Shares of Twitter (NYSE:TWTR) plummeted after the company reported its second-quarter earnings in late July. Not because it missed estimates for revenue or earnings, but because of management's comments on user growth.
Interim CEO Jack Dorsey noted that features designed to onboard new users have failed to ignite significant user growth. CFO Anthony Noto followed up by tamping expectations for future user growth by saying there won't be sustained meaningful growth in MAUs until the product starts to reach mass market.
Meanwhile, Twitter's transition to cost-per-conversion direct-response ads hasn't moved the needle much on ad prices. Cost per engagement increased just 6% in the second quarter. With user growth slowing significantly and ad prices mostly stagnant, Twitter will face pressure to keep up with Wall Street's expectations.
That's why you'll probably see more ads.
One in 20 tweets
Noto finally gave us some more finite details on Twitter's ad business, noting that it's still only one-third of the way to its long-term potential ad load. In November, at the company's analyst day, he mentioned that load would be around 5%, or one in 20 tweets.
Some users in the U.S. report that they already see more ads than that. A recent survey from Cowen and Company found over 50% of respondents already see Twitter ads at least once every 20 tweets. The average ad load according to the survey was just over 5%. Of course, those numbers rely on self-reporting instead of official metrics, so the ad load results ought to be taken with a grain of salt.
Nonetheless, Twitter will have to find a way to increase ad engagements if it's not going to see significant user growth. Twitter's average price per ad engagement was up just 6% in the second quarter after increases of 10% and 30% in the fourth quarter and first quarter, respectively. With analysts expecting 45% revenue growth next year, Twitter will need to find a way to increase ad engagements or price. With the slow improvements in ad prices, it's likely to rely on the former.
Since Twitter changed to a cost-per-conversion model, where advertisers only pay for specific user actions related to ads, Twitter can take further steps to improve conversions. That would enable it to see more ad engagements from the same number of ad impressions, and increase revenue that way.
Twitter is already taking steps to improve its targeting with the acquisition of TellApart in May. Additionally, it partnered with Google to include its ad inventory in DoubleClick Bid Manager, which enables advertisers to better track how well Twitter ads convert. If conversions prove higher than expected through DoubleClick, it could prompt some advertisers to spend more on Twitter ads.
Still, ad relevance on Twitter remains dismally low. That Cowen and Company survey mentioned earlier also found that just 29.1% of respondents found Twitter's ads even somewhat relevant.
Internally, Twitter is working on the problem with Cortex, its AI research team. Cortex is working on deep learning algorithms to "expose relevant content," with its primary focus on Twitter's ad system. Since Twitter's basic ad unit is just a tweet, the same algorithm can be applied to curated timelines like Project Lightning. Ads are nothing more than curated content inserted into users' timelines.
Temporary and real fixes
Increasing the ad load is simply a bandage for Twitter to stick on the real problem that ads aren't relevant. It's a quick and relatively easy solution for Twitter to keep up with analysts' expectations for revenue and earnings. But if Twitter wants to sustain its revenue growth well into the future, it needs to figure out ad targeting and how to convert ad engagements into actual sales for advertisers.
Otherwise, it will find itself largely limited to brand advertising, which still makes up the majority of its ad revenue. At the same time, Twitter is finding its user base surpassed by competition like Instagram and Snapchat, which are attracting significant brand advertising dollars away from Twitter.