Big Data Crashes Twitter Earnings

twitter-fail-whale- Big data
Unmanaged Data may Result in Loss

The unintended consequences of big data, real time data and loose data security all showed up for the Twitter Q1 earnings release this week and showed how not having good control of data can cause some bad things to happen.

What happens when information is released because it is accidentally made public? In this case the damage might seem small, only 18% shaved off the share price of Twitter, but this amounts to about a $5 billion dollar drop in valuation and millions changing hands for some investors. Given earnings were a miss maybe the drop in stock value would have happened anyway but the surprise element may have increased the impact of the news.

This is a great example to anyone wondering about data security and what it means to properly manage public or private data. This episode will pass but it will leave a blemish on Twitter’s reputation and should lead them to take a closer look at how they are managing data that is accessible publicly or meant to be secured.

In this case the data leak was caused when a company, Selerity, who provides real time content analytics that specializes in financial market data and sentiment, picked up the Twitter earnings release in a PDF posted on their public investor relations website with one of their web crawlers about an hour before Twitter pushed out the earnings release. They simply reported the information that was on a hidden URL but public information since it was not secured. (A great lesson for many non-technical marketing and PR people)

My favorite quote from Selerity comes from The Verge. “Any time a company’s earnings are due for release, we check the website periodically to see if the earnings are available. In this instance, I am assuming that Twitter mistakenly posted the earnings to the website early. But they did make the earnings available on the website.”

Earnings have been accidentally released many times before and this will not be the last. The good news is there are any number of simple ways stop or control data so it is not accidentally or purposely made publicly accessible by both using data publishing best practices as well as data management and security products. This is a good reminder to companies to review their data publishing, management and security practices and policies. Do nothing and your company could be the next one being talked about.