Independently, I’ve spent the better part of the last 25 years analyzing data and obsessing about digital culture. Up until joining SimilarWeb as a senior researcher 6 months ago, these were two parallel lines in my life, never intersecting and never coming close to each other. The position at SimilarWeb, therefore, held within it a promise, or at least an opportunity, to finally bring these two parts of my life together, but it also threatened to trivialize something I am passionate about.
The data I analyzed over the years came from a variety of sources: diffraction experiments and numerical simulations of electronic dynamics in quasi-crystals in my graduate studies, and signals of medical sensors as an R&D physicist. Regardless of the data source, at some point in each of these cases the abstraction took over and I no longer felt any substantial connection to the underlying process.
Joining SimilarWeb had me worrying that the same thing would happen, despite that fact that I am genuinely fascinated by “The Internet” and have been ever since I was forced to read about in printed magazines.
Happily, I can now say that my fears have been refuted.
When my twelve year old daughter and her friends started using the app musical.ly a few weeks ago I smiled to myself in satisfaction. I had seen it coming. I noticed the rapid growth of this app and saw the quick growth in installs, so by the time it started gaining popularity in her class I was already aware of the data behind its rise.
In a sense it felt like 2009 all over again. Back then, Twitter was a big issue in the blogging community, we all mocked it for being silly and useless, but then, one by one, we joined and got addicted. Yet, whenever I mentioned Twitter in conversation to real-life acquaintances, people were puzzled and had no idea what I was talking about. I had no solid data to support my gut feeling, but I trusted my Internet-savvy community of bloggers and early adopters. If something is trending among us, it is bound to become huge. A few months later came the Ashton Kutcher-CNN competition to hit one million followers and the same people, still puzzled, also admitted I was up to something.
A lot has changed since then, and the people setting digital trends now are no longer my peers in some imagined community. Though we like to keep thinking of ourselves as sophisticated early adopters, none of us even understands what Snapchat is all about. Snapchat, by the way, actually passed Twitter on US installs this past month.
Turning back to musical.ly, when I was working on creating an algorithm for daily app installs, I used several estimation techniques mixing mathematical models, data from our mobile panel, and publicly available data. No matter what approach I took, this obscure app, which I had never heard of, kept showing up at the top, consistently staying there, week after week.
The consistent rise of musical.ly’s metrics over a period of several months, as well as its high usage rate (20%-25% of app installers use it on any given day) could only mean one thing, I was witnessing the birth of an actual phenomenon. Musical.ly most probably isn’t the next Twitter and I might not even remember it in a year, let alone seven years, but at the very least, old and outdated as I might be, with data replacing intuition, I once again found myself ahead of the zeitgeist.
Such things happen on an almost daily basis – we come across a mystery, a possible irregularity in the data, which forces us to investigate its causes. Sometimes it’s an actual trend our algorithms caught before we learned about it, and sometimes it can be the result of a new technology or a new pattern of behavior which our algorithms were not yet trained to deal with.
Whenever something like this happens I am exhilarated by the opportunity to dive into the details, combining technological, sociological and behavioral analyses in order to realize what is actually happening, to stay in touch with the currents of digital culture underneath the mountains of data.