Facebook is having trouble meeting expectations.

The much-hyped social network made a tepid debut on the public markets on Friday, its stock ending just 23 cents above the IPO price of $38. Earlier in the week, it was dogged by news from General Motors that the world's fourth-largest advertiser would pull the plug on paid Facebook ads, citing concerns about their effectiveness.

In the end, it wasn't GM's dramatic 11th-hour pronouncement that got the stock off to a rocky start; technical glitches at Nasdaq, high expectations of a big IPO pop and an element of schadenfreude can be blamed for that .

But it does point back to the bigger question for investors and for advertisers: Just what should they expect from Facebook?

In GM's case, the automaker decided to pull its ads from Facebook after finding click-through performance trailed that of other online options, such as Google's AdSense, according to Reuters. But low click-through

rates will hardly come as a surprise to anyone who spends money on Facebook.

Facebook can be used as a direct-response tool, and sometimes quite effectively. Domino's, for example, opted to run a one-day promotion selling half-price pizzas in December exclusively on Facebook and reported that it drove 542,000 people to its online ordering site.

But the broader direct-response picture is mixed. Kantar Media Compete measured the effectiveness of Facebook content and ads in automotive, travel, tech and retail as defined by whether people exposed to them click on an ad, visit a brand page or seek out more content about the brand. In all categories, Facebook distinctly trailed most portals and search engines.

Perhaps that 's why Facebook is positioning itself as a branding tool—better for top-of -the-funnel consideration than bottom-of -the-funnel conversion—and urging marketers to focus on metrics like reach and "resonance" over click-through rates. Resonance, in Facebook speak, is a combination of recall, awareness and purchase consideration. And it might sound less like the web and more like a specific type of old-school media.

"TV kind of operates in the same way, and we're trying to demonstrate that the same measures that are moving through TV are moving through online in general, but specifically on Facebook," said Brad Smallwood, Facebook's head of measurement and insights.

Clicks, while easy to measure, can be a terrible proxy for success, said Ian Schafer, CEO of Deep Focus. "Brands need to spend money researching what kinds of behaviors make people better customers. ... It's going to be different for everybody."

Mr. Smallwood noted that Facebook is focused on partnerships with third-party-measurement providers, notably Nielsen, with whom it launched an online campaign ratings tool. The tool allows brands to measure the reach of digital campaigns in TV GRPs.

Mr. Smallwood's team has worked with major advertisers to gauge the "reaction" they're seeking, whether a concrete sales result or something more intangible like sentiment.

Movie studios want to know if Facebook leads to more ticket sales, so Facebook has used a predictive tool to estimate the opening weekend box-office results. It polls users on whether they intend to see the film, and then compares results between people who saw the ads and those who didn't.

Automotive, however, is more difficult because of the long length of the buying cycle. Marketers look more closely at "upper-funnel measures" and whether advertising had any impact on factors such as purchase consideration and likelihood to recommend.

Measurement is far easier for some categories. An online social-gaming company can measure its Facebook-ad success in a relatively straightforward way, said Simon Mansell, CEO of Facebook-ad company TBG Digital. But a brand like Heineken, which isn't looking to sell beer on Facebook, might measure whether it's lifting its "net promoter score," or the likelihood of people to recommend Heineken offline.

To gauge brand sentiment and the effectiveness of a campaign, a company such as TBG can deliver a survey within Facebook ads to both fans and non-fans. It then compares the two data sets over time. It can also enlist the services of a social-listening platform like Radian6 to do a sentiment analysis.

But ultimately, it's still more of an art form than a science.

Facebook could do certain things to improve a brand's ability to understand if the platform is working, Mr. Mansell said, such as making available certain types of anonymized user data. For instance, if a telecom company could match the email addresses of Facebook users who had liked its page against its own customer base, it could determine the churn rate (customers switching carriers) for its Facebook fans and whether that declining over time.

The reality is that a Facebook presence is a must-have for brands but Facebook advertising is not. GM, for example, spent $30 million on building and maintaining a Facebook presence for its various car makes and models and only $10 million on Facebook ads.

Most brands with big Facebook presences are not like GM and are spending on advertising, but they'll need to spend a lot more if Facebook is going to deliver on the ad-fueled growth its investors are expecting.

Facebook as a platform is still nascent and it may take years to understand the impact of messaging there.

"Maybe the real magic of Facebook is it's a long-term prospect, not a short term one," said Rob Norman, CEO of GroupM North America. It's "the idea of maintaining a relatively light-touch relationship with people and not obsessing over where they become a customer."

In other words, just as investors learned on opening day, marketers would be wise to look at Facebook not as a get-rich-quick play but as one that may pay off in the long run.