Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Four decades ago this Saturday, a computer company called Symbolics Inc registered the first ever dotcom internet domain name. It’s still alive today as an online relic, despite the company itself going bankrupt long ago.
If only it had invested more heavily in its pioneering website! At least, that’s the obvious conclusion from a fascinating new report from Wolfe Research’s quantitative analysts.
The team led by Yin Luo scraped the current and historical websites of the almost 800 companies that have been in the S&P 500 since 2013, a chunky 190gb data set of over 85mn individual web pages. They then set to work trying to find out if there any usable investment signals from a company’s web presence.
And lo, there was. FT Alphaville’s emphasis below:
We introduce two sets of stock-selection signals for a total of 64 factors using corporate website data. The first suite of features is primarily based on web meta data, such as the number and size of web pages, broken down into MIME types (e.g., HTML, octetstream, PDF, video). Empirically, companies with informative websites are associated with future outperformance.
Then, we conduct topic modeling on web contents, leveraging our QesBERT finance domain-specific language model. Firms that are leaders in their respective industry, raising unique topics, tend to deliver superior returns in the future.
. . . Finally, we develop a composite stock-selection model, using web meta data and contents. The Wave investable strategy achieves an impressive after-cost Sharpe ratio of 2.1x and a CAGR of 16% since 2020. As the quality of web data improves in recent years, model performance has also markedly advanced.
In other words, the larger, the deeper and the more sophisticated a company’s web presence is, the better its stock does.
This isn’t just a backward-looking thing. The signal derived from Wolfe’s database seems to have strong predictive power. And, rather than fizzling out as differences between corporate websites narrow, it has been getting stronger.
Wolfe’s quants highlight Nvidia as a good example of a company whose evolving web presence has proven a harbinger of good times, rather than just reflecting it.
. . . We can clearly see the focus of Nvidia’s business has transformed greatly, from merely a GPU producer for a small niche market of video game machines to today’s dominance in global AI advancement. A quick glance at its website tells us so much about Nvidia’s transformation. That is the power of conducting historical web analytics for investment.
Of course, this is not an iron law of internet economics. Berkshire Hathaway’s website is famously sparse, and it’s done prrrrettty well over the years. Nor does a website necessarily mean much for investors themselves. Renaissance Technologies’ website is also delightfully dull. And as with every systematic investment signal, this will eventually fade as people exploit it.
However, if there is one site that Alphaville would go sell-our-kidneys-for-OTM-calls long it would be this one, set up by the office of Nintendo’s founding family.
If you haven’t seen it before, do check it out. Words cannot do it justice. But the Yamauchi No. 10 Family Office had a good stab at it:
Seeking stability only clouds your soul. Fear of failing only kills seeds of innovation. We must create a more exciting future where people feel free to dream and leap into a world of possibilities. We must create a freer future where people are eager to truly live . . . We dare to dream. We dare to leap. We dare to live. And we dream that you do too.
