Taking a step back, the obvious problem is that Facebook and mobile are ill-suited for each other. The source of this divergence arises from the pattern differences of desktop vs. mobile social behavior (use case and access frequency) and incompatibilities between Facebook's advertising-based business model (awareness) versus mobile which centers around intimate local activity (discovery).
Until the recent acquisition of Instagram, Facebook's mobile strategy was to recreate the desktop offering, merely in a miniaturized version. However mobile centers on an intimate social subset or "core", not the "edges" of the social graph where Facebook excels by bringing everyone, including distant friends, together. Edge social activity only serves to convolute what needs to be a clean mobile experience. Welcome Snapchat, a clean 2-step (at most) experience totally focused on intimate sharing.
Despite the fact that Snapchat isn't generating revenue (yet), Facebook itself will soon need to reinvent its business model, to match how and where mobile is used, which happens to be all about real-time and hence local advertising. Banner ads and certainly "likes" don't translate usefully to either teens or local businesses, who would both happily take a transaction-driving offer/ deal rather than some national ad campaign.
So what's the end goal? Meshing local search with social context is the holy grail here. But Facebook simply doesn't have the requisite infrastructure on mobile. No maps, no local, and massive amounts of social noise that block effective algorithmic processing. Google is the strongest mobile player today, and bear in mind that Google created and incubated entirely new product suites over several years (Maps, Android, and more) to be able to deliver their mobile content effectively. (Note: where Google does fall short is in social relevance.)
Social Model Transfer to Mobile?
We know that all great companies must solve a basic common problem. While Facebook provided value by re-connecting users, the question has now become, "can Facebook even play a role in our mobile lives?"
Facebook-friends exist in a distinct category as virtual friends on the "edge" of our social networks (they're associated by 0 to 1 degrees but don't necessarily translate into our real world). Thus once having caught-up with them, many of these "friendships" quickly regress, much as in real life. As a result, the all-inclusive social graph approach generates low-frequency and weak-form communications.
Conversely, mobile revolves around an intimate set of close (or "core") connections that produce high-frequency and strong-form interactions. An analysis of Pew Research data shows the average person in reality has a small number of people in their core circle (6.5 friends, 3.1 work colleagues, 6.4 family members, plus a handful of others). These people represent 66% to 97% of our average daily engagement activity on mobile (via email, calls, texts, and social apps) yet are poorly facilitated by Facebook, and our interactions with them simply don't occur through that medium.
The initial attraction of having one's whole social graph online now does little other than stretch the limits of an acceptable signal-noise ratio. Facebook is not dumb to this (share of user time is quickly dwindling), and recognizes the power that Snapchat can have going forward, especially for potential revenue generation.
"Edge" versus "Core" Monetization Opportunities
The edge and the core do not have equal value in real-life, but what's so important about this? Simple, the value of data implied. Core networks are naturally personalized for us, and are our most powerful reference for thinking, doing and buying, which in turn means the highest likelihood of being monetized (a restaurant recommendation from one good friend beats even 100 reviews from mere acquaintances.) Because suggestions and referrals from the core network are far more accurate and relevant, monetization opportunities are maximized with inclusive data sets. This gives Facebook the angle it's always needed to attack local discovery and hopefully find a new source of growth revenue.
While edge graphs tend to be dormant, core inclusive graphs are dynamic. Because they're dynamic, they're much more action-oriented (i.e., transactional) and suited for mobile. Virtual communities are only monetize-able once they pierce the digital veil to the real world. While Facebook may be learning how to parse its massive data and algorithmically identify core activity, it can never supplant inclusive data and direct interactions among users themselves.
Screen real estate and content richness are non-issues in a desktop world, yet on mobile, this is proving a fundamental stopping point. On mobile, more yields less, and this is why Snapchat will become a huge phenomenon worldwide. While it's well known that smartphones are rapidly becoming prevalent (representing about 32% of handsets sold in 2011, with 60% annual growth), what is less talked about is that cheap "almost smart phones" represent 30-40% of all handsets sold (and are projected to comprise as much as 50% of existing phones globally). Almost-smartphones provide app and web access, but they aren't as fast, and don't like loading and processing rich content. Facebook has always risked losing vast populations of growth to competitors simply having faster and cleaner apps. (In app development, it's a well know tenet that every extra step of key press results in about 50% user drop-off, so this matters).
Local Discovery Requires Going Beyond "Likes"
Facebook's business legacy was building a platform for national brands to advertise and build awareness. However, of all mobile searches, searches for local results comprise 33%. Ads and 'Likes' served Facebook's purposes adequately, but don't transfer to local discovery. It's become painfully obvious that "likes" and "social ads" don't capture the power they were intended to. It's proven a vague and undefined value proposition for advertisers, as the average Facebook user will "like" as many as 200 things this year.
The distinct differences between the business models above, means that the social network is currently ill-equipped to handle monetization on mobile, and needs an entirely new product suite which may or may not need to be separated from its national advertising business. Translated in another way: "Is the target user acting on the referral of a core friend or some low-value edge contact? Product conversation and discovery (alone or in a group) need to occur naturally at the point-of-time users are interacting.
We're at a tipping point in business adoption of social media, and the market opportunity is huge: 79% of smartphone users use mobile to help with shopping, much of which gets bought locally. And growth? Mobile advertising spend in 2011 was $3 billion, rocketing to $20 billion in 2015. Yet Facebook has already lost ground in this battle. With its short "handles" and hash tags (i.e., @twitter and #twitter) that businesses can claim outright (versus needing to be found and friended on Facebook), Twitter (TWTR) has struck a take-over growth trajectory in the last 12 months. The opening segments of TV shows now feature Twitter handles or hash tag names. This is a major concern for Facebook, as Twitter's business model isn't yet formed but has shown major compatibility with mobile, and fills the lite, dynamic need in content delivery.
When you think mobile, you just don't think of Facebook. Will Facebook begin to lose market dominance, mind-share, and user attention as the next generation of companies take up the mobile mantle? One thing made clear from the past decade is the rapid and vicious evolution of social networking and online user behavior (think of the fast downfall of AOL which went from $226 billion in 2001 to about $20 billion in 2006). FB is burdened with legacy infrastructure that doesn't make sense for a mobile world. Declining engagement trends within the existing Facebook model currently speak for themselves. Although Facebook is a mammoth that can't turn or reinvent itself on a dime, it does have $100 billion of market cap to draw upon… and it has little choice but to invest and acquire aggressively and hope it bets well enough to stay relevant. Snapchat is worth $4 billion (and more), and Facebook should not hesitate.