Bitcoin signals and stats in mainstream media

There has been a lot of news lately surrounding a certain study funded by a third party. The study attempted an analysis of the Bitcoin blockchain for a number of purposes and came up with a few startling numbers that a lot of the “mainstream media” (if Bitcoin has such a thing as mainstream media) have been running wild with. The most commonly cited is that 78% of all coins are supposedly being stashed under the digital mattress. I think, however, that it’s fair to qualify the condition of that 78 % of coins as a bit more complicated than all that. The study fails to account for the fact that a huge portion of all coins ever mined was mined back when they were essentially worthless. 10,000 BTC bought a pizza, once upon a time, and CPU mining was the norm. Back then, losing several thousand coins to a hard disk failure or otherwise losing a wallet was like a $5 bill going missing from your wallet – annoying but not worth losing sleep over. So how many of that 78 % of coins are actually spendable but sitting idle and how many are lost to time? Like so many things with Bitcoin, the real problem is: we just don’t know.

People forget when dealing with Bitcoin that while it was designed with a certain kind of transparency in mind and it is infinitely more open and study-able than traditional currencies, it was also built with privacy in mind and to that end it’s very hard to get certain kinds of information about crypto signals and Bitcoin transactions. Of that 78%, we can certainly say that some are lost, some are savings and some are likely the cold storage accounts of large exchanges or merchant service companies – but we don’t know (and for the most part can’t know) which. All those stale coins are like Binion’s silver, clever financial analysis can certainly indicate that something’s missing, but the very nature of a thing being missing means we haven’t a clue where it actually is.

It goes beyond the design of Bitcoin, too. Bitcoiners themselves are often elusive beasts, difficult to track. Allow me to share a personal anecdote.

To understand my surprise, you must understand the difference between log analyzers like AWStats and client-side tracking like Google Analytics. Google Analytics is a little snippet of JavaScript that’s embedded into each page on your site and when someone views your site their computer runs that JavaScript and hands over all kinds of useful information like what kind of browser they’re using, how big their screen is, whether they’re on a mobile device and much more. The wealth of data gathered by Analytics can be quite valuable, but the method is flawed. Not every browser runs JavaScript the same way and some people disable JavaScript or otherwise block such tracking attempts, so some percentage of visitors to a given site simply won’t be counted – you can usually assume that your Google Analytics numbers are a bit on the low side. Log analyzers, on the other hand, dig through your web server’s logs, counting each and every time the server says it delivered a page to someone. Server side logs usually have far less useful information, but the information they do have tends to be more accurate. Armed with this knowledge, let me explain why my jaw hit the floor: Google Analytics said that I’d served up a bit over 60,000 pages in the month of September – AWStats was reporting just over 250,000.

Now my example may be a bit long-winded and I apologize for that, but it does go a long way toward making my point: Bitcoiners, you are some of the hardest people to gather information about I’ve ever met. If I can’t even get a solid number to tell my sponsors how many of you are viewing my stuff in a given month, I find it unlikely that any analysis of this community or our behavior will even be close to accurate.

But this can work both ways – sometimes a lack of data is just as telling as an abundance of it. I would also posit that this could make an interesting metric by which to judge Bitcoin adoption: Analytics parity. As more individuals outside of the highly-technical JavaScript-disabling community adopt Bitcoin, the percentage of you that remain uncounted should begin to decrease. Gather enough data points across enough Bitcoin sites with some non-Bitcoin sites as control and compare the percentage who go uncounted. If my site alone is anything to judge by, we’ve got a long long way to go.