On Display: Mobile
Ads, CTR, and the product-service line
tl;dr To have a successful display advertising business in the mobile era, the user has to view your product as an integral service.
Display advertising: the elephant in the room for most proponents of consumer software. These wonderful services, platforms and sharing tools are largely powered by a technological evolution of the billboards that line our highways. The billboard on the 101-N just south of San Fransisco doesn’t know I already have a Moto X in the car with me, whereas the Facebook ad knows my name and who my friends are. The Foursquare ad knows where I’ve been (and, increasingly, where I’m going next). The Twitter ad knows what I’m reading, what I’m saying and who I’m communicating with. The Gmail ad knows… almost everything. Much is made of adtech, segmentation, targeting and retargeting, exchanges, bid rates, varying conversion metrics, pricing models. But, at the end of the day, they are still ads: copy and/or images pimping a product.
Consumer software display advertising has generated the revenue for tech companies to build amazing, useful, and often free tools for consumers. The suite of services Google offers users for no charge in 2013 would have brought many enterprise software executives to tears if they had the premonition a decade ago. Younger companies are very aware of this sucess, and are eager to replicate it: Instagram, Pinterest and Tumblr all recently announced updates to their ad products. I think that’s great. But display advertising is notoriously ineffective, and product developers need to be careful. If we’re not pragmatic and explicit about why, where, and how advertising is applied to our products, it can become a band-aid solution that is bad for products, business models, and users.
Ads have low conversion rates that almost always deprecate over time. Andrew Chen calls this the Law of Shitty Click-Through. A famous eye-tracking heat map shows how effectively we’ve learned to cognitively ignore banners on web properties. I don’t mean they aren’t clicked; I mean most people do not consciously process their semantic content. Your brain has learned to implicitly dismiss them. This phenomenon spreads to any visual form of content that users regularly ignore: not useful at best, annoying or offensive in worse cases. It happened to news paper ads, then telemarketing, then banner ads, and will happen on mobile. This is not good news for marketers, especially now that we can actually measure conversion. The unquantifiable display ad business of last century, thriving through print placement, has already steeply declined.
Facebook will cannibalize their own mobile engagement by cramming News Feed with ads. Right now, mobile is constrained viewports combined with better personalization and contextual awareness, and a low barrier to switching apps or information sources. That equates to lower tolerance for content that doesn’t perfectly match user intent (can you imagine a pop-up ad in iOS?). My employer has chosen not to display any ads in native apps, despite doing so on the web. This isn’t out of some abstract altruism, but simply because the powers that be believe there’s no possible way to do it without significantly degrading user experience.
However, display advertising is alive and well and generates huge amounts of revenue for a large group of companies. Twitter’s IPO is another display ad IPO, Instagram ads are coming, Pinterest just announced Promoted Pins, and Yahoo will continue to ramp up advertising on Tumblr. Very smart and experienced product people still invest in display advertising, and, sometimes, it works. How does that fact reconcile with Chen’s Law and my apocalyptic prediction above?
To hang your revenue hat on display advertising, your product must be essentially ubiquitous to the consumer. It has to be baked into the service layer of their experience. In days when Google Search is on the home screen of 80% of the world’s mobile devices, it’s easy to forget Google was once mostly a web property. Its long-since become household parlance to “Google it,” and that translates well to mobile. Facebook is a mobile service, too: address book and sometime OTT messenger. Personal experience is a terrible excuse for research, but anecdotally I know scores of people that would quit Facebook if not for the repository of contacts. I’d love to see metrics on how many users have large graphs and use chat, but rarely share or actively comment.
*Aside: This could change in the future, as OAuth and cross-account sign-in protocols make graphs more exportable and universal (like OS-level integrations with social services). *
When a new piece of software (especially consumer) is in early stages, we love to ask: is this a product, or a feature? The platitude explores whether there is enough differentiation and utility to carve out cognitive space in a user’s daily life, or whether its just a novel bit of functionality that belongs inside something bigger.
For more mature products considering display advertising models, the question is: is this a service, or just a product?
Instagram has an highly engaged community that is 75% the size of Twitter’s (~150 million vs 200 million). In its early days, it would’ve been easy to dismiss it as a feature: photo uploading, gimmicky features, derivative follower model, minimalist liking/commenting system. But the results show it was an engaging experience, powered by smart technology like compressed uploading and “lying” UI: showing actions as completed in the interface while they’re still underway server-side. When they introduce ads, they’ll test the limits of their userbase’s engagement. Is it so widely adopted, and used, that people will put up with deprecated UX and targeting fails, like we all do on Facebook? Or will we take our flicking and double-tapping talents to Snapchat, Flickr, Vine, VSCOcam or one of the other countless alternatives?
It’s great to have a popular product, large and engaged user base, and growth metrics. But when you’re comparing revenue models, and exploring introducing ads, consider the utility of your product. In rare cases, the benefits of use are so great to the use that you can afford to show them billions of pieces of spam and collect handsomely on the minuscule percentage that converts. Twitter has crossed into this space, and, for the time being, keeps heading up to and to the right. Companies need to become profitable to survive, and startups have huge amounts of pressure to pay-out investors and employees. However, if you are comprising the utility you promised the user when you acquired them, it’s unacceptable. Don’t be the long-play version of promising one thing up front, then pulling a switcheroo once the user has invested time and personal data. It’s better to have “a few who love our product,” instead of a big crowd that just likes it. The proud minority have been able to introduce ads and retain engagement.
In many, many more cases, successful products have been abandoned by formerly engaged users in favor of the newer social flavor-of-the-month. Before you marry display advertising, consider how fickle a mistress an engaged community can be. Be sure that you’ve graduated from product, to service.