Facebook Ads API: Optimizing from Good to Great

Dennis Yu is Chief Executive Officer of BlitzLocal, a Webtrends partner that Builds social media dashboards to measure brand engagement and ROI, specializing in the intersection of Facebook and local advertising. You can reach him on Facebook, Twitter, LinkedIn, his blog, or good old-fashioned email at dennis@blitzlocal.com.

Continuing a three-part series on Facebook Ads API. In Part one, I looked at how Facebook Ads vendors have done little with their exclusive access to the API and instead have tried to game Facebook’s system through ad multiplication. In Part two, I discussed how the API can be used to gather the data needed to measure true performance. In this final part, I offer some ideas of tools that developers could create to improve the optimization of Facebook ads.

True optimization of Facebook ads implicitly requires that you track cost per action. If the action is a fan, then optimization is easy—you look at your cost per click and click-to-fan conversion rate. A 50 cents cost per click and 33% conversion rate means a fan costs $1.50. Lower your CPC or raise your conversion rate to get a better CPA. But if your goal is something deeper such as engagement, lead gen, or sales, then you cannot use Facebook data alone.

While certainly you can pass through a tracking code or parameter in your URL, you cannot tell when your ad hit Person A, who then influenced Person B. Except for where you can track all user activity, such as in an app, it’s currently impossible to directly measure the impact of word of mouth. Thus, the most powerful thing about Facebook is not truly measurable yet.

As a developer, you can still optimize against everything else. Why not build a tool where the user can enter their campaign items, specify a target Cost Per Fan and budget, and then hit “go”? What if the user could also establish a goal for engagement rate and newsfeed coverage, which then triggers pre-made Sponsored Post Stories? What if you provided the option to trade-off quality versus quantity, so that the client could see the engagement rate by fan cluster and then make a choice more informed than just going after the cheapest (and lowest quality) fans? Facebook might want to buy a company like that.

Given that the image is the most important part of the ad, why not build a tool that creates several variations of the primary image and tests for a winner based on CTR, CPC, and CPF thresholds? For example, if a brand has 50 locations, you’d have a tool that injects the city name into the ad and multiplies out 50 geo-specific campaigns? You can do it with vendors on AdWords, but not on Facebook yet.

We’ve all used SpyFu, Keyword Discovery, WordTracker, and other keyword research tools. What if you created an interest mining tool that was better at suggesting interests than Facebook’s own tool? We’ve seen some incumbent PPC vendors launch a Facebook version of their keyword tool. These are fundamentally flawed because a search keyword has intent at the time of search, while there is no immediate intent on Facebook--- ads are not triggered by an inquiry.

How about the equivalent of an AdWords Editor? I’d pay $20/month for a nice desktop tool that was faster than the web-based tool such as Facebook’s own or Marin’s (which we white label for).

How about building a tool that accepts the client’s content and then decides if it should be shown via an ad or organically via the wall? With a cursory knowledge of EdgeRank, you could create rules that automatically divert content that has a question mark to the wall, as well as use the ads Platform to amplify anything that has gone “hot” on your wall or anywhere else on the web.

How about a competitive monitoring tool that tracks how Company A is faring against Company B? If Company A was a client, he could set a rule to turn on ads if Company B gets more fans or has a higher engagement rate.

What about a tool that integrates ads and landing pages, such that the promise in the ad is automatically delivered in a tab on the brand’s page? How often have you clicked on an ad that was promising, but then were taken to the equivalent of a homepage? There is massive opportunity to be personalized here, versus the year 2001, when all ads, regardless of what product promoted, went to a company’s homepage.

How about integrating with Salesforce.com, Eloqua, Marketo or systems that collect leads? Most of these B2B companies are only just starting on Facebook, so they’re hungry to buy or partner with someone who can demonstrate a workable solution.

All of the above are tools you could build right now with existing API access. Now if Facebook really wanted to make developers happy, here are a few more things they could open up:

  • The ability to extract the user IDs for all your fans. One vendor has sworn that Facebook has given them this access, but my contacts are unaware of this. John Yi, if you’re reading this, I’m talking about you. By being able to see all your fans, you can figure out the quality of fans by vintage (when they joined and how they joined), overlaps (your true unique fan base if you have multiple pages), and interest analysis (what are the top things that your fans like?)
  • Fresher insights and Ads API data. Waiting 48 hours drives our clients nuts, especially if they see one fan count number on their page and another on their report.
  • Bring back conversion tracking. It was briefly live until last November and then killed for lack of usage and being buggy. Brands want to know what their users are doing in places other than just Facebook.
  • Remove the 20 user account limit. We have Google MCC accounts that have 70 to 80 client accounts, so why not allow the same on Facebook? In fact, why not copy what Google has done with the AdWords API? We know Facebook has only a couple dozen engineers on the Ads team, so we appreciate the prioritization challenge.
  • Including the “quality score” or other ad effectiveness metrics. While the argument we’ve heard is that doing so would open up the system for gaming, we believe the opposite is true. Give advertisers a target metric to strive for and they’ll optimize towards it, just like what is happening on Google now.

Where there are gaps, folks come in to fill them. For example, Facebook doesn’t provide EdgeRank, for the same reason that Google doesn’t provide the true pagerank of a page—so vendors such as EdgeRankChecker come in to fill the void. You could do the same on things like a Facebook Quality Score, using ad CTR and impression count to estimate where you are in the auction.

I see opportunity for dozens of tool vendors to thrive here on a much broader scale than the SEO and paid search vendors in traditional search engine marketing. The opportunity is bigger not because there is potentially more ad revenue at stake, but because this is where users are spending their time, because social advertising is more complex, and because word of mouth marketing requires deep Integration of other marketing channels.

Be sure to read the next Marketing article: 221 Marketing APIs: SendGrid, MailChimp and Eventbrite