Your SlideShare is downloading. ×
718 SW Alder St., Suite 200   www.blitzlocal.com   www.facebook.com/blitzlocalPortland, OR 97205                          ...
Executive SummaryAfter Facebook released the ticker, timeline, and other changes, some brandscomplained of a decline in or...
"Sponsored Stories", ad mechanisms, to directly combat the "EdgeRank"problem (competing for attention) and the time window...
Section 1: Organic Impressions are Declining                                                                              ...
There is a -0.56 correlation between organic and paid impressions (Figure 1.2).We expect this correlation to get stronger ...
Has Engagement Rate By Month Declined?                                                   Less Than 100,000 Fans           ...
(Figure 1.4). Despite the common perception that engagement tends to         decline as page size increases, the low value...
More People Are Talking About Brands                                                                                      ...
In addition to the amount of content generated on the Facebook wall each day,we see another trend emerge from the data tha...
Changes in Content Type are Hurting Total Coverage: Some types of contentgenerate more impressions than other types. For e...
Pages With Less than 100,000 Fans         Post Type Affects Organic Coverage                                              ...
Section 2: Paid Impressions Help                                                         Over the last 6 months 82% of imp...
The Majority of Impressions Are Organic                                      Figure 2.1                           Viral Im...
Pages with over 100,000 fans that have advertised in the last six months saw aboost in engagement rate and organic power. ...
Better paid Coverage Equals Better Engagement                                                                             ...
Section 3: The Cost-Effectiveness of Advertising is Decreasing                                                            ...
It is possible that the decrease in CTR may be due to sales-driven ad copy thatwill have lower CTR since it’s not engageme...
Though technically the cost of traffic is decreasing, it is still a bargain at $.34CPM. On average, $1,000 of ad spend hit...
From 2010 to 2011 the cost of traffic increased as well. The average CPC of2011 was up 34.8% from $0.69 to $0.93. Despite ...
CPC            2011              2010                                                        Figure 3.5                   ...
What Brands Are Saying“Year over year the amount of content available on Facebook has increaseddramatically. More brands a...
“At Gordmans, we were able to capitalize on the low CPC in 2011 and run amultitude of ads throughout the year. We found th...
Contributing StaffElliot Swan                                              Travis KingInteraction Designer                ...
Upcoming SlideShare
Loading in...5
×

What 120 Billion Facebook Impressions Have Taught Us - Dennis Yu

5,413

Published on

What 120 Billion Facebook Impressions Have Taught Us
Dennis Yu, CEO, BlitzLocal

We will go through key leanings by vertical, across 13,00 pages and 2,00 ad campaigns. How does the new Timeline affect brands and what are the three secrets to beating the Facebook EdgeRank algorithm to get more exposure in the newsfeed? Facebook admitted that 86% of fans never see what you post- a number that is only getting worse.

Walk away from this session armed with:
- an understanding of which Facebook metrics can be gamed
- benchmarks for your industry that you can measure yourself against
- three proven tips to increase newsfeed optimization
- a 110 page Insider's guide on Facebook marketing with real data and examples

Published in: Technology, Business
0 Comments
2 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
5,413
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
26
Comments
0
Likes
2
Embeds 0
No embeds

No notes for slide

Transcript of "What 120 Billion Facebook Impressions Have Taught Us - Dennis Yu"

  1. 1. 718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 1  
  2. 2. Executive SummaryAfter Facebook released the ticker, timeline, and other changes, some brandscomplained of a decline in organic impressions. Where is this occurring, andhow has the effectiveness of organic impressions trended against paid efforts?Our research spans data from a 7-month time period from June 1, 2011 toDecember 31, 2011: • 11,365 pages • 5,744,550 admin and user posts • 76,202,654 Facebook users spread over 75 countries • 91,009,142,252 organic impressions • 27,969,233,010 paid impressions • 1,142,652,653 viral impressions. • 120,111,027,915 total impressionsThis is a continuation of BlitzLocal research performed in January 20111, nowwith six times as much data and correlating organic data. Limitations of thisstudy are the exclusion of data from applications (“apps”) and open graphobjects, which likely have an effect on organic impressions.According to our data, the following factors affect organic impressions: • Engagement – Engagement affects a pages EdgeRank and how often Facebook will show posts. • Increased competition in the newsfeed – More objects in the newsfeed means fewer people are seeing brand content. • Content type – Some content types are more engaging. • Post length – Post length affects how many people engage with a post. • Post decay – Post timing is important as posts only live in the newsfeed for so long. • Sales related posts – Sales related posts drive less engagement and impressions.From our data we drew the following conclusions:1.) As more users generate more content via more apps and devices, brandmessages get drowned out in the noise. Organic impressions per page aredown 33% due to a decrease in frequency, not reach. We did not detect anyadjustment by Facebook to force brands to have to advertise to reach theirusers: theres just more "stuff" and Facebook has to apply a higher power filter.2.) Content that used to live for a day, may now live minutes in a usersnews feed. This concept is evidenced by decreased frequency of exposure aswell as the "half life" of interactions. Facebook created several forms of                                                                                                                1  http://www.allfacebook.com/surprise-­‐timeliness-­‐boosts-­‐facebook-­‐engagement-­‐2011-­‐06  718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 2  
  3. 3. "Sponsored Stories", ad mechanisms, to directly combat the "EdgeRank"problem (competing for attention) and the time window problem. These adsare effective in growing fans and amplifying the brands message. Facebookmetrics greatly underestimate the power of word of mouth.3.) While brands with paid advertising saw positive benefits in fan growthand organic exposure, the cost-effectiveness of Facebook ads declined,even as brands advertised more heavily.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 3  
  4. 4. Section 1: Organic Impressions are Declining From July to December 2011 organic coverage Highlights   (number of organic impressions per 1,000 fans) decreased 33%, from 836.1 to 608.• Organic  Coverage  is  down   33%  from  July,  while  paid   Some social mechanics to understand: Brands can Coverage  was  up  141%  in   increase their impressions by growing their reach November  and  121%  in   (the number of fans that see their messages) or December.   their frequency (how many times, on average, a fan sees these messages). Frequency is down• Post  decay,  increasing   competition  in  the  39.1%, while reach is up 12.5%, providing a net effect of a 31.5% decrease in impressions. newsfeed,  sales  oriented   Impressions per 1,000 fans are the product of reach posts,  and  changes  in   times frequency divided by total fans. We saw that content  type  may  have   the decline in frequency pushed impressions down contributed  to  a  decline   by 34.5%, while reach pushed impressions up by in  engagement,  but  we   less than 1.1% (Figure 1.1). In this same period, can’t  conclusively  make   this  assertion.   there was a spike in paid coverage driven primarily by the holiday season resulting in 934.3 paid impressions per 1,000 fans in Novght,ember (a 141% increase from October) and 691.9 paid impressions per 1,000 fans in December (a 121% increase from October). Reach Frequency Declines, but Reach Increases Figure 1.1 100 12 Unique Users/1,000 Impressions Total Impressions per User 80 10 8 60 Correlation=-­‐0.79   6 40 4 20 2 0 July, 2011 August, 2011 September, 2011 October, 2011 November, 2011 December, 2011 718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocal Portland, OR 97205 4  
  5. 5. There is a -0.56 correlation between organic and paid impressions (Figure 1.2).We expect this correlation to get stronger as Facebook rolls out ads in thenewsfeed, which will squeeze out organic impressions.Organic Impressions: Impressions in the news feed, ticker, and from fans visiting a brands page.Paid Impressions: Number of ad impressions.Why the Drop in Organic Coverage?Engagement may be decreasing: Our analysis should show the decrease inengagement rate to be seasonal, but without a full year of data on the newmetrics this cannot be judged conclusively. While engagement rate isn’tdefinitively down, impressions are down – meaning that Facebook is better atfiltering users news feeds. In other words, Facebook is showing fewerimpressions but still getting the same level of feedback.Facebook’s EdgeRank rewards pages in the news feed based on how manyinteractions they receive. Interactions are the number of likes, posts, andcomments a page has. For pages with more than 100,000 fans, engagementdeclined 21% from 3.8% to 3% between October and December 2011. Forpages with fewer than 100,000 fans, engagement also fell 26.0% from 16.0%to 11.8% (Figure 1.3). In this same period, active users fell by 27.3%.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 5  
  6. 6. Has Engagement Rate By Month Declined? Less Than 100,000 Fans Figure 1.3 100,000+ Fans December, 2011 Correlation=0.57 November, 2011 October, 2011 September, 2011 August, 2011 July, 2011 June, 2011 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20%Engagement Rate: Total Interactions/fan count over a 30 period.A shift towards sales-related posts in October, November, and December was apossible contributing factor to the drop in engagement and increase in fanremovals. Matching words like “buy,” “sale,” and “purchase”, we saw an 11%increase in sales-related posting in October, a 26.5% increase in November, anda 13.2% increase in December (Figure 1.4). Sales Related Posts Rise in October, November, and December 7% Figure 1.4 Percentage of Posts That Are Sales 6% 5% Related 4% 3% 2% 1% 0% June, 2011 July, 2011 August, 2011 September, October, November, December, 2011 2011 2011 2011Another possible contribution to the decrease in engagement was an increasein page sizes. There is a -0.12 correlation between page size and engagement718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 6  
  7. 7. (Figure 1.4). Despite the common perception that engagement tends to decline as page size increases, the low value of this correlation suggests a weak relationship between page size and engagement. It is possible that the first fans of a page are the most loyal. When the page gains mainstream presence it may have a more general audience, contributing to lower engagement. 3.0 Larger Pages Have Slightly Less Engagement Figure 1.5 2.5 2.0 Correlation=-­‐0.12   1.5 1.0 0.5 0.0 The data does not conclusively say engagement is down; rather it has gone up in the Fall and down in the Winter. Fourth quarter sales-oriented messaging may account for the drop, but without year-over-year data we are unable to confirm or disprove this hypothesis. If engagement is steady in light of impressions dropping, it may be a sign that Facebooks algorithms are doing a better job in selecting what to show users. Clearly there is a lot more content, but Facebooks filtering mechanism has improved. It is interesting to note that despite a decrease in engagement rate, there are more people than ever talking about brands. Between July and December 2011 there has been a 15.3% increase in People Are Talking About This (Figure 1.6). This could indicate a broader group of active users are interacting with pages, but less frequently. Different  Angles  to  Measure  Engagement   People  Are  Talking  About  This:  Total  number  of  people  who  have  had  a  brand-­‐ related  action  show  up  in  their  newsfeed.   Coverage:  Total  number  of  impressions  per  1,000  fans.       Unique  Reach:  Total  number  of  people  seeing  impressions.   Engagement  Rate:  Total  interactions  over  a  30-­‐day  divided  by  fan  count     Feedback  Rate:  Percent  of  people  who  saw  a  post  who  commented  or  liked.     Note: Facebook’s web insights provides up to 89 days worth of data, so brands must store historyda   themselves, use a 3rd party analytics like BlitzMetrics, or lose it forever. 718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocal Portland, OR 97205 7  
  8. 8. More People Are Talking About Brands Figure 1.6 People Talking Are About This/1,000 8.2 8.0 7.8 7.6 7.4 Fans 7.2 7.0 6.8 6.6 6.4 6.2 July,2011 August, 2011 September, October, November, December, 2011 2011 2011 2011Answering the Engagement Rate Question: Regardless of whether engagementrate has declined or not, People are Talking About and engagement rates arelow no matter how you cut it. This is likely less about Facebook algorithmicallyharming brands, and more about brands focusing on growing fans for the lastcouple of years and only starting to focus on engagement recently. Year afteryear we have seen a shift in attitudes and priorities of companies in regards toSocial Media. In 2009, 2010, and 2011 most companies fearfully expandedsocial budgets in order to not be left behind by competitors.2 Companies whoplanned social budgets for 2012 were looking heavily towards measurement andengagement.3Newsfeed Competition: Facebook cites 900,000,000 objects (events, pages,community pages and groups) that users can interact with on Facebook.4 Inaddition, there are 36.2 million user actions per hour showing up in the newsfeed. 5More and more content are competing for visibility, appearing in the availableimpressions in a user’s news feed. Similarly, ranking on Google is morechallenging today than eight years ago, given that more pages arecompeting for the first page of search results.                                                                                                                2 http://www.emarketer.com/blog/index.php/marketers-spending-social-media-wrong-reasons/3 http://www.emarketer.com/Article.aspx?R=1008667face4  https://www.facebook.com/press/info.php?statistics  5  http://www.jeffbullas.com/2011/04/28/50-­‐fascinating-­‐facebook-­‐facts-­‐and-­‐figures/  718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 8  
  9. 9. In addition to the amount of content generated on the Facebook wall each day,we see another trend emerge from the data that could reduce organicimpressions. From October to December, the overall time spent on Facebookdropped 15%, from 28:26 minutes to 24:43.6 With users spending less time onFacebook, and more content elements competing for display, the naturalproduct of these two trends was a decline in organic impressions.Post length could easily affect the number of organic impressions. On averagepage posts are 157.7 characters, while user posts are 121.5 characters andmobile posts average 104.9 characters. Posts between 140 and 159 characterslong are, on average, 13.3% less engaging than posts between 120 and 139characters and 9% less engaging than posts between 100-119 characters. Whatthis demonstrates, is that the content many brands are competing against for aspot in the news feed, is more engaging and readable (Figure 1.7). Longer Posts Perform Poorly Questions Non Questions Figure 1.7 160-179 140-159 Character Length 120-139 100-119 80-99 60-79 0.000% 0.005% 0.010% 0.015% 0.020% 0.025% Engagement                                                                                                                6    http://www.alexa.com/siteinfo/facebook.com#    718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 9  
  10. 10. Changes in Content Type are Hurting Total Coverage: Some types of contentgenerate more impressions than other types. For example, status posts haveweaker coverage on average, at 29.1% for pages with less than 100,000 fansand 54.8% for pages with more than 100,000 fans (Figure 1.8).We also see link coverage is decreasing, while status coverage is increasing. Apost with a link in it is more likely to be a sales-oriented post, especially inDecember. If true, then the increase in link posts may very well tie withengagement dropping in December. Brands Are Posting Weaker Content Status Videos Figure 1.8 Photos Links December, 2011 November, 2011 October, 2011 September, 2011 August, 2011 July, 2011 June, 2011 100 200 300 400 500 600 700 800 Organic Impressions/1,000 Fans  In the last 6 months there was an increase in the number of status posts. InJune, 43% of posts were status updates. At the same time the number of linksper month decreased 57%. On average, links have some of the highestcoverage, at 77% for pages with less than 100,000 fans and 69% for pages withmore than 100,000 fans (Figure 1.9). The decline in links and increase instatus posts affected overall organic impressions. Posting weakerperforming content ensured that fewer fans would see it.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 10  
  11. 11. Pages With Less than 100,000 Fans Post Type Affects Organic Coverage Pages With 100,000+ Fans Figure 1.9 Video Link Question Photo Status 0 100 200 300 400 500 600 700 800 900 Organic Impressions/1,000 FansPost Decay: Our analysis shows that 70% of interactions occur within 60minutes of a post. In November and December of 2011 brands posted anaverage of 1.19 and 1.17 times a day, respectively (Figure 1.9). In November70% of interactions, on average, occurred within 71.4 minutes, and inDecember 70% of interactions occurred within 70.2 minutes. Brands should beposting more to reap the benefits of potentially missed interactions. Brands Are Posting More Often Figure 1.9 1.4 Average Number of Posts Per Day 1.2 1.0 0.8 0.6 0.4 0.2 0.0 June, 2011 July, 2011 August, September, October, November, December, 2011 2011 2011 2011 2011718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 11  
  12. 12. Section 2: Paid Impressions Help Over the last 6 months 82% of impressions were Highlights   organic, 17% were paid, and 1% were viral (Figure 2.1). Even if total paid impressions were to double, • 82%  of  impressions  are   it would not have much of an effect on news feed organic  and  17%  are  paid.   competition. Other factors such as the number of ad • 77%  of  pages  are  not   slots (which Facebook recently raised to 7) could advertising.   also increase perceived competitiveness. • Large  pages  that   advertise  are  42.9%  more   Viral impressions have remained a small percentage engaging  than  pages  that   of total impressions. Over the last 6 months, their do  not.    Small  pages  are   share has decreased 82%, from an average of 3.3% in 21.2%  more  engaging.   July to an average of .6% in December. When compared with overall organic impressions, there is • There  is  a  strong  positive   a 0.69 correlation. Viral impressions are dependent correlation  between  paid   on fans seeing posts and sharing them. Because the impressions  and   overall number organic impressions decreased in engagement  rate.   November and December, the viral impressionsda   decreased as well. A sizeable amount of what is listed as organic might be considered viral, too. When we run ads we see that the organic and viral components increase. In the same way, when we run TV ads, we see more people searching for the brand name in Google. Not detailed in this research is the impact that Facebook campaigns, organic or paid, have on other channels. The true incremental boost of any social marketing effort is not attributable at the user level, so we have to correlate traffic by channel. 718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocal Portland, OR 97205 12  
  13. 13. The Majority of Impressions Are Organic Figure 2.1 Viral Impressions, 1% Paid Impressions, 17% Organic Impressions, 82%Only 23% of pages advertised on Facebook in the last 6 months, with theremainder relying on organic fan growth and impressions (Figure 2.2).Considering that only 17% of impressions are paid, this suggests that manybrands advertise only on a spot basis. Companies are running campaign-oriented ads, as opposed to persistent word-of-mouth ads, which Facebookrecommends. In 2012, we will see whether Facebook can wean marketers froma start/stop mentality, as is the case with most forms of advertising, and shiftto see Facebook ads as organic amplification with no end date. Room for Growth: 3 in 4 Brands Are Not Advertising Yet Figure 2.2 Brands Advertising, 23% Brands Not Advertising, 77%718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 13  
  14. 14. Pages with over 100,000 fans that have advertised in the last six months saw aboost in engagement rate and organic power. There is a positive correlation of0.47 between the number of paid impressions and engagement rate. Overall,pages advertising with more than 100,000 fans had a 42.9% higherengagement rate, while pages advertising with less than 100,000 fans had a21.2% higher engagement rate than those not advertising (Figure 2.3). Less Than 100,000 Fans Advertise To Get More Engagement Figure 2.3 Pages That Are Advertising Pages That Are Not Advertising 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20%Interestingly, the effectiveness of advertising appears to be mostly linear.There is a 0.70 correlation between paid Impressions per 1,000 fans andengagement rate (Figure 2.4). The more a brand spends, the higher theirengagement. We should note that while paid activity does not affect organicrankings in search engines, this does not hold true in Facebook.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 14  
  15. 15. Better paid Coverage Equals Better Engagement Figure 2.4 14% 12% 10% Engagement Rate 8% 6% Correlation=0.89   4% 2% 0% 0 100 200 300 400 500 600 700 800 900 1,000 Paid Impressions Per 1,000 FansPages that advertise get a 373% increase in organic impressions. This brings usback to an earlier point: the way Facebook defines viral traffic understates thepower of word of mouth. Share of Impressions For Pages That Advertise Figure 2.5 Viral Impressions, 0.4% Paid Impressions, 44.7% Organic Impressions, 54.9%Note: Data includes ad impressions landing on external URLs.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 15  
  16. 16. Section 3: The Cost-Effectiveness of Advertising is Decreasing Highlights   Facebook’s ad inventory is growing rapidly now that there are more users and more ads per page. • Between  October  and   They haven’t even begun to show ads on mobile, December,  the  average   which is 33% of all Facebook postings.7 CPC  increased  by  135%   and  the  average  CPM   The number of brands advertising is increasing increased  by  55%.   faster than Facebook is growing. As a result they • Both  the  Social  CTR  and   are driving prices up, decreasing the cost- non-­‐social  CTR  have   effectiveness of advertising. Traffic prices are declined  by  56.8%  and   going up only mildly, but the cost per click is 58%  respectively.   increasing at a faster rate than the cost of raw traffic because of lower CTR. • Between  2010  and  2011   the  average  CTR  rose  by   From October to December 2011, the average 6%  and  the  average  CPC   CPC increased by 135% and the average CPM rose  by  34.8%.       increased by 55% (Figure 3.1).  da   From October to December 2011, the ratio increased from 2.38 to 2.42 impressions per social impression. For 2011, Social ads were 56.7% more likely to be clicked on than non-social ads. Fewer social impressions brings down the average CTR, resulting in the increased CPC and CPM we see. Despite the Social CTR being much greater than CPM,  CPC,  and  CTR   the non-social CTR, both declined in the last three months of 2011. Between October and There  is  a  direct   December, the Social CTR decreased by 56.8% relationship  between   and the CTR decreased by 58% (Figure 3.2). This CTR,  CPM,  and  CPC.     change is explained by the effect that total paid Facebook  will  charge   impressions have on CTR, Social CTR and CPC. more  per  click  if  the  click   CTR has a -.77 correlation with paid impressions, through  rate  is  lower.     Social CTR has a -0.84 correlation with paid Brands  have  an  incentive   impressions and CPC has a 0.86 correlation with to  be  relevant  and  social,   paid impressions. Note that Facebook’s recent while  maintaining  this   change in the number of ads displayed (six as of mechanism  also  makes   November) could contribute heavily to the Facebook  more  money.     decline in CTR and the increase in impressions. da     da                                                                                                                   7 http://www.litmanlive.co.uk/blog/2011/05/twitter-facebook-mobile-stats/ 718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocal Portland, OR 97205 16  
  17. 17. It is possible that the decrease in CTR may be due to sales-driven ad copy thatwill have lower CTR since it’s not engagement-driven. The fact that CPMsincreased only slightly relative to CPC increases, indicates that the priceincreases are CTR-driven, not demand driven. A CPM increase means that addemand is outstripping Facebook’s white-hot inventory growth. CPC CPM CPC and CPM Increasing $1.20 Figure 3.1 $1.00 Correlation=0.97   $0.80 $0.60 $0.40 $0.20 $0.00 October, 2011 November, 2011 December, 2011CPC: Cost Per Click.CPM: Cost Per 1,000 Impressions. CTR Social CTR A Decline in Social and Non-Social CTR Figure 3.2 0.09% 0.08% Correlation=0.97   0.07% 0.06% 0.05% 0.04% 0.03% 0.02% 0.01% 0.00% October, 2011 November, 2011 December, 2011Non-Social: Clickthrough Rate of all ads that are friend of fan targeting.Social CTR: Clickthrough Rate of friend of fan ads.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 17  
  18. 18. Though technically the cost of traffic is decreasing, it is still a bargain at $.34CPM. On average, $1,000 of ad spend hits 658,823.5 unique users (reach timesfrequency equals impressions).Advertisers Suffer Ad BurnoutUnlike search ads, Facebook ads burn out quickly if brands are not refreshingthem weekly, if not daily. Because more than half of Facebook users log indaily8, brands that practice "set it and forget it" show their ads dozens of timesto the same users. Of course these ads dont perform.Though prices are increasing, not all industries fared the same. When split upby category, CPC has a positive correlation of 0.41 with total impressions.There is a negative correlation of -0.23 between impressions and CTR. Thisappears to support a connection between impressions and CTR, as well asimpressions and CPC. For example, retailers had a 67% increase in CTRbetween 2010 and 2011. Over the same period, total impressions declined32.7%. Automotive companies, on the other hand, had a 66% decrease in CTR,while seeing a 73% increase in impressions in the same period (Figure3.39). Bigbrands do not do a better job in advertising compared to small biz, despitetheir tools, bigger budgets, and general sophistication. The Relationship Between CTR and Impressions 0.18% Figure 3.3 0.16% 0.14% Correlation=-­‐0.23   0.12% 0.10% CTR 0.08% 0.06% 0.04% 0.02% 0.00% 1 10 100 1,000 10,000 100,000 Impressions (In Millions)                                                                                                                8 https://www.facebook.com/press/info.php?statistics9 Figure 3.5 is based on 4,202,841,334 ad impressions from earlier research in 2010 plus an additional8,662,942,836 in 2011  718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 18  
  19. 19. From 2010 to 2011 the cost of traffic increased as well. The average CPC of2011 was up 34.8% from $0.69 to $0.93. Despite the increased CPC, the averageCTR increased by 6%. There were few exceptions: • Retailers in 2011 paid less than half the CPC of what they paid in 2010. This can be explained by a 67% increase in CTR due to an 89% increase in total social impressions. This reflects the introduction of sponsored like ads, making it easier for brands to utilize FOF targeting. • The CPC for telecommunications companies decreased by 21%, despite a 7% decrease in CTR. Decreasing competition in this area led to fewer impressions, and less competition over ad inventory. • The CTR and CPC of restaurants remained the same between 2010 and 2011. Decreasing competition demonstrated by fewer impressions balanced the natural increase in the cost of traffic.   CTR 2011 2010 Figure 3.4 Overall Tabloids and Blogs Media and Entertainment E-commerce Travel Automotive Correlation  Between     Non-profit 3.4  and  3.5  =-­‐0.81   Cosmetics Retailers Restaurant Consumer Packaged Goods Misc. Local Services Financial Services Telecommunications Conference and Events Internet and Software Health Care 0.00% 0.02% 0.04% 0.06% 0.08% 0.10% 0.12% 0.14% 0.16% 0.18%718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 19  
  20. 20. CPC 2011 2010 Figure 3.5 Overall Health Care Internet and Software Telecommunications Financial Services Local Services Conference and Events Misc. Restaurant Retailers Consumer Packaged Goods Cosmetics Travel E-commerce Non-profit Automotive Media and Entertainment Tabloids and Blogs $0.00 $0.20 $0.40 $0.60 $0.80 $1.00 $1.20 $1.40Social Brands Fare BetterThough total impressions certainly played a part in changes to CTR and CPC inboth 2010 and 2011, the type of brand played a much larger role in the overallCTR and CPC. Social brands like retailers, restaurants, tabloids and blogs had ahigher CTR and lower CPC. At the other end of the spectrum, non-socialbrands such as Health Care providers and Telecommunications companiescontinued to perform poorly in comparison.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 20  
  21. 21. What Brands Are Saying“Year over year the amount of content available on Facebook has increaseddramatically. More brands are beginning to leverage this channel and while thenumber of consumers on Facebook continues to grow, it is going to get harderfor brands to attentively engage consumers. Here are some key pointshighlighted in this report.Successful brands need to focus on increasing interaction effectiveness withtheir consumer base. You do this by increasing the frequency of exposure,paying close attention to what is being discussed, and focused advertising.Interactions with consumers will occur as you post and more than 70% of theinteraction occurs within the first hour. So keep up your efforts on postingmore often and monitoring right after you post versus posting and “going tobed.” Also, a post with a “question” tends to drive increased interaction.Don’t forget that advertising does work. Your engagement rate can go up by21% to 43% by knowing what to say, when to say, and of course how you saythings on Facebook.Brands have a significant opportunity to leverage social. The key to success isto understand how much of the social data is public, how to collect the datawithout alienating the consumer, and how best to quickly use the data that iscollected.” Sundeep Kapur, Author – www.emailyogi.com, Digital Evangelist - NCR"This research by Dennis and his Blitzlocal team provides a measured view, andhelps makes sense, of what we already knew - we are drowning in data. It willtake the use of effective big data filtering platforms like Blitzmetrics to allowbrands to effectively sift through the data noise as well as successfully engagewith the right conversations and the right consumers." Lee Bogner, Consulting marketing technologist and social business analyst - @leebogner, Social Business Ambassadors and Board Director, Social Media Club LI. Reach718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 21  
  22. 22. “At Gordmans, we were able to capitalize on the low CPC in 2011 and run amultitude of ads throughout the year. We found that our paid advertising actsas a catalyst in driving additional organic impressions; because of this, we sawan increase in both our organic and paid performance last year.” Veronica Stecker, Manager of Converged Channel Advertising – GordmansAbout the Data Set:This report data spans a 7-month time period from June 1, 2011, to December31, 2011: • 11,365 pages • 5,744,550 admin and user posts • 76,202,654 Facebook users spread over 75 countries • 91,009,142,252 organic impressions • 27,969,233,010 paid impressions • 1,142,652,653 viral impressions. • 120,111,027,915 total impressionsLimitations: It is important to note that this study lacked data on applicationsand open graph data, which are increasing.On January 18, 2011 Facebook announced the Open Graph going live on theFacebook developers blog. In the coming months it will be interesting to seetheir effects on organic impressions and engagement. Industry results werebased on categorizing campaigns in our dataset, which means some industrieshad only a single brand and may not be representative of an industry’sperformance overall. Our analysis does not include premium ads, since theyare reported in separate systems.For further questions about this research, please emaildennis.yu@blitzlocal.com or message us at facebook.com/blitzlocalYou can find some of our data on organic and paid impressions here:https://blitzmetrics.com/dashboard/?account=wsj&page=wsjAbout BlitzLocalBlitzLocal is social analytics company founded in 2006 by Dennis Yu andheadquartered in Portland, Oregon with 32 people. Their product,BlitzMetrics, measures social for brands in retail, CPG, and network TV.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 22  
  23. 23. Contributing StaffElliot Swan Travis KingInteraction Designer Manager of Facebook MarketingJustin Dunlap Carrie MedinaDirector of Engineer EditorDennis Yu Ken PennyChief Executive Officer VP of OperationsGlossary of Terms:Coverage: Impressions per 1,000 fansEngagement: Page interactions divided by page fan count.Frequency: The number of times each user is seeing a post/page.Organic Impressions: Impressions in the news feed, ticker, and from fansvisiting a brands page.Paid Impressions: Number of ad impressions.CPC: Cost Per Click.CPM: Cost Per 1,000 Impressions.CTR: Clickthrough Rate of all ads.Social CTR: Clickthrough Rate of friend of fan ads.Social Impressions: Impressions for friend of fan ads.718 SW Alder St., Suite 200 www.blitzlocal.com www.facebook.com/blitzlocalPortland, OR 97205 23  

×