which involves asking consumers to link their Amazon accounts to their social-media profiles, could make Meta more attractive to advertisers and let Amazon attract more shoppers from outside its web store.
The unprecedented partnership between the two companies also could help them fend off challenges from TikTok, which has launched a US e-commerce marketplace, as well as such Chinese upstarts and Temu and Shein.
US shoppers will see real-time pricing, delivery estimates and product details on select Amazon ads running on Facebook and Instagram, according to an Amazon spokesperson.
“For the first time, customers will be able to shop Amazon’s Facebook and Instagram ads and check out with Amazon without leaving the social-media app,” the spokesperson said.
The partnership between Amazon and Meta should lead to better ad conversion for Amazon, given the reduced friction in the purchase process. Allowing shoppers to buy Amazon products directly within Instagram and Facebook removes extra steps previously required to complete purchases. This streamlined experience is likely to increase conversion rates on Amazon ads run through Meta's platforms.
Additionally, the integration provides more first-party data for Meta. By enabling direct on-platform purchases, Meta can now link specific ad exposures directly to resulting sales. This gives Meta valuable conversion data to optimize ad targeting and measure ROI, complementing the targeting capabilities of its AI systems.
Overall, the partnership is mutually beneficial - Amazon gets a smoother purchase experience leading to higher conversions, while Meta gains purchase data to strengthen the effectiveness of its ads. Removing friction for users while providing better measurement for advertisers is a win-win.
That’s all for this week. If you’ve made it this far, thanks for reading. If you’ve enjoyed this newsletter, consider subscribing or sharing with a friend
I welcome any thoughts or feedback, feel free to shoot me an email at portseacapital@gmail.com. None of this is investment advice, do your own due diligence.
Tickers: META 0.00%↑ , OpenAI, AMZN 0.00%↑