I’ve been recently thinking about the potential of newer technology to accelerate a phenomenon that I call “source consolidation.” It seems to me that this will only speed up even further as technology progresses, for example as generative AI becomes more powerful and prominent. In this post, I talk about what I mean by source consolidation, how it relates to tech and generative AI, and what its implications can be.
What Do I Mean by Source Consolidation?
To discuss this, we need to discuss how discovery works in the physical world, and how it is different in the digital world.
Imagine you want to buy something, say some groceries or food. In the physical world, you’d go to a physical store, and you’d walk around in the store until you found what you wanted. However, as you walked around, you could see all sorts of other products being offered on the shelves that you walk by. You essentially discover these products in the process of shopping for what you originally wanted. As you discover these offerings, you can potentially become enticed by them, and you could potentially walk out of the store with a lot more than what you initially intended to purchase.
This kind of discovery is necessarily an ever-present part of physical shopping: to get to the item you wanted, you need to walk through some physical space, and unless the item happens to be something sold in the front of the store, you’re going to discover other products along the way as you walk. This effect is enhanced further by the fact that stores often congregate within shopping districts, so that you’d discover more stores along the way as you walk to your intended one.
However, discovery can work differently online, in the digital landscape. There, you often have the ability to search, where you can directly reference exactly what you’re looking for and see results for just that, without having to do the online equivalent of walking by other products and shelves in the process of getting to yours. Of course, some websites can still try to present you with recommendations for other items you’d want, which they’d want to do in order to increase sales. However, these recommendations are generally determined by algorithms that are trying to assess how similar or complementary those items are to what you’re already buying (the algorithms try to analyze your buying behavior to determine what else you’re likely to purchase too.) Even with such recommendations, over time your buying habits would probably congregate towards a smaller set of items and search queries that you buy more repeatedly.
Furthermore, e-commerce sites like Amazon and eBay expose a much wider variety of sellers around the world to a much wider variety of buyers, so you may be less sure of the quality of unfamiliar sellers online. Thus, you’re probably more likely to keep going with some trusted sellers, rather than try out new sellers and new products every time that carry greater risk with them.
What this means is that over time, your sources for the items you own — the products you buy online and the sellers you purchase from — become consolidated into an ever-shrinking group. This is what I mean by “source consolidation.” I’d imagine that online, you’re more generally more likely to reorder the same basic things often than you are to buy newer and more exotic products — especially compared to, say, if you were walking around inside a novelty shop.
This example discusses shopping, but the effect of source consolidation applies to all sorts of other content too. For example, think about information, like the news. If you trusted a certain site to be your source of information, you may just keep putting that one site into your address bar, as opposed to going out of your way to find other websites, which could potentially have looser journalistic standards. Why take that risk when you already have a perfectly good source? Then, your sources for news online end up consolidating, as opposed to if you typically got your news by walking by a magazine rack.
Filter Bubbles
This idea of source consolidation is related to the phenomenon of filter bubbles. People often talk today about how social media can enhance echo chambers and create bubbles in which other perspectives are filtered out. This is in many ways similar to the point we made about product recommendations; in general, algorithms that are designed to maximize similarity and complementarity can end up enhancing these kinds of filter bubble effects. This can apply to information, like on social media, or to products, like on e-commerce platforms.
However, I think the phenomenon of source consolidation goes beyond recommendation algorithms. For example, as we discussed above, the inherent nature of a search bar makes it so that the sites you get your content from end up consolidating over time. And online, there is so much content out there, being produced by literally billions of people. It seems like it would be impossible to organize access to this content without some kind of search feature. As a result, it seems that the inherent nature of online activity would lend itself to source consolidation, regardless of recommendation algorithms (although increasing the diversity of recommendations could potentially reduce the level of consolidation.)
Furthermore, with a search interface, the cost and effort for switching sources is higher, since in order to switch you need to find out about alternative sources, and discovering alternatives is not immediate if you don’t even know what other sources to search for. Thus, even if the few sources you use do fail in quality sometimes, unless the failure happens often enough, you may be less likely to switch from them.
Technical Products
The effect of source consolidation can be even greater for technical products that require more specialized knowledge to configure and evaluate. In that case, consumers may decide that they would not spend the time and effort to compare their current source with another one, especially if making such a comparison entails deeper study and analysis, which may require a greater technical background. This can apply to say computer-related technology as well as other technologies that require specialization to configure and evaluate.
The Paradox of Choice
In cases where source consolidation is less prevalent, consumers may perceive the resulting spectrum of choice as a greater benefit. However, this may not always be the case: there is a much-discussed “paradox of choice,” where it actually may be to a consumer’s advantage that there aren’t too many “levers to pull” or “things to decide” when trying to choose a new source.
Effect on Monopolization
If sources become consolidated, then it stands to reason that those sources end up drawing most of the user base, away from their competitors. This makes it so that it’s easier for online providers to acquire a much larger slice of the pie more quickly, leading to greater monopolization. If someone is used to a particular source, why would they switch sources, especially if the effort to do so is much higher?
Opportunity for Stronger Niches
Contrasting with the previous point, an opposite effect can also apply online: greater connectivity means niche groups can come together more easily, expanding the reach, success, and longevity of niche platforms. These can end up taking users’ attention away from monopolistic providers, spreading the focus of Internet users collectively among multiple niches.
Further Source Consolidation with Generative AI
Generative AI may have the potential to accelerate source consolidation further. Earlier, interfaces like Google Search would present 10 or so links, and despite the more targeted focus, consumers were as a result exposed to more sources. However, if a generative AI product like a chatbot were to generate a single answer from a single source more often, that can intensify the effect.
