In this essay I assert that the level of personalisation we've seen thus far on the web is insignificant in comparison to what is coming. Additionally, I describe a transition to a new model of consumer tech and more broadly, data services. In so doing, we see how the biggest tech companies of the day will sink into irrelevance, lest they evolve. We get a sense of which organisations will benefit from the transitition and I give a short treatment on how I think you can personally protect yourself and benefit from the revolution.
There are 3 principles around which this revolution turns—
Today, a select few companies have accumulated the right to build on our personal data. You know the names – Google, Facebook, Amazon, Apple, Netflix etc. We've exchanged our personal data, a machine-readable form of our preferences, for services – some more frivolous, some less.
This has generated a landscape in which our data is split across dozens of different pockets.
It's precisely this model which limits the potential of the current iteration of the web. There are countless initiatives underway which are driving a paradigm shift on the axes of these principles. I want to describe a few of them from my perspective here, and try to give a picture of how this all might culminate in the web's next big leap.
Thus far, data has aggregated around service providers. Google, Facebook et al have built their monopolies around the fact that you don't know how to store and maintain your own digital identity. So you spread your data across the services you use. This has enabled them to accrue all the data and has meant that virtually noone else can compete. They know who you know because you told them who your friends are. They know what you want because they've hired the top people in the world to build algorithms out of the data we've given them.
This has been the greatest advantage Google, Facebook et al have built. But at this point it's also revealing a gargantuan weakness.
Concretely, you're already seeing this with DeFi. This new class of financial services shows quite clearly where the future of personal data is going. You bring your wallet – i.e. your personal data store – and selectively unlock it to bring to the UI to life. Beyond the quality of the UX and featureset, there's zero lock-in.
Google, Facebook et al haven't only achieved their lock-in through UI though. Largely it's come through being able to develop the strongest algorithms. Having the most data has been the unfair advantage of the past 20 years – and that's how these companies have built their "moats". By the way, this extends far beyond Google and Facebook – this is the default consumer data model of the web, and shows up in almost every service model.
So I think what we're seeing is a transition from this model where multiple businesses store our data and hold the keys to it, to a model where the data is stored in a decentralised way and individuals hold the keys. Projects like Urbit, 3Box and Blockstack are giving a glimpse of this. Instead of landing on a site and using MetaMask to fill it with your financial data, you'll do so with your personal data store. I was actually involved in creating one of these personal data stores, very similar to 3Box, in 2013/2014 with Jens Elstner called Keepzer. Alas we were much too early – #timingmatters.
This new model – one of personal data stores and self-sovereign data – will likely create a huge opportunity to customise the way the web looks and works for you. No longer are you only able to see your data through the lens of the UI that Facebook and Twitter make available to you. If you're unhappy with the fact that the Twitter Protocol forces you to filter your timeline by its awful Top Tweets algo, switch to its forked, alternative UI and have Latest Tweets by default.
This radical change isn't just limited to "traditional" Web 2 services though. Increasingly, you can imagine this touching a huge range of other, critical areas of our lives. For example, what if you could rock up at a hospital with all your healthcare data from birth? What if you could rock up to a school with your entire learning history? Every test question you've ever got wrong, every one you've ever got right1. This kind of thing is a reality on a web which reorganises data around the individual, rather than the service provider.
Honestly, this doesn't only apply to personal data. It equally applies to businesses. In my own business I constantly wonder why the hell all my data is strewn all over the place – conversations on Slack, financial records on Xero, inventory data on Shopify. This stunts the potential of the services I use, means that my data cannot be used as collateral and makes the business wildly less sellable, IMO. There is so much that's wrong with this and I kind of can't believe nobody is talking about it, but maybe I'll save that for another post.
On top of these changes to the visual/interaction model of the web, we are also likely to see a huge change in the way data is processed. Data processing is likely a much bigger deal than UI-level changes because this is really where the current lock-in lies. I think the biggest innovation here is around multiparty computation and trusted execution environments that you see on the likes of Enigma.
With trusted execution environments, you can process an individual's data without the people processing it being able to see it. This isn't merely a privacy benefit – it's a radical shift in private data economics.
The critical takeaway is that, given this data model, people are going to be able to truly lease out and sell access to their data. This can't happen currently because 1) people don't have their personal data in their control and 2) if they did have it there would be no way for them to share it without the data processor copying it – immediately voiding its scarcity and market value.
It's precisely this ability to maintain scarcity which makes Web 3 so compelling. It allows a much broader set of data processors to build their models. The data of the world's individuals becomes an enormous, rich bed of potential for building radically more personal and targeted data services.
That's all great, but the real shift this creates is that it makes it possible for individuals to generate value from their data. The intuition is simple – I'm a developer and I want to make a valuable new machine learning model to help people sleep better. I'm going to sell this model as an app. I want to use your data to train my model. I do so, in a secure 3rd party environment where I cannot see the data and dilute its value. The training is complete, my model is better, your data retains its scarcity value. I am now more likely to be able to sell my model. There has been an accrual in value, and you have contributed to that. So you should be rewarded, obviously.
The big question, which I find fascinating, is precisely how you should receive that value. My intuition, which I explored heavily with Jens Elstner, is that it will be based on an equity share model much more so than cash in hand. That is, you as the data provider will be rewarded with a future share of my returns in the form of equity – likely tokenised – in The Business of Me. This is because there's a time delay between me supplying my data and you realising its value in the sleep improvement market. Anyway, this is something else I'd like to give a much deeper treatment to in another article.
You can naturally see how this changes the whole economic equation for the individual. Why on earth would I pick a Web 2 service, which effectively steals the value of my data, over a Web 3 service which helps me personally realise its value? Web 3 will lead me to treat my data with an entirely different reverence. My digital behaviour suddenly becomes a source of material value. Time will tell how much spending power this value brings. At minimum though it should lead to a much great ability for me to contribute to the creation of services which benefit me personally. This is the source of hyperpersonalisation in the new era.
I've become very conscious of this economic equation since starting my business. It's so difficult to spend money as a consumer 2. What I mean by this is that there's no way to measure a personal return on consumption. As a business, I can say "this behaviour made money in the past, so I can safely assume it'll do so in the future." Kelly-criterion style. But in my personal life, there's no way to do this 3. Because if I pay to see a film, there's no objective value on the other side. Note that I'm not talking necessarily about financial value, but domain-specific value – something I can cover at another moment.
Everybody knows how much harder it is to start a consumer business. Consumers are "irrational". With the new model of the web I sense that this might change. It doesn't necessarily mean that our grannies are going to start calculating their "data returns" on this brand of pineapple versus the other, but I suggest there's a decent chance that the recommendation engines of the Web 3 future will be calculating precisely this on their behalf.
There's incredible potential for applied research – aka data science – in this new data processing model. Potential on a scale the world has never seen before. I saw the Apple Health study recently with 400,000 peoples' data sourced from Apple Watches. You get medical researchers drooling over these kind of numbers – it's difficult to get 40 good patients for a heart study 4. But imagine if researchers are sitting on an enormous substrate of health data – data of incredible quality, ensured by the economics described above.
The example I'm most familiar with is genomics research. A lot of my thinking about this came from my involvement with GeneOS, a project for which we won the EOS Hackathon last year.
A separate, equally interesting dynamic is to consider what this means for the supply-side of data services. What motivation do businesses have to provide services that create high quality data? For example, why would 23andMe sequence your data if they can't sell it on to pharmaceutical companies 5?
Well, the business models for these supply-side cats are really interesting. They will likely operate on cryptonetworks, like Bitcoin. And they'll operate as miners. So they'll be rewarded for 'mining' personal data on your behalf. In all likelihood for their efforts they'll receive stake in the network as well as stake in your personal equity, which you'll use to pay for their services.
With GeneOS, we were creating a personal genomics data network. The supply-siders of the network were genome sequencing laboratories. They sequence your genomes in exchange for 3 things – tokens in the network, tokens from your personal equity, and maybe some cash too. Data scientists then pay to use the protocol to improve their models on your data. The protocol takes a fee, generating cashflow for the network. The network's equity increases, realised in its token. In all likelihood, the individual's personal equity also appreciates because the data scientist's model begins to generate cashflow as it's sold in their app. So the sequencing laboratory makes money when—
I don't think it's as simple as betting against all the current consumer tech companies and buying everything-Web3.
The current consumer tech cohort will likely continue to make incredible returns for upwards of a decade. Microsoft continued to reward its stakeholders handsomely despite losing out to the web and mobile. And don't bet against these companies leveraging their resources in the Web 3 world. Microsoft has also reinvented itself very successfully with its cloud and productivity offerings. Imagine how Google and Facebook might clean up on a data marketplace like Ocean Protocol. Or a Bloomberg with its trading data on Numerai.
Something to be wary of, though, is that these companies' "resources" are often actually your resources. Google and Facebook have built their monopolies on you. With Google there's often been speculation about how it might "become evil". It's worth being concerned that it will when its business meets its first true threat.
It feels like strange investment advice, but given what I outlined above, I think one of the most astute investment decisions going forward may be to move your private data stores into your own control. Learn how to publish on a self-hosted blog. Move your audience into a format that's actually in your control – email maybe? Adopt Web 3 services as soon as you can, and have the big companies delete your data.
In terms of actual investing, there are a large range of areas here to be extremely excited about. They run the gamut from data creation, through storage, processing and the area which I somehow haven't really touched upon in this essay, effectuation – software services that get shit done on your behalf. Concrete investments weren't the focus of this essay and I don't really think I have bandwidth to weave it in here, but will make it the focus of future work.
2 What does this mean for central banks and governments who are in large part responsible for growing consumer demand? What happens when consumption and demand finds a natural market-mediated balance?
3 Obviously there's internal satisfaction to the stuff we buy and do, and these will remain priceless. Priceless, objectively unmeasured.
4 Turns out that heart data wasn't actually that valuable because the data is hugely skewed towards wealthier, younger demographics. Doesn't degrade my point though.
5 Yes, this is where they make their money.