Public opinion about bitcoin ATM business varies, but mostly it is split into two radical sides: either “machines are not used by anyone and hence operators don’t make any money”, or “the fees charged are usually so huge, probably operators do a lot of money”. Obviously, both extremes do not correspond to reality, as if it was not profitable, there would be not that many businesses doing this long term, on the other hand, the fee size is dictated by market and there are plenty of various costs and risks associated with this business.
In this post we analyze historical data available on the website about operators growth and closure of activities. We analyze activity of total 870 operators on the website, since October 2013, when the first publicly accessible ATM was installed, and up to the end of January 2020.
The chart above summarizes information, that is used as a source for analysis — number of ATMs installed by each operator on the 1st day of each calendar month.
The first metric we check is a survival rate, which is a percentage of operators that didn’t close activity. Each operator is assigned to a group depending on how many machines in total they had at peak level: 1-10, 11-30, 31-100, or > 100.
|Maximum Locations||Closed Operators||Active operators||Survival Rate||Average Months
The left part of the table shows how many operators were closed and how many are still operating machines, with split into groups based on scale of activities. The largest percentage of operators closed activity is within the smallest “1-10 machines” group. This is expected behavior, as many operators close on small scale, means they could not scale business, or decided to sell hardware / business to larger companies (there is a centralization trend on the market for the last 2-3 years). If operators increase the number of machines in their network the probability of closing business reduces a lot. So the switch from below 10 machines to 11-30 machines group reduces the chances of closing business from 40% to 17%. And the more there are ATMs, the smaller this probability is. Among operators who reached “100 ATMs and more” level, there was no single case of quitting business (100% survival rate).
Another metric covered in the table, is the average length (in months) of activity before closing business. So for the smallest group “1-10 machines” it is 18.6 months on average when business is closed. This is within reasonable range, as 1.5 year should be sufficient to test and understand whether this activity is worth pursuing. The more machines the operator had, the longer the business was operated. This can be explained that reasons for closing business when you had 50 machines and 5 machines are different. In the former case, it is highly likely not the reason of business being not profitable or inability to scale, as operator reached already a level with a large number of machines. The reason here could be regulatory framework change, or acquisition by another larger business. And this explains longer period of activity before closing. With 50 machines it will be on average a little more than 4 years before closing. It is worth remembering that the whole industry exists for 6.5 years only.
The following histogram chart shows distribution of number of months before closing activity by 319 operators (closed ones):
It is seen that peak value is about 9-10 months, so this is a period after which many realize whether they want to proceed with this business activity or quit.
Statistics of operations at the end of first year
The following table shows data about similar activity of operators, but split based on years since 2014 up to 2019 (by start of activity) and provides a comparison of how metrics changed over time with industry getting more mature.
|Year Started||Number of operators started||Number of operators active after 1 year||Survival rate after 1 year||Average Number of ATMs after 1 year||Maximum Number of ATMs after 1 year|
Data for 2019 is not filled, while 1 year period will finish only in the end of 2020 , however, number of operators who started activity in 2019 is final now. From the table it is seen that during bear market times (like 2015, 2016 and 2019) the number of entrepreneurs entering business was reducing. Year 2017 and then 2018 as a result of December 2017 price spike had the largest number of new operators, who started bitcoin ATM business.
Survival rate after 1 year was lowest (78%) also for operators who started in 2017, probably due to many rushed to enter hyped market, and then understood what it is worth to consistently run such a business.
Average number of machines after one year of operation was consistent at 3-3.5 ATMs per operator.
While the largest increase by one operator in 2016 was from 0 to 70 machines within one year, and in 2018 this metric was even larger at 78 machines. In 2017 this number was at 40 machines maximum. The reason could be that new operators were busy with supporting existing machines operations, while demand in 2017 (especially second half of the year) was probably largest over previous periods. Operators could concentrate on quality growth, rather than quantitative expansion. In the periods of bear market, operators increase volume through expansion of number of installed machines. In this way, they also get prepared in advance for the next growth cycle to better serve demand.
The following charts illustrate distribution of the ATMs number by operators at the end of 1st year after operations started:
It is seen that absolute majority of operators still had only 1-2 ATMs at the end of first year of operations and only few operators scaled business to tens of machines.
In the article we check some metrics of operating bitcoin ATM business based on information on the website. In general, operating a network of bitcoin ATMs is like any other business. It might be successful or not. Out of all operators added on the website, 37% closed and discontinued business. However, with scaling business the success rate increases and business becomes marginally more profitable. Out of 17 operators who reached 100 and more ATMs in their network, none has closed business.
After 1 year on average there are 3 ATMs run by each operator, while majority of operator still have only 1 ATM at the end of year one. On average it requires 8-10 months for operator to decide and discontinue business, while the larger operator becomes the longer it takes on average before closing operations.