It’s been a turbulent few months from Yelp. Rumors of a potential acquisition have created volatility for shares of the reviews site, compounding investor concerns after a first quarter earnings report that showed lower than expected profits and sluggish user growth sparked a sell-off in April — the largest sell in its three years as a public company.
In a note to clients a few days after the April sell-off, Bank of America argued that Yelp’s “old growth story” was over. From a balance sheet perspective that may be true, but the company’s growth story is far more complicated than much of the discussion in the financial community allows. The shift in traffic from desktop web to mobile app usage has placed a fundamental pressure on its existing model while also creating a potentially compelling growth story once that transition is complete.
Most importantly, the apparent slowdown in Yelp’s growth does not necessarily reflect the value of the company to consumers or advertisers. Instead, it exposes the growing pains of a mobile ecosystem, which lacks the necessary infrastructure to accommodate the weight of more mature companies.
A user on mobile is not worth the same as a user on desktop
Two key metrics — total unique visitor growth and local advertising accounts – are driving most of the concern for investors. The company added 6200 new local advertising accounts in the past quarter, representing the second consecutive quarter in which account growth decelerated. The slowing growth in accounts, paired with relatively flat spending per account, is an issue, but one that could be chalked up to a poorly executed sales force reorganization as well as a shift in its sales model from bundled packages to cost-per-click models.
However, the more existential concern about Yelp centers on its user growth. The growth of total unique visitors, a metric that accounts for both mobile and desktop visits, has declined on a year-over-year basis for the past seven quarters, down to a little over 7% from 40% two years ago. Mobile application downloads are accelerating, growing a little over 10% from the previous quarter — the fastest growth rate in over a year — but the growth is not making up for losses in desktop traffic.
But the story is far from linear. The decline in total reach overshadows the vast differences in the lifetime value of mobile and desktop users, and underscores the challenging reality facing publishers when mobile is the core business — not a secondary metric to discuss on earnings calls. The absence of an organized system through which properties can move traffic from one application to another on mobile — the equivalent of the links and search engine model on the web — is deeply limiting the company’s ability to grow its user base.
Fragmentation is not endemic to mobile — it’s a sign of its immaturity
It’s difficult to determine the actual value of a mobile user (versus one on desktop), but it’s reasonable to use searches as a meaningful measure of engagement. Mobile represented a little over 65% of searches for the company, while mobile traffic, measured in both mobile web and app, is still slightly lower than its desktop counterpart. That means that each mobile visitor accounts for roughly 30% more searches than a desktop visitor.
What accounts for the disparity? On the one hand, it’s logical to assume that the frequency with which a user engages with a property does increase when a user downloads a mobile app. If a user has the Yelp app on the phone, it’s logical that they would search more frequently than if they were on the web where the could use multiple services.
But the higher search-per-user ratio isn’t just about search frequency increasing, it’s also about the denominator — overall traffic — decreasing. This decline in user acquisition does not necessarily reflect a decline in demand for Yelp’s product, but rather demonstrates the absence of a meaningful channel through which the company can reach consumers who are not, directly, look for its site on mobile. The inability to easily link between apps, and move traffic from one property to another, has left the millions of applications available on iOS and Android effectively cut off from one another creating a much different calculus for building a local search business today than in the days when desktop dominated.
To complete a given task on the web, say “find a restaurant,” we may visit three or four properties before making a decision: from Google to Yelp to an owned website. Conceptually, the data layer of the stack and the algorithm layer were kept separate, allowing both the data and the algorithms to improve at a much faster clip.
Yelp has thrived because it had two channels: its direct search business where a user would search Yelp content on a Yelp property and a larger indirect business where the search functioned would be farmed out to a search engine such as Google. That indirect audience, where Yelp served solely as a content company, allowed Yelp to build an audience for its content at an exceptional rate and with relatively fractional customer acquisition costs. The company’s own search product was important for core users, but traditionally, its relationship with Google was economically as, or more, important than its direct business.
We’re seeing the collision of search and data across the entire industry. In the local services category, a handful of new players have popped up pairing a unique data set with a search brand. Porch.com, for instance, aggregated project data from service providers across the country to create its own services search product.
The critical question for investors in Yelp (as well as anyone with a stake in mobile) is whether the consolidation of search and data reflects something unique to mobile, or is a momentary effect on an immature ecosystem. What’s increasingly clear is that the high cost of user acquisition on mobile (caused in large part by the absence of a system for managing inter-app traffic) reflects the immaturity of the mobile ecosystem — not something unique to the device or the platforms supporting it.
Here’s my argument: There’s nothing endemic to mobile that would promote an environment in which search and content are not distributed the way they are. In fact, the growth of personalized search — the idea that the searchers past behavior should influence the results they receive for a given keyword — makes it even more reasonable to have a single search property distribute traffic to millions of separate content sites.
Deep linking initiatives will inevitably solve the fragmentation causing the problem. Meanwhile, companies like Yelp will return to a place where their third-party content business — where users search for their content using another company’s software — will drive a majority of its traffic.