Highlights: First Quarter of Unpacked
Bold product moves, AI things accelerate, Tech policy in the front seat, Interesting things continue. Bird’s eye view of everything published on Unpacked in the past quarter.
I published the first ever Unpacked piece “Meta, News Outlets and the Canada Online News Act” towards the end of July. I wasn’t really sure what to expect but I was excited to write it. Seventeen articles in and it’s been an incredible journey so far. This post is a bird’s eye view of everything that was published on Unpacked in the past quarter.
Bold product moves - Threads, OverflowAI, Uber Ads & Instacart IPO
Tumult at Twitter continued post the Elon Musk acquisition despite some promising Twitter (X?) moves around the launch of a creator revenue sharing program. Meta fully capitalized on the opportunity and launched Threads. While Threads had an impressive start, there were clear gaps in their approach to building the product as a text-based everything for everyone. I argued that in the history of social networks, there is very little precedent for a social network that started with a generic / broad focus and ended up being successful, and Threads needs to pivot to focus on a specific user motivation (social or interest / curiosity-based). I hope to do a refresh on this in the coming quarter.
Another fascinating story was that of StackOverflow, which has been disproportionately hit by the generative AI wave. The company has so far executed on a strong pivot through a multi-pronged strategy. They launched Overflow AI which powers conversations-based search using their own data, and also plugs seamlessly into developer environments. The product also extends its capability by plugging into an enterprise’s data sources, in turn helping StackOverflow double down on its enterprise offering and minimizing reliance on eyeballs-based advertising. They have also been one of the first companies to embrace a data licensing program for large language model providers.
The other notable product bet we saw pan out incredibly well for companies was advertising. Uber announced their first ever profitable quarter and Instacart went public, both thanks to their advertising revenue. Very similar businesses in many ways, they leveraged advertising as a way to turn their low margin operations into profitable businesses. An important pattern with both their ads products was the ability to have formats that are “direct response” / “performance” based and can drive measurable results for advertisers (UberEats sponsored ads, Instacart product ads for consumer packaged goods companies).
AI things accelerate - Startups, Questions about moats & Regulation
Unsurprisingly, there was more momentum around AI. In one of the early Unpacked articles, we talked about the AI value chain - infrastructure layer (model providers like OpenAI), platform layer (data platforms like Snowflake) and applications layer (startups, features within existing products).
I argued that most of the value from AI is going to be captured at the applications layer, while the infrastructure and platform layers get commoditized quickly. Zooming in on the the applications layer, the article further dug into startups in YCombinator’s W23 batch and uncovered some interesting trends: focus on specific problems, integrations with existing software software, LLM customizations, new UI interfaces, data silo-ing for enterprises to name a few. While all useful applications, the article pointed out how several of these startups lack long-term moats.
This became more apparent in a follow-up piece where we zoomed in on how SaaS companies are launching their own AI features, clearly challenging the sustainability of several moatless AI startups.
One of the encouraging things to come has been early conversations about regulation. AI regulation in the US is still in its early days and there are two types of regulatory constructs under consideration today - 1) broad bills in the Senate which cover a wide range of issues and might be difficult to get consensus on, and 2) non-binding, broad frameworks listing out AI principles but without much specifics agreed upon. The article on AI regulation made the case for a more focused approach to AI regulation that is less of a “bundle everything into one bill” approach, and more of a targeted approach that regulates specific mechanisms tied to meaningful AI risks.
Tech policy in the front seat - Anti-trust trials & EU v. advertising
This has arguably been one of the biggest quarters for tech policy. Two huge anti-trust cases moved forward and as part of Tech Policy September, we went deep into both DOJ v. Google and FTC v. Amazon, including an analysis of the full complaints filed by regulators.
I’m incredibly impressed by the DOJ’s case against Google - their most convincing argument focused on the deeply anti-competitive deals that Google essentially forced Android device manufacturers to sign by withholding critical Android functionality. While I have been critical of the FTC in the past (OpenAI's web crawler and FTC missteps), I was impressed by their arguments against Amazon - in particular, they made a nuanced but convincing argument about how Amazon’s practices prohibited sellers from providing cheaper prices on other marketplaces (even if they wanted to), which suppressed price discovery and therefore harmed consumer welfare.
The other tech policy topic in limelight was EU’s war on behavioral advertising. That article dug into the latest enforcement decision against Meta, and how it essentially shuts down the only possible pathway in which Meta could get users to opt into behavioral advertising by default. This goes one big step beyond what Apple’s privacy changes do - Apple’s policy only applies to cross-app advertising while EU’s policy applies even to ads within a company’s own app. I argued that the EU probably went a step too far by essentially killing the only real monetization option that free digital products had, and this might have several second degree effects including companies reducing their product investments in the EU.
Interesting things continue - BNPL market & end of Intuit’s lobbying
Lastly, I covered a few companies and markets that I was personally fascinated by. First was Klarna - it’s a great story of how first mover advantage helped them cement their position in Buy Now Pay Later (BNPL) market. The article also brought out some interesting data that quashes myths about this market - it’s not as predatory as it is accused of being, and the pandemic altered BNPL spending behavior towards more essential category (vs people splurging on expensive discretionary things). We also looked at some regulatory challenges they might run into, and more importantly, competition from companies that own massive distribution (read: Apple).
Which brings us to the latest piece. The story of Intuit’s lobbying showed regulatory capture at its worst. E-filing in the US first started in 1986 - thirty-eight years later, the IRS announced that they will pilot a direct tax filing system in 2024. Intuit systematically (and in all credit, impressively) lobbied for twenty years to keep its biggest competitor (uh, the IRS) from launching their own free direct tax filing software and it’s finally coming to an end.
Concluding thoughts
Well, not much! I’m psyched about how far Unpacked has gotten. I’m encouraged by all your engagement and it keeps me motivated to publish weekly. Thank YOU for being here. If you liked Unpacked so far, tell a friend or colleague about it.