What is a Facebook App? You’ll get different answers based on who you ask, but surely the most common answer — from most agencies and their clients — will still be something like, “an application that sits inside Facebook.” I’d like to suggest that it’s all about the open web (which these days more or less means “the web outside Facebook.”)
A majority of people in the marketing industry still seem to believe that Facebook development is all about Canvas Apps (or Tab Apps). Even if those of us who work in agencies claim that we don’t really believe this, our business models give us the lie: how much revenue accrues to us from what we might call Tab build?
I’d like to pause for a moment to share a personal mantra:
Facebook marketing has everything to do with building Facebook into your business, and almost nothing to do with building your business inside Facebook.
There are any number of drawbacks to building apps inside Facebook; but I’d just like to highlight two here.
- apps still don’t work on a majority of mobile platforms (and there’s little real sign that they ever will)
- Facebook’s branding dominates theCanvas, and they serve ads around your content. You might choose to believe that this reassures the audience in some way (if you carefully ignore people’s widely-published privacy concerns, that is. Personal mantra 2: people trust their friends, not Facebook.)
Furthermore, you have to drive traffic to Tabs. This isn’t strictly a drawback — but if you’re going to drive traffic anywhere, why wouldn’t you drive it to a space you own (your site, for example), rather than a space you merely lease (a Facebook Canvas with competitor ads down the side)?
The smart money these days would seem to be those who are using Facebook’s Open Graph to its fullest, allowing them to embed much of the key functionality of Facebook (finding friends and sharing stories, for example) into their own open web and mobile apps, and allowing those apps to integrate back into key Facebook distribution points: the News Feed, the Ticker, and (importantly) the Ads platform.
The user benefits from an improved experience (no more lengthy data capture forms to fill in! find my friends! brag about my triumphs and achievements!) and the app developer from increased reach, and peer-endorsed marketing messages — leading ultimately to improved user conversion rates, and other marketing efficiency goodness.
As examples of this, I’d like to cite DrawSomething and Instagram; both powered by Facebook, both sufficiently popular that I suspect that none of you had to click on those links to discover what I’m talking about, but neither has a particularly large following on their Facebook Pages (and I’m sure neither of them cares.)
To explore how developers are migrating their Facebook offerings to the open web in greater depth, I wrote a quick scraper to pull data from the excellent AppData’s list of top Facebook apps, in order that I could draw this chart.
Each column represents a hundred apps, with the most popular hundred apps (ranked by Monthly Active Users — or MAU) at the far left. The pink bar represents the proportion of those apps that aren’t Canvas apps3, and the grey bar represents the proportion that are.
From this we can see that fewer than half of the top hundred apps are Canvas apps (most of these are social games.) The remaining apps all have a life outside Facebook on the open web or as mobile apps.
The non-Canvas apps’ share drops off a bit beyond the first hundred, but they do continue to represent more than a quarter of all apps.
It’s a bit more dramatic if we look — not at a simple count of apps — but rather at share of MAU.
So here’s the thing. Let’s all stop building functionality in the leased environment of Facebook, and instead learn how to build out our web apps as part of Facebook’s Open Graph.
- For more clarity: in my analysis, I split out apps which were classified as appearing solely on the Facebook Canvas platform from those that were multi-platform (which might include Canvas), or single-platform non-Canvas (e.g. mobile clients.) ↩