This week, Parag Agrawal, who recently succeeded Jack Dorsey to helm Twitter as CEO, published a thread explaining the company’s approach to fighting spam on the platform. That spam, primarily perpetrated by “bots,” or representations of Twitter users controlled by software, had been cited as a point of concern by Elon Musk, who has agreed to buy the company. Musk replied to Agrawal’s thread in a characteristically colorful fashion, questioning how advertisers could “know what they’re getting for their money” if the prevalence of bots cannot be quantified with absolute certainty.
So how do advertisers know what they’re getting for their money? This is fundamental to the financial health of Twitter.
— Elon Musk (@elonmusk) May 16, 2022
It’s important in considering this question to understand the composition of Twitter’s advertising revenue because not every advertiser has the same motivation or goals in purchasing ad impressions. I’ve written extensively (and sometimes, dismissively) about brand advertising as a practice, but ultimately, there are two advertising tactics that can co-exist in a sort of yin-and-yang symbiosis under the broader umbrella of performance marketing:
- Brand advertising, which seeks to use advertising impressions to build awareness for a product that will result in incremental purchases over some enduring, indefinite timeline;
- Direct response advertising, which seeks to use advertising impressions to generate attributable, incremental conversions (often: purchases) within some specific amount of time from exposure to an ad.
As I propose in What is performance marketing?, brand advertising is not antithetical to performance marketing: brand advertising and direct response advertising might both be utilized within a performance marketing framework simultaneously, or individually at different points in a product’s or a company’s lifecycle. The effectiveness of brand advertising is generally more difficult to measure than direct response, and sometimes that opacity of efficiency is used as a crutch to avoid accountability by brand marketing teams. “Building a brand” is an imprecise and nebulous endeavor, and unimaginable resources can be invested into it. Brand advertising can be pursued within a performance model, but very often it isn’t.
At an analyst day in 2020, Twitter revealed that 85% of its advertising revenue is generated from brand advertising campaigns. This, on its own, is not a cause for concern or necessarily an objectively unfavorable revenue structure. But the context of this announcement was that Twitter was wholly committed to changing that structure by improving its advertising platform such that it could manage more money from direct response advertising campaigns. Twitter made the point that it felt that direct response advertising represented an opportunity for the company: the economics of direct response advertising index to the value generated for advertisers, and thus direct response revenue can scale in a way that is untethered to DAU growth. Brand advertising is simply a function of the size of a platform’s user base: with brand advertising campaigns, advertisers pay for reach on a CPM basis.
Elon Musk is correct: bots prevent brand advertisers from quantifying in a reliable way the number of people — real, breathing humans — that see its ads. But Twitter has a bigger, more pressing problem than bots. Twitter isn’t able to manage direct response advertising campaigns at an appreciable scale. Fixing this, and welcoming more direct response advertising spend to the platform, is a far more compelling revenue opportunity for the company.
Because for the most part, direct response advertisers don’t care about bots. Direct response advertisers optimize towards business objectives: purchases, registrations, etc. Bots may muddy a direct response advertiser’s understanding of its CPM metric, but CPM (and reach) is primarily a vanity metric for direct response advertisers. Direct response advertisers are focused on “bottom-of-the-funnel” metrics that aren’t influenced by bots. This is why, while annoying and unacceptable, the consistent drumbeat of revelations that this-or-that big ad platform inflated its reach metrics doesn’t tend to perturb direct response advertising budgets. I wrote an op-ed on this topic in 2019 that was not warmly received.
Last December, when Jack Dorsey stepped down from the CEO role at Twitter, I wrote a piece titled Can Twitter become a direct response advertising behemoth? in which I questioned whether Twitter possessed the constitutional inclination and technical capacity to expand its diminutive direct response advertising business. From the piece:
The most pressing question for Twitter’s new CEO, to my mind, is whether the company can create a direct response advertising behemoth that can rival other large platforms like Facebook and Google…Twitter’s O&O advertising platform is currently something of a chaotic, anachronistic system, and the company sold the ad tech asset, MoPub, that best captured direct response ad spend on mobile. If Twitter sees advertising revenue, but especially mobile advertising revenue, as an integral component of the company’s growth, then it should devote whole teams to advertising attribution and measurement. Nothing else should really matter to the company: growing direct response marketing revenue should be Twitter’s first and most critical aspiration.
Why hasn’t Twitter done this? There may be structural reasons: the Twitter feed is perhaps more frenetic and not as friendly to direct response ads as other product surface areas. But if the format isn’t the obstacle to building an important direct response business, and actually the hindrance is cultural, then addressing that is almost certainly a more lucrative initiative than utterly eradicating bots, which may not be possible and seems to be something of a white whale for Musk.
