Facebook Continues to Experiment in Local — But Its Focus Is Long-Term

Street Fight Magazine

Facebook, the putative sleeping giant of local search, has made a few attention-grabbing forays into the local arena over the last four years or so, including the launches of Graph Search and Nearby Places. I’ve been writing myself about the local potential of Facebook since 2012. Each time Facebook makes a new announcement, we prognosticators opine that the social network is about to finally capitalize on its huge user base and its goldmine of affinities linking engaged users to local businesses vying for consumer attention online.

Often the talk turns the monetization. The argument here is that Facebook has been chasing advertising revenue since its May 2012 IPO, and that local search opens up a potentially massive market of small business advertisers. Thus its local product launches are supposedly designed to win users over to local search as a core Facebook activity.

The trouble is, this hasn’t happened, at least not with products like Graph Search and Nearby Places. Graph Search, launched to much fanfare in January 2013, has now been folded rather unassumingly into the basic Facebook search platform. The big deal with Graph Search was that it was supposed to support structured queries that mirror what we do in real life when we ask friends whose opinions we trust to recommend a local plumber or a restaurant in an unfamiliar town. One can only assume that Facebook didn’t find much of a user base for the idea. Support for structured searches still exists, but you don’t hear much about it these days. And notably, local business results on the Places tab do not surface any sidebar advertising.

Nearby Places, the local search feature on the Facebook mobile app, seems hidden in a corner as well; on iOS it’s the tenth and final item on the feature list under the More link. Findings like that of the Neustar/Localeze/15 Miles local search study from a couple of years back suggest that people use Facebook for mobile local searches to a surprising degree, but I’ve often wondered exactly what kinds of local searches these are; in my own experience, it’s so much more convenient to use Google or Apple Maps if you need the nearest Starbucks that I can’t imagine anyone turning to Facebook for that reason. I suspect that Facebook’s high mobile local numbers come more from name-in-mind searches where users are looking to connect to specific local businesses, for instance in order to check in during a store visit.

Check-ins are obviously significant from a local engagement perspective, but they don’t represent anything close to the full potential of Facebook as a local portal. I’m certainly curious to see if that potential might be realized by the launch of the company’s new Services portal last month, though I’m also skeptical. One issue with launches like Graph Search, Nearby, and now Services is that Facebook seems content to let them vie for themselves out in the boondocks rather than granting them the inevitable usage lift of dedicated space in the Newsfeed. Consider Google Places as a counterexample. Google has carefully but progressively ceded a big portion of the search page to its local property, in what might be viewed as a cleverly calibrated move to grow the user base for local and thus win an increasing share of ad dollars for locally targeted keywords. By contrast, as far as I can tell there’s no way to even find the new Facebook Services page from the Newsfeed or your user profile, at least not yet. I got to it the way a lot of us probably did, by reading about the feature in a blog post and then typing facebook.com/services into the browser. Not exactly an invitation to the masses. Also important to note: if Facebook’s goal in local is monetization through advertising, why is Services not only orphaned from the Newsfeed but, from what I can tell, entirely ad free?

Some may note that I haven’t mentioned one other local experiment on Facebook’s part, the Places Directory launched in November 2014. Structured in the form of a city guide that helps you discover great places in your area, the Places Directory still exists but, like Services, you can’t get to it easily from the Newsfeed. In fact, if you type a local query like “hotels birmingham al” in standard Facebook search, then click the Places tab, you’ll get local results, but in a different, more Newsfeed-like interface than the same search atfacebook.com/places. The listings themselves are different as well, as though Facebook is employing totally unrelated ranking algorithms. Do the same search atfacebook.com/services and you’ll see a third interface with listings ranked according to yet another method. It’s fair to say that the search experience across Facebook’s local portals is intentionally fragmented.

Or rather, that the company isn’t trying to build a unified local search portal at all. Instead, they’re launching trial balloons, turning developers loose to experiment with the company’s local data in order to surface features and nuances that might eventually be useful as subtle additions to the Newsfeed. That’s what they’ve done with Graph Search, so if the same thing happens with Services and the Places Directory, it shouldn’t come as a surprise. Shortly after it launched Newsfeed in 2006, Facebook made the feature into the centerpiece of its user experience, adding features but jealously guarding the Newsfeed’s primary role as destination and site of user engagement. The orphaned status of its local portals testifies to that unwavering focus. All of which leads to the conclusion that Facebook recognizes the value of local but sees that value in a broader context and in terms of a long game rather than a land grab.

The Six Quality Score & Ad Rank Questions Everyone Has For Google



1. How Do I Increase My 1–10 Quality Score?

I should start by pointing out that Quality Score (1–10) is a lot like age: just a number. This question should really be phrased, “How do I make better ads for my users?” Quality Score, that 1–10 number you can see on your Keywords tab, isn’t the best way to measure how well your ads are performing. In fact, it isn’t even used when calculating your Ad Rank. (Real-time, auction-specific quality calculations are used instead.)

What is helpful, though, is the system’s ratings for each of the components of your Quality Score (1–10). Are you below average, average or above average when it comes to your three quality components? Based on those ratings, you’ll have a sense of where AdWords thinks that keyword/ad combination can be better. A full list of specific steps to take for each of those three ratings can be found in this article in the AdWords Help Center, but here are some of the highlights:

Component to
Actions to consider taking
Ad relevance
  • Speak more directly to the intent embedded in a user’s query.
  • Move keywords to smaller ad groups with more targeted creatives.
Expected CTR
  • Write more compelling ads (by highlighting unique benefits, testing out new calls to action and being specific in your text).
Landing page experience
  • Send traffic to landing pages related to someone’s query.
  • Improve conversion rates. Conversion rate is not used to calculate quality, but it’s a good proxy for landing page experience.


2. How Does Ad Rank Affect My Actual CPC?

The max CPC that you’ve set for your keywords enters the Auction and contributes to the actual CPC that you end up paying. You’ll ultimately end up paying what’s minimally required to hold your ad position and any ad formats shown with your ad. It’s explained really well in this video by my building-mate, Hal Varian. The better your other components of Ad Rank (quality/extensions), the lower the CPC you’ll typically end up paying to achieve different positions in the Auction.

3. Why Is Average CPC Not Indicative Of What I’m Paying?

Because averages suck. I know “suck” isn’t really business professional phrasing, so I’ll instead say, “This metric can be useful, but it’s easy to misinterpret.” I stole that language from an old Inside AdWords post about average position, also by my building-mate, Hal Varian. It’s a great post that’s well worth a read, as the lessons still apply today. (And also, man, has the results page changed a lot in four-and-a-half years. The screenshots in that post are showing their age.)

Your max CPC enters each auction, and depending on a host of factors, you can pay any amount up to that max CPC. Your actual CPC can even be over that max CPC if you have bid adjustments in place, which you probably do. I’ve included this table in a couple of other blog posts, so let’s go for a trifecta, as it applies here, as well:

Max CPC Total Clicks Clicks with an actual CPC of $0.01 Clicks with an actual CPC of $0.11 Avg CPC
$0.11 10 9 1 $0.02
$0.09 9 9 0 $0.01


One actual CPC near your max CPC can really throw off your averages if you almost always pay on the lower end of that spectrum. The headroom between your bid and the price you typically pay can vary widely, and that makes average CPC a metric that requires some context. Your max/actual CPCs appear across different, much more complicated distributions than I’m talking about in this chart, and that makes averages even less useful than what you see here.

adwords bid simulator

Look at the bid simulator report for some of your keywords to see how a new max CPC can throw off your averages. Raising your max CPC can get you additional clicks, but your average CPCs might increase by a higher rate.

If you’re curious about the CPCs you’re paying for different types of clicks, segment your reports and unpack performance a bit. Geos/devices/match types/audiences/network. Segment by things like hour of the day to see if lunchtime clicks cost less than dinnertime clicks. There are so many variables in play, and the average of all of those can lead to a sometimes useful, sometimes misleading metric.

4. How Does Google Determine What Is A Good User Experience?

User experience is hugely important when measuring quality. A wide variety of metrics are used to calculate good user experience with your ads, and that variety has evolved over time (and continues to evolve). You won’t be able to get a comprehensive list of those factors, but the good news is that the most important ones are already available for you to evaluate.

CTR is the clearest one in terms of compelling ads. Plenty of other things go into ad quality, and quality is controlled for the position where your ad appears. But if things are relatively stable in your account, you should look at CTR and see if things are improving over time. That’s a great way to gauge user experience.

Similarly, as I said above, conversion rate is an excellent proxy for landing page experience. If people are completing your desired actions when they visit your site, then you’re most likely delivering a good experience.

5. Is There An Account- Or Domain-Level Quality That Is Taken Into Account In The Auction?

The Auction doesn’t use any type of internal, account-level quality measurement. Quality ratings are about a user’s query and the ad that’s serving against that query. When you’re launching new keywords, those keywords learn from other similar keywords in your account, but that’s only at the beginning. And even then, it’s often because new keywords share ad copy elements with established keywords. It doesn’t take much traffic before keywords are judged on their own merits.

This is one myth that’s refused to die, so please spread the word: There’s no such thing as account-level quality.

As for domains, because they appear in the copy of your ads, users can see them, and domains can affect user behavior just like any other element of creative copy (i.e., they can influence an ad’s CTR, as many who advertise on their brand term may already appreciate). I wouldn’t recommend changing domains as a quality optimization tactic or anything like that, but do acknowledge that your domain may affect user experience.

6. If I Create A Mobile-Preferred Search Ad, Will This Increase Quality?

That depends. Are you writing an ad that does a better job connecting with mobile users? If so, you should generally receive better quality ratings. Enabling mobile-preferred creatives simply to have mobile-preferred creatives won’t increase your quality.

Quality is all about user experience, and a good experience on mobile is often different from a good experience on desktop. Your mobile-preferred creatives should be created for mobile users. If they deliver a better user experience than your standard ads, then quality should improve.

For Effective SEO, Content Is King


Marketers rely on search engine optimization (SEO) to improve search rankings, website traffic and lead generation, and in June 2015 research by Ascend2, 89% worldwide rated SEO successful at achieving these objectives. Further, in February 2015 polling by Econsultancy, fully 73% of in-house marketers and 76% of US agencies worldwide said SEO provided excellent or good return on investment (ROI), the first- and second-highest responses.

Most Effective vs. Difficult SEO Tactics to Execute According to Marketing Professionals Worldwide, June 2015 (% of respondents)

More than seven in 10 respondents said SEO effectiveness was improving, and content was by far the most important factor in this. Fully 72% cited relevant content creation as the most effective SEO tactic. Unfortunately, though, it was also the second most difficult, at 46% of respondents.

The second most effective area, keyword and phrase research, proved to be much easier at executing; however, this stresses content’s importance again, as these terms must relate to the content produced. Putting manpower toward frequent website updates as well as link building also ranked highly for effectiveness, though the latter was the most difficult to execute. While website URL restructuring was the least difficult, it was also the least effective.

Outside experts can help marketers maximize SEO efforts and effectiveness. As such 81% of Ascend2 respondents outsourced some or all of their SEO tactics.

Q4 2014 data reaffirms the importance of showing up at the top of search results. Looking at the clickthrough rate (CTR) of US desktop organic search results on Google, average CTR for branded keywords in position one was 45.7%, vs. 11.6% for second and 5.4% for third. Similarly, CTR for nonbranded terms was 24.7% for spot one, vs. 12.8% and 7.4% for Nos. 2 and 3. Mobile showed similar results. CTR for branded keywords in position one was 46.4%, vs. 10.8% in second and 6.1% in third. For nonbranded terms on mobile, first position saw an average CTR of 27.4%, dropping to respective rates of 17.6% and 13.9% for Nos. 2 and 3.