How to fix Twitter

If Twitter is broken, it shouldn’t be that hard to fix. In an IPO-anniversary piece, the Wall Street Journal painted a picture of a company in disarray, with regular strategic re-sets, executive departures, sluggish user growth, and a confused CEO. The story rings true, but most of Twitter’s ailments are self-inflicted, and curable by a dose of focus.

Twitter set itself up for the perception of failure by creating and then changing meaningless or confusing metrics, inviting faulty comparisons with Facebook. If CEO Dick Costolo can’t get the story straight, relatively new CFO Anthony Noto – ex Goldman Sachs – should be able to spin the Street. But not with a message that says Twitter’s mission is building the world’s biggest audience.

Twitter needs to embrace the fact that it is a broadcast medium with an audience that’s reached the critical mass necessary for a healthy advertising business. Sure, audience growth is nice, but Twitter’s reach already supported revenue that more than doubled last quarter, to over $360 million. It’s well on the way to becoming a $1.5 billion company, and 85 percent of its ads dollars are mobile already. Just stop talking about “monthly active users,” and “timeline views per MAU,” and “ad dollars per 1,000 views,” a concept that only vaguely resembles CPM or CPC (another metric Twitter doesn’t differentiate clearly).

Define the audience properly

Twitter’s $175 million net loss was driven by $170 million in stock-based compensation, so Twitter really needs to get a bang out of its engineering and sales buck. After focusing for most of its life on its “MAUs,” Twitter now talks about them being a 284 million core – that’s the group that only grew 20 percent. That core has a surrounding circle of drive-by or non-logged-in viewers Twitter thinks is one to two times that size, and an undetermined number of additional viewers who see Twitter content via syndication partners.

Besides the fact that the first thing a media company needs to do is count its audience, I’d also define Twitter’s audience segments differently, and support them with products and revenue streams accordingly. Twitter users look like this:

  • Communicators. A rather small subset of active users actually uses Twitter conversationally. They’re the digerati, journalists, and activists who are complaining about potential algorithmic content promotion. They create much of Twitter’s content, and they’re probably using TweetDeck. Let them manage their feeds and opt out of unchosen content sources. Except for ads, of course. But the ads they see should be targeted to influencers.
  • Broadcasters. The entertainers and media types that don’t converse much are broadcasters, and create most of the rest of Twitter’s content. High-volume broadcasters should pay for that privilege, and many do.
  • Viewers. These are the masses that use the website, click through from elsewhere, and see content on partner sites or apps. Go ahead and tune their feeds, but first offer them “channels” of pre-packaged people to follow, much like Sulia originally did, and somewhat like what Hash is doing now.

Twitter is well equipped to serve up content and ads to the masses, and ads to the digerati, targeted by expressed interest and consumption context. Social media marketing is hot and mobile advertising is about to explode. What Twitter needs to spend R&D and sales efforts on is inventing appropriate content marketing and ad formats, and refining that targeting. And properly count the masses.

Double down on data

The interest graph generated by each of those types of user will ultimately be extremely valuable. I’d argue that Twitter following, viewing, searching, and interacting is more predictive of real interest and purchase behavior than are Facebook Likes, if not quite as ad-friendly as Google searches. Yet data licensing and Twitter’s nascent mobile ad network generated only $40 million this quarter, 11 percent of revenue.

Twitter bought data reseller Gnip in May for about $130 million and announced a relationship with IBM for analytics expertise and a pipeline into CRM and other enterprise functions. More like this, please. I’m sure Salesforce, Oracle, and Microsoft can afford higher licensing fees, if traditional media companies cannot. And Twitter should focus a consultative sales team on Facebook’s top agency partners.

Jettison distractions

Twitter is once again courting jilted developers with its Fabric platform, that includes Digits, an innovative, telephone-number-based log-in scheme.

Well, not everyone should be a platform. Twitter, as a media company, is more like Spotify than like Facebook, and Spotify recently abandoned its platform aspirations.

While Digits could be cool, and would help Twitter support an ad network outside of its own properties, Twitter should probably leave identity management to serious players. Facebook, Google, Linkedin, and Microsoft all correctly demand an authenticated, real identity at the core of their services. Someone should offer the ability to easily anonymize or “persona-ize” that identity depending on context. But that’s likely beyond Twitter, which has been at the identity game for years, with nothing to show for it.