A brief history of Bitcoin “wallet” growth

There has been a lot of investment and press coverage of the overall Bitcoin ecosystem.  So what kind of growth have some of the larger companies historically had?

Even though it is not an accurate measure of growth or adoption (see Measuring Interest and Not User Adoption), a lot of discussion on social media typically uses self-reported “wallet” numbers as a valid metric for traction.  Ignoring the fact that there is nothing in the network that can be described as a “wallet” (there are no real “payment buckets,” since addresses are essentially just UTXO labels), for simplification purposes, we will talk about what are typically referred to as wallets.1

A brief history

As mentioned in a working paper last spring, Coinbase began 2013 with ~13,000 wallets and on February 27, 2014 announced it had reached 1 million.

Similarly, Blockchain.info had roughly ~13,000 wallets in August 2012 and reached 1 million in January 2014.  On April 14, 2014, Blockchain.info reached 1.5 million wallets.

Yet it is unclear how many are active or actually have any bitcoins in them (similar uncertainties surround Coinbase wallets).  More on that later.

Fast forward to the present day, Blockchain.info recently announced that it had 3 million wallets and Coinbase now has 2.5 million (note: the about section on Coinbase also states there are 2 million “users” though that is unclear if they are active, KYC’ed users with an actual balance or just a registered empty account).

Altogether, Coinbase purportedly added 1.5 million new wallets over the past year and Blockchain.info supposedly doubled its own wallets.

Sounds like real consumer traction?

Or, maybe not.

Why?  Because there is no cost to open or create a wallet.  In fact, for “best practices” users are supposed to use only one address per transaction due to privacy and security concerns.  Thus, consequently the growth in wallet creation could be a skewed metric.

Internal usage

According to media reports, merchants accepting bitcoin for payments globally increased from ~21,000 in January 2014 to now over 100,000 as of February 2015.  Of that total, Coinbase states it has 38,000 merchants and BitPay claims 53,738 merchants accept bitcoin payments through them.

What does this “growth” actually look like?

coinbase offchain transactions

Above is a chart covering the past year from Coinbase which illustrates the daily off-chain transaction volume, the transactions that take place within the Coinbase database.

While it is unclear if all of this activity represents merchant processing, vault movements, etc., the trend over the year is actually relatively flat.  Perhaps that will change in the future.

Can we be sure that this flatness is missing actual merchant activity?

Four-and-a-half months ago, in October 2014, Brian Armstrong and Fred Ehrsam, co-founders of Coinbase, did a reddit AMA.  At the 31:56 minute mark (video), Fred discussed merchant flows:

One other thing I’ve had some people ask me IRL and I’ve seen on reddit occasionally too, is this concept of more merchants coming on board in bitcoin and that causing selling pressure, or the price to go down. [Coinbase is] one of the largest merchant processors, I really don’t think that is true.  Well one, the volumes that merchants are processing aren’t negligible but they’re not super high especially when compared to people who are kind of buying and selling bitcoin.  Like the trend is going in the right direction there but in absolute terms that’s still true.  So I think that is largely a myth.

What about Blockchain.info?

blockchain my wallet transaction volume

Above is a chart measuring the internal transaction volume over the past year of the “My Wallet” feature (the product name of the user wallet) from Blockchain.info.

Earlier this week, Blockchain.info claimed that “over $270 million in bitcoin transactions occurred via its wallets over the past seven days.”

But this is probably not accurate.  Organ of Corti pointed out that the 7 day average was indeed ~720,000 bitcoins in total output volume (thus making) the weekly volume would be about “5e06 btc for the network.”

Is it valid to multiply the total output volume by USD (or euros or yen)?  No.

Why not?  Because most of this activity is probably a combination of wallet shuffling, laundering and mixing of coins (e.g., use of SharedSend and burner wallets) or any number of superfluous activity.  It was not $270 million of economic trade.

Blockchain.info’s press release seems to be implying that economic trade is taking place, in which all transactions are (probably) transactions to new individuals when in reality it could simply be a lot of “change” address movement.  And more to the point, the actual internal volume looks roughly the same as has been the past few months (why issue a press release now?).

Is there another way to look at this?

blockchain my wallet number of transaction per day

Above is a chart from Blockchain.info that visualizes “My Wallet” transaction volume over the past year.

While Blockchain.info has seen transactions per day roughly double over the past year (from 25,000 to 50,000), without doxxing where those bitcoins go, it cannot be said that a doubling of economic activity, or that bonafide consumer traction has taken place.

Has there been any “exponential” growth, adoption or traction?  Probably not.  Again, perhaps that will change, but consumer usage could simply continue to grow at a linear fashion or maybe even less as well.

There may be a number of reasons, perhaps the average consumer is still someone who buys and holds bitcoin as a speculative investment and has no need to actually spend it with the available merchants.  But this is a topic for another post (see also Zombie activity).

ChangeTip

ChangeTip was founded on December 17, 2013.  It is not generally seen as a wallet, like the services above, in fact it currently bills itself as a micropayment service (e.g., “tipping”).  However, users need a ChangeTip wallet — which is provided for free through its platform — in order to perform their tipping services.

While their “Offsite storage wallet” (cold storage) is publicly accessible, below are three charts culled from Changetip real-time usage stats which has been broken the last couple of weeks (or the API they were collecting data from is broken; or both).  The time period is from between December 6, 2014 and February 17, 2015, covering ~73 days including Christmas and BitPay’s “Bitcoin St. Petersburg Bowl.”

changetip total number of tips sent

The chart above visualizes the total number of tips sent on the ChangeTip platform .  In just over 2 months it increased from: 119,740 tips to 187,071 tips.  During this 73 day period, approximately 67,331 tips were sent which is roughly 922 per day.

changetip total usd tipped to date

The chart above visualizes the total USD tipped to date (at current exchange rate).  During this time frame it increased from: $54,767 to $111,963.  Thus $57,196 was sent in tips during 73 days, roughly $783 per day.

changetip total numbers of users

The chart above visualizes the total number of ChangeTip users during the same time frame.  It increased from 45,851 users to 67,469 users.  According to this data, 21,618 users joined ChangeTip during 73 days, which is approximately 296 new users per day.

Altogether this comes to a grand total as of February 17, 2015 — 67,469 users have sent 187,071 tips totaling $111,963.  The average user has sent 2.7 tips altogether, with each tip worth about $0.60 (just under 60 cents to be precise).

Perhaps this trend will change — in addition to its usage on Twitter and Reddit they have added support to Slack and Youtube.

But then again, maybe tipping is not a really accurate, useful or desirable signaling mechanism (recall that micropayments is not a new idea).  And while speculative, a lack of traction could be one of the reasons why — after 3 months since Coinbase first launched their own — it recently dropped their own tipping feature (e.g., the engineering resources consumed more than the service generated).

Future research and conclusions

What about Android Bitcoin wallets?  Last October a github user put together a short comparison of the top 10 Bitcoin wallets by number of downloads.  What we saw then was a power law distribution: growth in downloads among the top 3 but a relative plateau for others.  More striking was that there was linear growth, not exponential.  Future research should also take into account the corresponding amount of deleted wallets and inactive wallets.  Note: last May at the Dutch Nationaal Bitcoin Congres, Mike Hearn described this comparison of downloaded vs deleted wallets at length, see his presentation (video) starting at 11:30m.

Bitreserve, which incidentally also launched in October 2014, provides a public transparency stats page which could serve as the beginning of a “best practices” for the industry.

Why is this important?

We have previously looked at BitPay data (which was flat).  Circle and Xapo have not publicly released much data at this time (incidentally, breadwallet is actually ranked higher at #4 in the Apple Store than both Coinbase and Xapo).  Yet from the data above it is increasingly clear that actual user numbers should not be conflated with wallet creation numbers.

Aside from movement into P2SH addresses, it is hard to really say where large, sustained organic growth is occurring.  Perhaps it is only a matter of time, maybe it is “early days” as some say.  Or maybe it is a reflection of other economic development constraints.

Update:

I received an email from Andreas Schildbach, creator of the Android Bitcoin wallet, and a portion of it is posted below (with his permission):

Install count is at 700k. Perhaps an interesting metric is that on GitHub, it’s forked 384 times (and starred 371 times). A lot of these forks made it to the Play Store.

Update 2:

I received an email from Wendell Davis, creator of the Hive Wallet.  According to him, all the Hive Wallet stats are open and accessible.  He also pointed to a similar, smaller discussion on reddit last fall.

Update 3:

BitcoinPulse has been tracking the total amount of downloads for the Satoshi bitcoind client (the reference client); over the past year there has been a linear increase in downloads.  Arianna Simpson pointed out that MultiBit, as of March 2014, had at least 1.5 million downloads.

  1. I would like to, again, thank Andrew Poelstra for crystalizing this point for me. []

2 thoughts on “A brief history of Bitcoin “wallet” growth

  1. Speaking of conflating, just cause a honey badger don’t care doesn’t mean he don’t count… and accurately.

    Great insights and accountability. No doubt such verifiable reporting strikes fear into anyone attempting to stretch their numbers through vague mischaracterizations.

    Great objective work and valuable to anyone attempting to assess the market opportunity and competitive landscape.

  2. Breadwallet is ranked higher because it’s so easy and fast to set up. You just download it and bam it’s usable. Coinbase and xapo you need to sign up, put in your email, come up with a password. Verification is most cases? Breadwallet all you need to do is copy the phrase or even in most cases do it when you’re home.

Leave a Reply

Your email address will not be published. Required fields are marked *