Twitter’s ad crisis: poorly targeted ads, fewer advertisers
Twitter shares have dropped by 20% and its ad targeting issues may be largely to blame. According to The Verge, Twitter admits it has been showing more ads to users recently – perhaps to make up for current revenues falling short of Q3’s expectations.
Twitter uses a second price auction to determine which ads are shown to its users. Advertisers place bids to compete in auctions that fit their target audience, and the winner with the highest bid pays the price of the second highest bid.
What’s unique about Twitter is they have slightly modified this so that the winning bidder has to pay “incrementally more” to beat the second highest bid. ie. if you bid $2 and the second-highest bid is $1, you’d just pay $1.01 because +$0.01 is enough to beat the $1 bid.
As we learned in lecture, second price auctions are an optimal model to follow because advertisers are encouraged to bid truthfully and it has proven to be successful for eBay and other platforms. This has proven to increase social welfare because the winning bidder gains a positive payoff by paying a lower price for something he values higher ie. he saves +$0.99 when he would have been willing to pay that much more for the advertising slot.
It is also interesting because Twitter offers “real-time suggestions on what other advertisers are bidding”, which transforms the sealed-bid auction into an open one. Yet as we discussed in class, it is still optimal for advertisers to bid their true value even if they’re aware of what competitors are bidding.
Another fascinating fact about Twitter is it not only factor the price an auctioner bids, but also a “quality score” of the ad based on resonance, relevance, and recency. These matter to advertisers because they often have target audiences they want their product to reach, so strengthening these characteristics will increase the likelihood of converting a Twitter user to a potential consumer of their product/service.
In addition, Twitter accounts for engagement and impressions of an ad through how many people like, retweet, or interact with the tweet, which connects to networks because it relies on the interconnectedness of multiple factors to determine a score or weight of an ad.