Robinhood Outages Anger Customers

Heavy trading on Wall Street apparently was too much for the Robinhood app to handle on Monday, once again.

The company reported issues with equities, options and crypto trading just before 10 a.m. Eastern Time. By 10:25, most of the problems had been resolved.

Trading on Robinhood has been functional for new orders with the exception of fractional equities since at least 10:25 AM ET. We’ll continue to update our status page with the latest https://t.co/mON07oWvHy.

— Robinhood Help (@AskRobinhood) March 9, 2020

The incident is a case of bad deja vu for investors using the app.

Whatever the problems were at Robinhood last week, they didn't seem to figure them out in time for today's turmoil. https://t.co/ZJ8NqMNFNY

— Nathaniel Popper (@nathanielpopper) March 9, 2020

Robinhood customers potentially lost millions of dollars when the company’s app went down last Monday, March 2, the day the Dow had an historick uptick.

Another outage occurred for a few hours on March 3.

“A service like this that is unavailable during big market movements is a liability and not an asset,” noted Rob Enderle, principal analyst at the Enderle Group.

“People divest liabilities,” he told the E-Commerce Times.

Angry customers already were thinking along those lines when last week’s outages occurred.

#RobinHood bunch of crooks. Can’t sell my positions… for last hour. Server not working… I’m clearing all of my money out today and closing the account….. pic.twitter.com/r0T5flOn99

— PDXOptions823 (@PDXOptions823) March 2, 2020

Closing account and staying with my other brokers tasty trade and TD. I assume everyone else is doing the same. Robinhood has always had problems and today was just an enormous fk up. Lost a lot today because I couldn’t close my position.

— Moses (@Moses795k) March 3, 2020

A Twitter feed advocating a class action lawsuit against Robinhood — Robinhood Class Action — had more than 7,600 followers as of mid-day Monday.

https://t.co/uDcHB0qoWR

— Robinhood Class Action (@ClassRobinhood) March 9, 2020

Robinhood blamed last week’s outage on excessive stress on its infrastructure caused by an unprecedented load, which triggered a failure of its DNS system.

A Crumb of Comfort

Robinhood offered US$15 in compensation to its Gold customers following last week’s fiasco.

Robinhood Gold membership costs $5/month and members are charged 5 percent yearly interest when they borrow more than $1,000. The interest is charged daily.

That compensation offer was inadequate, said Ray Wang, principal analyst at Constellation Research.

“Folks missed out on the biggest trades, and their losses were in the hundreds of thousands of dollars to millions,” he told the E-Commerce Times. Still, “it’s better than nothing.”

It will be impractical for Robinhood, which claims 10 million customers, to compensate them all.

The company should compensate only those users who were impacted adversely, Enderle said, but “determining what is enough, given the number of people that will likely try to game the issue and get compensation they aren’t entitled to, is problematic.”

Robinhood “is a thin-margin company, so their ability to fund a large compensation pool without putting the firm in financial distress is very limited,” he pointed out.

Funds should be directed to addressing the problem so it doesn’t recur, Enderle suggested, because “we have a volatile market, making it likely we’ll see more huge spikes.”

Technical Reasons for the Crash

It’s been known for years that sudden spikes in traffic can crash a network. Many corporations use virtual machines, and others have made provisions to tap into AWS or some other cloud service provider to beef up their server capacity if necessary. However, neither of these steps will help if the corporation’s network infrastructure can’t handle the increase.

“Hypervisor software that can clone VMs on demand are ultimately limited by the computing capacity that is available,” pointed out IDC Research Director Mike Jude. Further, each VM comes with redundant application instances, further increasing the load on the network.

Cloud services are another possible solution, but “they depend on data links, which can introduce latency into every transaction,” Jude told the E-Commerce Times. “On a very active trading day, such latency can actually introduce costs.”

Robinhood “needs to do some real systems analysis, identify the choke points, and look at adding capacity to those points,” he suggested. Ultimately Robinhood may need to redesign its software to handle load increases more efficiently.

One option could be to solicit advice from massive multiplayer online role playing game designers, Jude said. “Active trading is just another MMORPG — where you want to see an instant response when you karate chop a villain. Delays cost virtual lives.”

However, it’s inefficient and inexpensive to design for the odd extreme event, he noted. “They can beef up their infrastructure, but at some point it’s simply silly to add capacity that might never be used.”

Robinhood could design a system that taps additional capacity from the cloud and add telecommunications to access those cloud resources, “but it may not be enough when the markets go crazy,” Jude said.

Everybody’s Getting Hit

To be fair, E-Trade and TD Ameritrade also had problems last week.

Forty-four percent of E-Trade customers had problems placing orders, while another 44 percent had problems logging in, and 11 percent reported problems with the company’s website, according to Downdetector.

Excessively high traffic “performed like a DNS attack,” Enderle said. “Online brokerages often have this problem while traditional brokers tend to bottleneck at contact — too many clients calling at once.”

The outages are not all the brokerages’ fault, Jude remarked. “Markets are all electronic and highly complex. Outages happen.”