CitizenNet's Facebook Audience Planning and Purchasing Playbook

Iina Hanka | Download | HTML Embed
  • Jul 23, 2014
  • Views: 34
  • Page(s): 14
  • Size: 974.79 kB
  • Report

Share

Transcript

1 The Audience Planning & Purchasing Playbook for Facebook Advertising Moving Beyond the Fan: How to use Facebooks new Audience Insights Tool to Double your Facebook Advertising Results by Growing your Audience without Decreasing Response Rate

2 Contents Audience planning is Vital to Success with Facebook Advertising 1 The Problem with Only Targeting Your Fans 1 What about Lookalikes? 3 The Problem With Only Targeting Content on Display Networks 3 Targeting and Creative 3 A How-to Guide for Audience Discovery and Planning 4 Discovering your Target Audience 4 Building the Audience Plan 6 Example8 Building Out the Campaigns 9 Ongoing Optimization 9 End of Flight Check 9 How CitizenNets Technology works 11 More interest categories. 11 More powerful scoring 11 More detail 11 References12 CitizenNet CitizenNets patented technology leverages innovative processes in natural language processing, semantic analysis, and machine learning to analyze billions of data points from across the social web and deliver actionable insights for Facebook Advertising. Clients include Live Nation, Oakley, Sony, House of Blues, as well as other industry leading brands and the agencies that serve them. Solutions include both self-service platform licensing and fully-managed service. CitizenNet is a Facebook Ads Preferred Marketing Developer. www.CitizenNet.com July 2014

3 I f youre only talking to your fans youre missing the big opportunity on Why Facebook should be a bigger part Facebook. Before now it was hard to expand an audience intelligently, of your Marketing Mix without scaling too broadly based on demographics, or guessing at additional interests based on intuition. Using Facebooks New Audience Insights In considering channels, modern marketers have many options and may wonder, How tool and this guide, you can make data-driven decisions and enjoy reliable much of my budget should I spend on growth. At CitizenNet, we have used these same techniques to dramatically Facebook Ads? increase audience size without sacrificing response rates. The key is moving beyond Facebooks interest targeting and finding un-expressed affinity. When Currently, many brands are allocating 10-20% done correctly, these predicted affinity audience clusters drive both long- of their digital marketing budget to Facebook (this can be larger for brands that have term and short-term revenue, because you can identify new market segments additional budgets for content creation, and and increase overall awareness without sacrificing results. This guide and the community management). However, some included Excel Worksheet will provide a step-by-step solution to unlock the marketers are achieving breakthrough success power of Facebook Advertising using the new Free Audience Insights Tool. by allocating larger portions of their budget towards Facebook ads. The reasons are simple: Facebook has reach and frequency, the targeting and accountability that comes Audience Planning is Vital to Success with Facebook with digital media, and high impact creative. Advertising Reach and Frequency Headlines about Facebook Marketing can give you whiplash. Many brands Facebook reaches 71% of all American highlight strong success stories, while others publicly decry the platform. Which adultsA and 73% of 12-17 year olds. 64% of Facebook users visit every day.B Despite is true? Is Facebook a revolutionary marketing channel or just a fad? Based reports of Facebooks declining popularity on the results seen with many brands, its safe to say Facebook can be a game with younger demographics, Forresters latest changer IF it is done right--so whats the difference between brands that are reportC cites Facebook use among 13+-17 doing it right and seeing success and those that are doing it wrong and crying year olds is pervasive and a vital part of their foul? The answer is Audience planning. social networking activities for 89% of the group. With consumers spending so much time online, and with so much of this time Without successful audience planning, brands are either targeting too broadly aggregated on Facebook it is a must for any or too narrowly. For example, if an Education brand typically targets Adults brand of any size. Aged 18-49 and applies that same targeting to Facebook they can achieve a huge number of impressions, but they arent well targeted enough to generate Targeting and Accountability an acceptable level of response or return on investment. Facebook has huge As users spend time on the Facebook penetration with every demographic in the US (see sidebar Why Facebook platform, it captures information about their should be a bigger part of your marketing mix). However, unlike Paid Search interests and behaviors to provide one of advertising, which delivers consumers who are very close to making a purchase, the most targetable audiences ever available to a marketer. In June of 2014, Facebook Facebook reaches consumers much earlier in their decision journey. In order announcedD they would even begin to use to achieve a high return on advertising spend, advertisers must zero in on browsing history to improve targeting. subsets of their typical demographic target who are more likely to respond. The Facebook offers segmentation across alternative to audience planning is to rely on other targeting tools. Facebook geographic and demographic variables as well Fan targeting, Facebook Lookalike targeting, or display advertising targeted as audience groups from 3rd party research companies including Acxiom, Datalogix and to relevant content are three popular methods but each of these has their Epsilon. This is unprecedented access to limitations. research that might normally cost advertisers thousands of dollars. Other online sites are plagued by ad viewability issues and click The Problem with Only Targeting Your Fans fraud, but Facebook provides significant insulation against these problems, ensuring The easiest subset with the strongest predicted response is often a brands own ads are seen by actual human consumers fans -- Facebook users that have already sought out and liked the Facebook (not bots). This targeting combined with the page. This is the most common targeting strategy, but there are three big detailed reporting for optimization makes Facebook a reliable source. problems with this. High Impact Creative 1. Your fans arent new customers. Using the same Education client example, students often dont like a schools page until they are Facebook has one of the most integrated advertising environments of any online enrolled in a program. Facebook can be a great platform for driving channel. Users go to Facebook to read the incremental / repeat purchases, but many brands would like to News Feed and ads are IN the News Feed accomplish other objectives and look just like Native Content. This FACEBOOK ADS PREFERRED MARKETING DEVELOPER 1

4 2. Many people dont Like Pages. Even if the fans of your page are rich dramatically decreases banner blindness territory for finding new customers or driving incremental sales, the as long as the ads are relevant to the user. Facebook wisely limited Video Ads in the scale is severely limited because many people dont Like pages. newsfeed until users were used to seeing Clearly likes are not a good way to target potential customers. video content from friends, family, and Facebooks internal targeting algorithms are very good at targeting entertainment outlets. Now video ads are people who have expressed affinity (checking in at a store location, or truly native content and consumers actively positively mentioning a brand in a comment or post). However, this is engage with them. Within mobile, the ad units completely fill the display, making them still a very small subset of the potential audience, and often too small to particularly powerful. achieve the scale necessary for success. Mobile 3. Likers Like Lots of Pages. It turns out that those people who like pages, tend to like lots of pages, across lots of categories. Since most ComScore reportsE that in May of 2014, Mobile Apps surpassed the desktop browser brands are using interest targeting (likes) to build Facebook ad plans, and all other platforms for time spent online. (targeting their fans, or their competitors fans) the demand for these Facebook is the #1 mobile app with usage by likers is high. Because Facebooks ad marketplace is auction based, this 76% of the USF and 24% of ALL time spent drives up the cost to reach these audiences, making them inefficient. on mobile devices.G Facebook is bigger than Googles mobile search, Twitter, Instagram EVERYTHING. A Facebook campaign is a mobile campaign! Audience Feedback Facebook is the worlds greatest focus group. 32% Companies no longer have to recruit panelists of people have NEVER liked a brand and rent a room with a one-way mirror to find out what their customers think about a new on Facebook. product. Brands can test messaging strategy and targeting within Facebook and receive instant feedback in the form of actual user 22% comments, as well as metrics like engagement of those that have liked a company in and conversion rates. These insights can save money and improve response when applied the past dont do so anymore. to all other media channels, both online and offline. 66% Social Advertising budgets have been increasing by nearly 30% for the past several of users like fewer than 5 brands1 yearsH, and the first brands to make a serious commitment to Facebook have made major gains in their verticals. FACEBOOK ADS PREFERRED MARKETING DEVELOPER 2

5 What about Lookalikes? Facebooks built-in lookalike targeting is a great tool for moving beyond a core audience or page fans, however even this is too limited. Facebooks lookalike targeting hinges on identifying specific earmarks of your audience and then finding those same markers on other Facebook users. On the surface this sounds good, but the problem is the markers that are the most visible to Facebook may not be the ones that are most predictive of response. For example, Facebook may expand along dimensions of demographics or behaviors that have no bearing on response. Not all 18-24 year olds are likely to have the same response to an ad from an online education company. Lets assume Education level (High School Diplomas but No College Degree) and employment (Employed) are the key attributes for responsive Audiences in this case. Even if Facebook correctly identifies those factors, there may be significant differences between those who are employed, but are working in different types of jobs, and Facebook may not identify this nuance. Similarly the same type of job might indicate a different response rate depending on geographic location (Restaurant employees in a metropolitan city may be more career oriented than those in a rural area). Ultimately, Lookalike audiences are too limited, and dont necessarily target an audience that is likely to convert, just an audience that looks like another audience. Audience planning uses data to predict new audiences who dont look like your current customers but who do respond to your offers. The Problem With Only Targeting Content on Display Networks One way companies solve the targeting issue is selectively targeting content on publishers or ad networks. The belief is that this captures users who are expressing interest in the topic, and so are better qualified prospects. There are several problems with this approach. Usually the dominant brands in any category have such a strong presence on all the relevant pages that its impossible for a mid-tier brand to have any share of voice. This pushes those brands to second tier sites with questionable content, which then casts a negative light on the advertisers. Furthermore, users are often closer to a decision when they begin researching in this focused manner, and they are no longer gathering options to consider. By the time they are reading specialized content, they are only evaluating from a few choices which means advertising for a new brand is a wasted impression. Facebooks scale means its impossible for one brand to own all the relevant impressions. With proper targeting and audience planning, brands can concentrate on their best prospects even more efficiently than with other display options. Facebooks News Feed content is largely based on a users interests and past behaviors, so it is always high quality. Facebook offers the opportunity for a brand to engage with a prospect early in their consideration process, even CREATING interest with new customers. Targeting and Creative The key to relevant creative is understanding the audience and their unique needs. Digital media creates an opportunity for true Facebook offers the opportunity for 1:1 targeting, where each offer can be customized specifically to a brand to engage with a prospect the person reading it, but without well-defined audience planning, early in their consideration process, creative can be too generic or too specific. For example, some Education prospects might respond to messaging about a better even creating interest with new career and more money. Other targets respond better to messaging customers. that highlights convenience and student support. Proper audience planning and targeting ensures that the right message reaches the right customer at the right time. Facebook Advertising with proper Audience Planning solves all these problems. The platforms scale reaches a large audience and (when properly planned) delivers prospects with strong interest, who can be targeted with high-impact creative. With proper audience identification, compelling creative and efficient frequency brands can create and cultivate the interest and desire that grows revenues and delivers the success stories brands are seeking with Facebook. Clearly Audience planning is the key to success on Facebook Heres how to do it! FACEBOOK ADS PREFERRED MARKETING DEVELOPER 3

6 A How-to Guide for Audience Discovery and Planning The list of the biggest advertisers on Facebook reads like a Whos Who of Successful Marketers2. Whether your objectives are focused on direct response initiatives or branding and demand generation Facebook Advertising can be one of your most successful channels. Discovering your Target Audience Audience Discovery uses two very important concepts. A Core Audience and a Predicted Affinity Audience. A Core Audience is a group of people who have clearly expressed affinity for a product or service, and are representative of other potential targets. Typically, a Core Audience would be the existing fans of a relevant page. A Predicted Affinity Audience is the group of people we are predicting (based on the analysis described below) will have a strong response but are outside of the core audience. Predicted Affinity Audiences create scale, without sacrificing response rate. This group has an Affinity for the product, but its an Un-Expressed Affinity because they havent yet liked the products page. The Facebook Audience Insights Tool, released in late June of 2014, is a treasure trove Predicted Affinity of information about any audience. Put in just about any keyword or interest category and you will see a ton of demographic, psychographic and life stage information. You Audiences create scale, can also use this for insights on custom audiences, etc. We will use this tool extensively without sacrificing in this guide. As you build research on Audiences and create different parameters, be sure to save them in Facebook! That will make it easy to target them when you build the response rate. campaigns later. Continuing with the education vertical, lets say we are targeting registrations for an online business course focused on small business owners (not an executive MBA). Whats a good example of a core audience? Well known publications, blogs, and other specific resources are usually great magnets of curated audiences. Why not Entrepreneur magazine? We will use that, along with the Facebook category of Entrepreneurship. This gives us a reach of about 8m 9m monthly active users. This is still a bit broad. Looking at the other categories available, lets use the Work and Financial sections to hone in on our target: FACEBOOK ADS PREFERRED MARKETING DEVELOPER 4

7 Here we will select those users who work from a home or small office, and who have a Net Worth under $1M. Notice how adding these settings drops the reach to 1.5M 2M people. Well use this as our core audience. Sowere done, right? No. This targeting probably wont work too well and stopping here is why some brands fail with Facebook Advertising. First off, Facebook targeting isnt terribly precise. While this audience may seem like entrepreneurial small business owners, we really dont know if they would be interested in taking a business management course. Its safe to say there is some subset of these 1.2M people that would apply, but we dont know which subset or how to define those people - yet. The audience itself can guide additional segmentation. By looking at other interests within this core audience we can find some appropriate ways to segment the audience further to find the winners. The Page Likes tab gives us such a list and ranks other pages that our core audience may also like. Lets take a look at these pages. They seem to fall into the following buckets: 1. Entrepreneurial publications: Entrepreneur, SUCCESS magazine 2. Self-Help/Personalities: Brendan Buchard, Anthony Robbins, Zig Ziglar, Robert Kiyosaki, Sandi Krakowski, and Jim Rohn 3. Facebook for Business 4. Retailers: Diane Von Furstenberg, Homegoods, and Zappos 5. Health: Dr. Oz, Cleveland Clinic This brings up a subtle but very important point about this list: Do not simply target these interests! This page shows topics that our core audience may be interested in, but that doesnt mean they all want to signup for an online business course. A few of these categories demonstrate this. Diane Von Furstenberg is a renowned clothing designer and businesswoman. It makes sense shes on this list as she is an outspoken advocate for women starting and running businesses. However, the bulk of the Diane Von Furstenberg fan page is people who are really into fashion not entrepreneurship. How do we know this for sure? Try the following with these interest categories: Entrepreneurship: 950k Diane Von Furstenberg: 325k Together: 1,250k This means that there are actually very few people who like both, and thus it may be a bit dangerous to assume a lot of similarity between these groups. FACEBOOK ADS PREFERRED MARKETING DEVELOPER 5

8 Building the Audience Plan This leads us to the next step, building the Plan for targeting Audiences. In the supplied Excel file, you will find a worksheet that helps you capture all the relevant details and a number of formulas as well. Heres a snapshot of the Worksheet CitizenNet Audience Planning and Purchasing Worksheet *Expand Sections using the "+" above Audience Definition Audience Score Budgets Actual Performance Facebook's Audience Insights Pricing Summary Performance Relative Audience Cluster Name Audience Cluster Type Interests Work color CPM/Core CPM Score Cluster Budget Peformance Reach Frequency Headroom Actual CPM Projected Budget to benchmark Entrepreneur (magazine), Office Type Home Net worth under 100,000; Core Core $3,578.40 $0.78 -22% 665,280 1.8 454,720 $1.49 $1,524.45 Entrepreneurship Office, Small Office 100,000-1,000,000 SUCCESS magazine, Inc. Office Type Home Net worth under 100,000; Media Predicted Affinity 112% 2.11 $1,216.35 $0.83 -17% 302,940 2.7 37,060 $2.86 $238.48 Magazine Office, Small Office 100,000-1,000,000 Brendon Buchard, Anthony Office Type Home Net worth under 100,000; Personalities Predicted Affinity Robbins, Zig Ziglar, Robert 112% 0.48 $1,216.35 $1.24 24% 336,600 3 3,400 $1.27 $9.72 Office, Small Office 100,000-1,000,000 Kiyosaki Diane Von Furstenberg, Lean In, Office Type Home Net worth under 100,000; Fashion Businesses Predicted Affinity 127% 0.24 $2,187.00 $0.98 -2% 213,840 1.2 326,160 $3.60 $2,641.90 Nasty Gal Office, Small Office 100,000-1,000,000 The worksheet is broken out with Rows for different Audience Clusters (including the initial Core Audience and multiple Predicted Affinity Audiences) and Columns that describe aspects of that Audience. The Audience Definition and Audience Score Section can both be expanded. Well reference much of this example data throughout this guide. Audience Definition The first section of columns Defines the Predicted Affinity Audiences. Audience Score The second section of the worksheet Scores the Predicted Affinity Audiences based on Size and alignment with the Core using Duplication, Relevance, Similarity, and External Signals. The Summary Score is simply a weighted sum of each of the dimensions. The higher the value, the more likely the audience will perform well and respond to advertising messages. Combined with the Campaign Parameters table, this score creates budgets to be used for targeting these Audiences as part of a media plan. The Pricing in this tool can be either CPC or CPM. Although Discovering Audience Clusters to target is vital to this process, Scoring them is equally important. This worksheet has some detailed instructions and examples built in to help explain how the scoring works. Every Column in the Scoring section has comments. Simply hover over the header row for that column for a detailed overview of the values contained in the column. This FACEBOOK ADS PREFERRED MARKETING DEVELOPER 6

9 will help you see which columns require your input and which are calculated Direct Response Campaign Planning for you. All the columns that require inputs are marked with an arrow. The This worksheet uses Facebooks Relevance, Similarity, and External Signals calculations are particularly complex, recommended CPMs for estimating budgets requiring input from a couple different sources, so those calculations are further for each Audience Cluster and also building explained with screenshots on a separate Tab within the Worksheet. the media plan. Facebook also offers Cost Per Click bidding models in executing campaigns. An advertiser can easily build campaigns Actual Performance using CPC bidding to hit certain benchmarks. The final section is used to track actual performance of different audiences for It is worth noting, Facebook optimizes to ongoing optimization. their yield which means ads that get clicks get served more and at a lower cost to the advertiser. Its often a good idea to begin with a higher bid, to get better impressions, and then when an ad has a proven history of performance, lower bids. Additionally, on a separate tab in the worksheet there is a table for Campaign Parameters to be used in setting goals *this example has inserted values. Using this worksheet as a guide, follow these steps to develop an audience plan for Targeting and Optimizing Predicted Affinity Audiences. 1. Campaign Parameters: Fill in the desired Frequency and total budget for media (See Sidebar: Direct Response Campaign Planning). 2. Audience Clusters: For each Audience Cluster (Row), add audiences, completing details about the audience definition, and scoring parameters. Based on Campaign Parameters and the Audience Cluster Score, the worksheet will automatically calculate and recommend a budget for this cluster. 3. Audience Plan. Continue to Add Audiences until you have exhausted all available targets you can develop, or you have run out of budget. Tips: To increase reach for initial testing, reduce some of your restrictions on the audience. For example, if you think the probability of response from someone who likes Entrepreneur magazine is similar regardless of gender, then dont restrict to just one gender. Facebook currently has a 200-interest limit for each ad; so dont use too many small interest groups in building an Audience Cluster. FACEBOOK ADS PREFERRED MARKETING DEVELOPER 7

10 Increasing or decreasing Frequency can impact planning. Typical best practice is to use a Frequency of 3 or higher, but simple offers can succeed with less, and more complex messages may require more. The example worksheet provided has several values filled and can serve as a handy reference guide. Data is accurate as of July 2014. Example In this example we chose three audience clusters for expanding beyond the core audience. The Clusters are Media, Personalities, and building on the Diane Von Furstenberg interest Fashion Businesses. We maintained the Facebook targeting to Home Office and Small Office and net worth under $1M as a way of finding those segments of these Audience Clusters that are the most appropriate for testing response. Facebooks targeting will look at obvious connections, but using Audience planning, marketers can leverage insights to find new connections. Platforms like CitizenNet can also help with making Semantic connections (leveraging semantic meaning instead of nominal characteristics). As a marketer you can use other External Signals to drive this type of discovery. For example, Amazons Directory on Top Books for Small Business and Entrepreneurship delivers some interesting insights. Currently (July 2014), the #2 book is #GIRLBOSS by Sophia Amoruso, the CEO of Nasty Gal clothing company. Combining this insight with the fact that Diane Von Furstenberg was a top interest category indicates there is overlap and opportunity. Clicking on the product page for #GIRLBOSS shows Sheryl Sandbergs book, Lean In, as a Frequently bought Together recommendation. Lean In is a targetable interest on Facebook. #GIRLBOSS is not, but Nasty Gal is keeping our Office Type and Net Worth filters in place delivers potential targets from this page and helps it make sense for audience expansion. As we move into Audience Scoring we examine the overlap between the different groups (Core and Predicted Audiences) using shared % with the Core as a sanity check we see the Fashion Business Cluster has 13% Duplication almost as much as the other, more intuitive clusters. To round out the Media Cluster and provide some additional reach, we again use Amazon as an External Resource. Looking for other recommended magazines for SUCCESS Magazine, Inc. Magazine comes up. Well add that to the Cluster as well and pull in that interest groups details from the Facebook Audience Insights Tool. Other external recommendation engines (e.g. Google Insights) provide additional related interests for expanding Audience Clusters. FACEBOOK ADS PREFERRED MARKETING DEVELOPER 8

11 Building Out the Campaigns Once the plan is complete, the setup of these campaigns on Facebook is pretty easy. Weve already done most of the legwork: each cluster represents a segment of our audience that we predict will be interested in our online education offer. The people within a cluster are similar to each other in some respect, but are different enough from people in other clusters to allow for testing and optimization. The best way to execute this is to make each cluster an adset on Facebook, with the initial budget dictated by the budget column on the worksheet. You can test creatives by putting them in the same adset with the same targeting Facebook will use their own optimization algorithm to weight the best performing creatives. If you want to get a true picture of performance youll have to set up separate adsets for different creatives and split the budget between them The benefit of Audience yourself. Planning is purposefully You have already saved the audiences in the Facebook Audience Insight tool, so its just finding new customers to a matter of loading them from the Audiences section of Power Editor, filling out the grow your business. rest of the campaign details, and then its off to the races! Ongoing Optimization Optimization windows are highly dependent on the volume of impressions and daily budgets. Typical best practice is to make changes weekly, but to check campaigns daily and ensure ads are approved and delivering impressions etc. During these checks you will begin to see variation in performance between different Audience Clusters-- this is good! Remember, the benefit of Audience Planning is purposefully finding new customers to grow your business, so it is speculative. Thankfully, Facebook controls the spend (roughly) to a level amount each day, so exposure is under your control. For most campaigns, we suggest a weekly cadence: allow your campaigns to run for a week without changing targeting or creative, but optimize for performance by allocating money from your worse performing ad sets to your better performing ad sets. After the first week, dont change budgets more than about once a day, because it may cause campaigns to under or over-spend. End of Flight Check Once your campaign is done, its time to take a look at what worked and what didnt. This can be captured in the Actual Performance section of the Excel worksheet. (Weve put some illustrative numbers in the spreadsheet for these campaigns these are representative of typical performance we have seen): Lets take a look at some of these findings: The Core beat our benchmark by 22%, which is great. It means that Entrepreneurship is a correct core audience to base our decisions on. The Media cluster also did well. From a scoring standpoint, it seemed that Shared Interests and External Signals were great predictors. It also makes good intuitive sense. The Personalities cluster performed the worst, from a scoring standpoint, this cluster had the highest Relevance, however. What gives? Well, although 47% of Personalities would consider themselves entrepreneurial, these followers may be more into broader life coaching themes, and arent too interested in business courses. FACEBOOK ADS PREFERRED MARKETING DEVELOPER 9

12 The Fashion Business cluster also performed at our historical Expressed Affinity vs. Unexpressed benchmark. Congratulations! You just found a non-intuitive audience Affinity with Unexpressed Affinity! These are people that normally dont associate themselves with the self-declared entrepreneur bunch (this Moving beyond Fans with Expressed Affinity to targeting new predicted audiences with has the lowest Shared Interests score). This cluster is more expensive Unexpressed Affinity can double advertising but it converted within our constraints. Best of all, it added another half results on Facebook. As discussed elsewhere million new potential customers that respond better than average! in this playbook, Fans with Expressed Affinity represent too small of an audience and This is the key to getting the most out of Facebook Advertising. Identifying non- often cost too much for a campaign to be successful. Predicting new audiences has intuitive audience clusters with Unexpressed Affinity opens the floodgates for numerous benefits. finding new customers (efficiently) and growing your revenues. Many brands have not yet mastered the So what do we do next? A few options: techniques described in this guide for developing new audiences. There is still Double Down on your Success. We have audiences leftover, as a great deal of un-tapped potential within indicated in the Headroom Column. Headroom is the number of users Facebook and early movers will have a in this audience cluster who have not yet seen the ad. We can set-up decided advantage. Just as the first brands to additional campaigns to reach the rest of them. There may be some master Paid Search generated above market returns till their competitors caught up, the erosion in performance as Facebook may begin to increase Frequency same opportunity exists now within Facebook. with the highly active members of this audience at the expense of Human powered audience prediction can be additional reach, but there may be lots of potential left. tricky, but using the strategies and tactics Tweak the Scores. From these results External Signals correlated described in this guide, any brand can find best with performance, whereas Shared Interests and Relevance had new productive audiences on Facebook. no correlation. Audience Similarity could go either way maybe There are many tools available to assist different attributes are better predictors? Once the score is tweaked, in Audience Prediction and Optimization. you can find new, better audiences armed with insights from previous CitizenNet is focused on solving this problem, campaigns. leveraging several U.S. patents in natural language processing and semantic analysis Hunt. Not happy with the results? Then you need to revisit your to predict and discover new audiences for strategy. If your Core Audience isnt working well, then you will need to brands. This technology leverages insights hunt for new targets to base your strategy off of. about the semantic meaning of an audiences Refine. You may be able to further improve performance by taking your interests to predict new audiences likely to best clusters and splitting them further for example, experiment by have the same response. segmenting by gender, income, geography or other dimensions. For example, if off-road biking enthusiasts Improve the Creative. This is potentially the most powerful. Now that show a strong response to a particular offer, you have discovered this new Fashion Business audience and are CitizenNet evaluates the characteristics of armed with all this knowledge about the clusters, tailor your message that group against other potentially strong to them! See through their eyes and focus on what machines cant do: matches. Rather than only examining the nominal aspects of this group, CitizenNet make a human connection! examines the semantic aspects of the group. A nominal analysis would only focus on biking enthusiasts of different varieties, since off-road describes biking, the main characteristic of the group. A semantic analysis recognizes that off-road may in fact be the most salient characteristic of the group, and so trail running, or hiking interest categories would be explored along with bike racing and other categories. If you are interested in scheduling a product demo or learning more about CitizenNets technology and how agencies and brands are using this for both self service and fully managed solutions, please contact us [email protected] CitizenNet.com. FACEBOOK ADS PREFERRED MARKETING DEVELOPER 10

13 How CitizenNets Technology works As you can see, manual Audience Planning and Optimization is a fair amount of work. The good news is that CitizenNets platform is designed to automate all of this work, and more thoroughly than an army could; heres just a few examples of the benefits of using technology. More interest categories. In our example, there was a lot of labor to create only a limited number of interest keywords. For each campaign, CitizenNet collects and tests 10,000+ different interests all pulled through current conversations on Twitter, Facebook, Youtube, and other social platforms. More audience attributes In our example, we had several different attributes that described the audience (the columns in the spreadsheet). CitizenNet builds a description based on over 50 different features. More powerful scoring In our spreadsheet, we use a simple linear score, but thats not terribly accurate. High overlap between the Affinity Audience and the Core Audience might diminish importance. The CitizenNet platform, based on cutting edge, deep- learning algorithms, builds much more complex non-linear models that more closely approximate human behavior. More detail In our example, we build our score off of very limited information. Is it too early to tweak the weighting after just one flight? If results change over time, is it because of our tweaking, or is the audience changing? CitizenNets algorithms leverage the billions of historical data points gathered running other campaigns and the platform builds in sophisticated tests for statistical significance by running mock predictions behind the scenes. This is how CitizenNet, on average, finds Predicted Affinity audiences that are twice the size of the Core Audience but have the same response rate. There you have it! The complete How-To Guide for doubling your results with Facebook Advertising. Its not a quick fix, but building and testing audiences within Facebook will grow your business and provide valuable insights on targeting and creative that can be used across your media. We welcome your feedback on this guide and hope it leads to ever greater results for you and your business. To your Success! The CitizenNet Team FACEBOOK ADS PREFERRED MARKETING DEVELOPER 11

14 References 1 http://www.exacttarget.com/resources/SFF10_highres.pdf 2 http://www.businessinsider.com/top-advertisers-on-facebook-2013-11 A US Census and Facebook ad Platform Statistics B http://www.pewresearch.org/fact-tank/2014/02/03/6-new-facts-about- facebook/ C http://blogs.forrester.com/nate_elliott/14-06-24-facebook_still_dominates_ teens_social_usage D http://adage.com/article/digital/facebook-web-browsing-history-ad- targeting/293656/ E http://www.comscore.com/Insights/Blog/Major-Mobile-Milestones-in-May- Apps-Now-Drive-Half-of-All-Time-Spent-on-Digital F https://www.comscore.com/Insights/Market-Rankings/comScore-Reports-May- 2014-U.S.-Smartphone-Subscriber-Market-Share G http://www.comscore.com/Insights/Blog/Major-Mobile-Milestones-in-May- Apps-Now-Drive-Half-of-All-Time-Spent-on-Digital H http://www.slideshare.net/eMarketerInc/emarketer-webinar-7-social-media-ad- trends-for-2014 FACEBOOK ADS PREFERRED MARKETING DEVELOPER 12

Load More