Last month, Benedict Evans penned a blog post that talked about the limitations and consequences of the mental models that that we all use to frame innovation. We search for degrees of sameness to determine how much alike the new thing is to the old. For example, it was easy to understand cars as a new-and-improved horse and buggy. This mental model was so compelling that it spawned the most ludicrous metric of the modern era: “horsepower.”

One of the reasons why advertising on Facebook and other social platforms has lagged what results indicate should be a lot more spending is that mental models have induced marketers to perceive it as something other than what it is.

Many of us first tend to see Facebook through the lens of its consumer experience: connecting with friends and sharing experiences, photos and videos. The early days of spending on Facebook sought to increase the things consumers liked about their favorite brands on Facebook. Companies thus formed teams to build and maintain communities of fans and get more likes.

Marketers could advertise on Facebook, too. This made it familiar to digital marketers in charge of placing banners online. Positive results were elusive, though. Remember when GM slashed its Facebook marketing budget? One of its executives said, “I think we’ve learned through our experience that the paid advertising is not the best way of activating Facebook.”

This was just three years ago.

Then, Facebook changed by making it harder for a brand’s organic posts to appear in a user’s news feed. To make sure those posts got noticed, brands would have to pay to promote them.

That spawned Facebook’s auction system and private ads API integration. The mechanics of how marketers buy and launch ads looks similar to what Google offers. That led search marketers to begin comparing Facebook to Google.

About the same time GM punted on Facebook, the company published its first mobile app, and soon after that, ads start appearing in the news feed for the first time. These ads occupy just about all of a phone’s screen. They look just like a friend’s post, which is why some started calling it “native advertising.” If a user doesn’t like an ad, she can just skip it, but Facebook makes that less and less likely by offering up an accumulation of all the data it collects on users so that marketers can target just the right customer.

As a result, mobile marketers begin recognizing that Facebook is an important option for them, too.

Today, I bet that if you ask a typical CMO what Facebook means to her company, you’ll get a hodgepodge of an answer. Is it a community? An alternative to paid search or desktop banners? A native play? A mobile marketing platform?

At Ampush, my employer, we have a clear view on what makes Facebook so valuable to marketers, and it reflects some of these perceptions.

• Facebook is mobile: by far, most of its usage happens on smartphones.

• Facebook does display, but better than open marketplaces: thanks to superior targeting and ad units that are perfectly suited to smartphone screens.

• Facebook performs at all levels of the consumer funnel: like TV, it can help introduce customers to products they would otherwise never know to search for. Like search, it can sniff out intent based on a user’s likes, behaviors and interests.

Facebook changes frequently, so this mental model may need changing, too. Still, we think it’s better than any of its alternatives. Once more CMOs start to see it this way, look for spending on the platform to increase substantially.