How quickly does the structure of digital media use change, and how well have incumbents been able to defend their position and remain competitive through waves of technological innovation and changes in user behavior?
After an interesting workshop discussion yesterday I had a look at this question using data on share of time spent online in the United States as the metric. The slide below provides an overview.
The top ten platforms by share of online attention together account for about half of time spent online (a “moderately concentrated” market by normal HHI conventions).
These platforms are on average more than twenty years old.
Four platforms with roots back to the 1990s/web 1.0 (Yahoo Mail, Google Search, Amazon, and Outlook Live) account for 9.73% of all time spent online.
Five platforms with roots in the 2000s/web 2.0 (Gmail, YouTube, Facebook, Google Docs, and Reddit) account for another 39.23%.
No platform launched in the 2010s has cracked the top ten.
ChatGPT is the only platform launched in the 2020s that have cracked the top ten, and it accounts for 1.47%.
Incumbents tend to struggle with disruptive innovations (Microsoft is the only company founded before the internet with a platform in the top ten, and no telco or legacy media company feature), and it is possible that technological change means that, as a leaked Google memo from 2022 put it, ”We Have No Moat: And neither does OpenAI”.
That said, as long as there are economies of scale, network effects, and data network effects – plus sometimes (ab)use of market power – many incumbent platforms have over the last thirty years been able to maintain their positions. (This has been an argument Matthew Hindman in particular has made in academic research in this area, e.g. here.)
The last thirty years thus simultaneously demonstrate (a) that real, structural change does sometimes happen, and can happen fast (as in the 1990s and then again the 2000s), but also that (b) while some end up much diminished when overtaken by the pace of change, many incumbents are able, by hook or by crook, to maintain their position.
Thanks to David Lazer & co for their work at the National Internet Observatory which I have relied on here.
(Added: I should have made explicit above that the above data is on desktop browsing, so there is a significant issue with capturing mobile, app-first 2010s social platforms and games, skewing the picture. I still think/hope it is useful.)
rasmuskleisnielsen.net 