I recently explained why I still won't touch Twitter (TWTR) -- even at a near-40% discount to its IPO price of $26. Its monthly active user (MAU) growth remains sluggish, its ad revenues are declining, its new initiatives are floundering, and its bottom line remains deep in the red.

But there's still one bright spot in Twitter's bleak quarterly reports: its "Data Licensing & Other" business, which posted double-digit annual revenue growth in every quarter since its IPO in late 2013. Let's see what this business does, and what Twitter might do with it in the future.

Two Twitter bluebirds in a speech bubble.

Source: Getty.

What does the Data Licensing business do?

Twitter's Data Licensing business sells a "firehose" of over 500 million daily tweets to various companies. These customers include analytics firms, which analyze the data to identify consumer trends; news agencies, which filter the data for breaking news stories; and even high-frequency trading firms, which use trending tickers and stories to place trades.

Simply put, Twitter sells its river of information to companies that try to turn its tweets into actionable data. Back in 2010, Twitter's Data Licensing business brought in 75% of the company's revenue, with the rest coming from ads. But in 2016, just 11% of its revenue came from the Data Licensing business.

How fast is the Data Licensing business growing?

The slowdown in Twitter's core advertising business makes the Data Licensing & Other unit's growth look impressive. But as seen in the following chart, its sales growth has also been waning over the past five quarters.

Data Licensing & Other:

Q1 2016

Q2 2016

Q3 2016

Q4 2016

Q1 2017

YOY revenue growth






% of Twitter's revenue






Source: Twitter quarterly reports.

To expand that business, Twitter forged several major analytics partnerships over the years. Its most promising deal was with IBM (IBM -1.95%), which integrated Twitter feeds into its cloud-based analytics platform for enterprise customers in 2014. Big Blue also trained 10,000 IT consultants to be experts in tweet-based analytics.

Will Twitter ever divest it?

Expanding the Data Licensing business seems logical for Twitter, but the unit also represents a double-edged sword for the social network. For example, the use of Twitter's firehose feeds by high-frequency trading firms possibly contributed to "flash crashes" in the stock market by plugging "raw" (unfiltered) firehose feeds into their trading algorithms.

A robot reads a book.

Twitter has an ongoing bot problem. Source: Getty.

Twitter's desire to grow its Data Licensing revenue could also encourage it to overlook its bot, spam, and fake accounts -- which all keep its stream of tweets flowing with automated retweets. Of course, Twitter could merely be complacent or turning a blind eye to these issues to report better MAU growth.

A recent study by the University of Southern California and Indiana University found that 9% to 15% of all Twitter's MAUs were likely automated accounts -- which could strengthen its firehose, but discourage cautious companies from buying ads.

Therefore, it might make sense for Twitter to sell the business. A potential suitor would be Alphabet's (GOOG -1.00%) (GOOGL -1.05%) Google, which already acquired some of its developer tools earlier this year. Google already links directly to tweets in search results, thanks to a lopsided deal in 2015, so licensing those tweets alongside its own search data seems like a natural next step. For Twitter, parting with the data mining and licensing aspect of its business might enable it to focus more on expanding its core advertising business, which posted negative year-over-year sales growth over the past two quarters.

The key takeaway

It could be a tough call for  Twitter to sell its Data Licensing business, since it's a high-margin business that generates better growth than its core advertising business. But doing so might streamline its core business and enable it to post better top and bottom line growth.