Request For Startups

alternatively, things I want handled.

Here’s a (growing) list of miscellaneous problems I want handled (not necessarily by me). I will keep adding to it.

If you’re working on one of the below problems, feel free to reach out. I may also add a thread for solutions. Keep in mind that many of the problems likely have working approaches (e.g. grid resilience is an active research field), but I’m highlighting the problems here because I think they are underrated for their worth. [1]

Expectations: 2021 will be more precedented than 2020. Reality: see Context.

Context: Topics below were drafted before the events of January 6, the debates on social media de-platforming, and the exodus from WhatsApp to Signal and Telegram. I stand by my below comments, especially my views on ads as a monetization model (Telegram, please don’t). The post was longer than intended, especially Point 2 where I tried to avoid going down the adtech. rabbit hole, but failed. If you’re not interested in me lamenting the demise of products due to ads models, you should go ahead and skip number 2, and read only the bolded. I lightly edited the below for clarity in light of recent events. Corrections always welcome.

---

The Problem List

1. Digital resilience and preservation. The problem here is the fragility of digital content systems.

   • Resilience. If electricity goes out, how long does it take for us to restore communication and power? How easy is it to cause outage? Can we turn our phone off for more than a day, two days, a week, a month? See Nashville for a recent example for why I’m worried about this.

   • Content preservation. Much of historical accounts is now preserved digitally. We take crawler services like Wayback Machine for granted - what will happen to content in 50, 100, 150 years? Who will decide on what to and when to maintain? Are we ok with the implicit privatisation of libraries (implicit in a few people end up deciding what digital content to archive, and people default to these accounts due to a lack of alternatives)? On a personal level, if something bad happens to your cloud storage provider, do you have content backups? See Azure and Gmail 2020 outage for a recent example of not having access to a scarily large amount of your own content. Deleted Twitter replies is a minor but frustrating example of this. [2]

2. Decreasing reliance on AdTech for digital monetization. The problem here is that ads lower content and experience quality, and that there are not many great monetization alternatives for content-driven companies. See summary at end of bullet point. One possibility is transitioning to subscription-based models (e.g traditional media approach on digital monetization or Substack itself). Would you pay Facebook and Twitter a monthly fee if they promised no ads, and no data sharing?

   Ad models create problems in digital content distribution. When I talk about ads here, I mean exclusively platform ad displays, and not sponsored creator posts. Lots of people have made smart observations on this topic. I won’t try to substantiate that ads enhanced social media polarisation (beware: USC paper linked in the Wired article is a draft) or monopolised product marketplaces. I also won’t try to substantiate that ads monetization necessitate privacy trade-offs (DuckDuckGo is one such counterexample). That this article needs to exist or that Apple adding user privacy features cost companies ad revenue hopefully say enough.

   For one reason or another, advertisement is the default primary monetization model for content-based tech. companies. These companies include Google (70.7% total revenue from ads), Tencent (16%, mostly WeChat), Facebook (98.5%), Twitter (86%), Snapchat (62.6%), ByteDance (60%), and Reddit ($119 million, unk. total) for a total combined ad revenue of $ 214 billion USD est. 2019.

   

   Global Digital net ads revenue, ranked by size.

   To my untrained marketing eyes, ads bundled with content and search isn’t the most ideal solution for users or advertisers (see Search case study). Proctor & Gamble cut their digital ads. spending by $200 million in 2018, with little to no impact on its revenue. Given that intent to buy is difficult to proxy well, some doubt is cast on the effectiveness of digital ads for advertisers.

   Ads fund much of our digital content. If ads become unbundled from media and search, how we consume content would be vastly different. Mark Zuckerberg himself published a 2019 note citing a desire to find a non-ad based monetization model for messages that allows for more privacy (ironic given the things we know about WhatsApp in 2021). I’m not sure if the answer here is trying other monetization models (e.g. subscriptions), or a separate, ads-only platform for users with intent to buy. Whoever cracks this can change digital content and platforms as we know it.

   • Case Study: Better Search. Google (88.14%) and Bing (6.18%) take up roughly 95% of the worldwide search engine market in 2020 [source: statista]. At a glance, it seems unwise to challenge search incumbents, given the sheer size of compute and engineering ingenuity that has gone into the making. However, there are unchallenged assumptions in the current landscape that could pave way for a much different future.

   • Ads are annoying for users. AdBlock extensions are popular for a reason. People shown a given Google ad click on average only 2% of the time. The difference to me between a relevant search result and platform-generated ad spot is that the former has every intent to buy, while the latter does not necessarily measure intent. It’s possible to show only product search results, and sell bidding spots for those exclusively (e.g. Amazon but for every store), without embedding it into the rest of platform content. When there is no intent to buy, ads are annoying for users.

   • Ads aren’t effective for ad purchasers without user intent to buy, which is often not the case. Even the most committed, financially viable user may not want to click away to a vacuum ad link on YouTube (the same is not necessarily true for sponsored YouTube creator content). See previous P&G example. If as a seller, you need to spend lots on ad budgets, 200 SEO factors to filter intent to buy, and multiple site pop-ups to force users to abandon their ad-blockers, maybe you’re forcing an impossible sell. It shouldn’t have to be this complicated.

   • So why is search (and media) bundled with ads? Brief History Tangent. There are lots of other problems associated with ad-based search I won’t mention here. Personalized digital ads has now become a complicated central art form to campaigns and marketing, but when and why did it get this way?

     • Rise of ad-giants: Sergey Brin and Larry Page in their original 1998 paper (paper which presented Google for the first time to the world) did not like ad-based business models.

       “For this type of reason and historical experience with other media, we expect that advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers …

       In general, it could be argued from the consumer point of view that the better the search engine is, the fewer advertisements will be needed for the consumer to find what they want. This of course erodes the advertising supported business model of the existing search engines. However, there will always be money from advertisers who want a customer to switch products, or have something that is genuinely new. But we believe the issue of advertising causes enough mixed incentives that it is crucial to have a competitive search engine that is transparent and in the academic realm.”

       This attitude shifted (largely seemingly thanks to Bill Gross and GoTo.com in one way or another). In the first Google shareholder letter, the founders remarked that “Advertising is our principal source of revenue, and the ads we provide are relevant and useful rather than intrusive and annoying.” Ad-based search has since seemed to be taken for granted as the way of life (even for engines such as DuckDuckGo).

     While I was not in the rooms where media startup founders decided to choose ads for monetization, from what we know it seems that even champions of ads are wary of its limitations. Perhaps it is time to question the default.

   Tl;dr: Ads models bundle separate problems that may have better individual solutions. It is the default monetization model, but perhaps for a lack of a better alternative. For an industry that moves fast and breaks things, perhaps it is time to move fast and break free from ad models.

3. Governance OKRs. The problem here is that we don’t have a good incentives framework for performance outside of polls and approval. [3]

   • Consensus on need, no consensus on action. Why are there problems that a majority find important but aren’t incentivized or paid to work on (e.g. why does it take so long and so much energy for negative externalities such as climate change to be recognized and regulated)? The tempting default seems to be inaction when there is a consensus on need but no consensus on approach.

   • Governance Performance Dashboards. I will preface this with 1) governance problems are not only in the government 2) I don’t believe that optimizing metrics is the end-goal, but rather a good barometer for what’s missing 3) not everything that can be quantified should be quantified. 4) My knowledge of government systems is North American-centric.

     To my knowledge, much of the elected legislative branch relies on public opinion (e.g. election results, polls, social media) as one barometer for performance. There aren’t many other easily viewable performance metrics meant for representatives to indicate how much progress they’ve made towards their platform. While plans for 90 days in office and general plans exist, overall progress made through sponsored bills or initiatives in even a single week in politics can be difficult to track. We wouldn’t fire employees for underperforming if they were not given clear, mutually understood goals, but we fire legislators (by not electing them) even if they aren’t given clear avenues for regular feedback outside of polls. Though far from perfect, a legislator dashboard where core campaign platforms each have a quarterly quantitative metric (e.g. how many clauses drafted towards a certain goal) may be a place to start.

     On the other hand, not having clear performance metrics also prevent the public from keeping legislators regularly accountable come election time without editorialized summaries. While report cards exist (thanks to Nik Marda for some of these links: 1, 2, 3, 4), they may not come at the frequency or consistency desired. One other example for increasing citizen input outside of polls is crowdsourced policy (e.g. Stanford’s Participatory budgeting platform project.)

4. More to come.

---

[1] A note on what I mean by “worth”. There are very obvious, big problems (climate change, poverty, lack of healthcare access) that will do much more good for people than they cost. These problems are not included here, although I want them very much handled. I also tried to stay away from problems that are solution-first (e.g. insisting that tech. such as AR is the avant-garde solution to some big problem if we hack long enough on the tech. itself). I am not dismissing work that has no immediate use, nor positive externalities. You need people to come up with the abstract notion of groups and rings so a century later someone else can invent asymmetric cryptography. The intention here is to be specific enough people can start chewing away at problems that have big consequences.

[2] This is explicitly not a comment on content moderation on private platforms or decentralized web hosting. Those discussions deserve better treatment than one given here. People better informed than me have spoken at length on this issue (e.g. speakers at Josh Constine’s panel on moderation). In particular, Alex Stamos posted a thread on foundational work in this area.

[3] I would love to be proven wrong here, so please feel free to send suggestions to the contrary.

Affiliation: I’m not speaking on behalf of organizations I work(ed) and invest(ed) for. I link content where I want to, not because I’m sponsored, told to, or necessarily agree with author stances.

Thank you: Thanks to people who have subscribed! I more or less expected Substack to be a nicely formatted place to shout into the void as I try to write more in 2021, but am pleasantly surprised. Feel free to drop me a note, thoughts always welcome.

Corrections: The Google paper publication date was originally cited as 1988 instead of 1998. This has now been fixed. Thanks to a reader for noting this correction!