June 29, 2012

NEXT POST
Facebook's Bogus Sales Pitch Facebook has to sell a lot of ads to satisfy Wall Street. Unfortunately the latest sales blitz aimed at brands and agencies is lame. Facebook’s huge user numbers, nimble targeting using 15 possible variables and average 80 cent CPC are enticing. Facebook had 151 million unique US visitors in April 2012. Facebook attracts 7 out of 10 Americans online. The average person had 28 sessions for 379 minutes during the month, basically once a day, accounting for an average time spent of 7:09 hours. Although, with a 0.04% CTR, 4 out of 5 users say they’ve never been influenced by an ad on Facebook. The initial attraction gets stopped dead when you realize that the system has been purposefully constructed and rigged to harvest ad dollars not necessarily to facilitate brand communication. Like everything else Facebook does, they know better. Their mission is to reinvent advertising. Brands spent money, big money in some cases, to accumulate large fan bases. Then Facebook limited access to those fans (and their Newsfeeds) by inserting the Edgerank algorithm to filter content. Estimates are that only 7-16% percent of brand communications actually reach fans’ newsfeeds. What’s worse, according to PageLever, as the fan base increases, the number of impressions created by distributed brand content decreases! The brands that most effectively used Facebook to build audiences get the shaft. Facebook hijacked the fans that brands paid to accumulate. Now Facebook will sell access to fans and guarantee reach levels using a new series of ad units, like Sponsored Stories, based on a new Reach Generator tool. It’s a closed coercive cash-generating circle. Great for Wall Street but not so helpful to Madison Avenue. Lauren Fisher, writing on the Simply Zesty blog, observed, “By squeezing more and more from brands, they’re making the platform an unattractive option for social marketing from a cost basis if nothing else.” Most large brands on Facebook are trying to understand whom they attracted and if there is a relationship between fans and customers. Since Facebook won’t share the data that brands helped generate, brands are left to either guess, test new 3rd party eCRM tools or rely on Facebook’s vague representations to dissect their fan bases. Most brands know that fans are much more likely to engage and probably buy than the average Joe. But most also suspect those fans are multi-channel customers and that they may already be reaching and persuading them through e-mail, branded websites, loyalty programs and general advertising. Since nobody has convincingly studied the impact of Facebook engagement in isolation, savvy marketers cannot determine the relative or incremental marketing value of Facebook. Some suspect they are duplicating efforts and essentially preaching to the choir. To try to make a strong sales case, Facebook collaborated with ComScore to create a series of reports titled “The Power of Like.” The latest version merchandises Facebook’s Like button and new ad units as ways to reach friends of friends and achieve stronger reach with the implied endorsement of people you know. Calling this phenomenon “amplification” the report scored different brands on their relative success at virility. “Most leading brands on Facebook achieve a monthly earned Amplification Ratio of between 0.5 and 2.0 meaning they extend the reach of their earned media exposure … by 50-200%.” The implication is that exposure to Facebook’s premium ads had something to do with this. The report goes on to cite specific, though specious, cases of successful brands activating consumers on Facebook. The first case focuses on Amazon, Best Buy, Target and Wal*Mart during Holiday 2011. Citing these leading brands as active in social media and offering special deals via Facebook, the results “showed significant higher rates of purchase among fans” and modest increases in spending among friends of fans. But this ignores several key factors, which could easily inflate these results. These are well known, heavily advertised brands Awareness, preference and purchase intent already existed Most focus on multi-channel customers, who are known to purchase more and more frequently based on RFM models There is no source data on purchases Exposure to Facebook ads was not isolated During the holiday period each brand was blasting away with offers and messages in many channels. It strains credibility to attribute the results exclusively to Facebook ads. Moreover we don’t know whether purchase data was self reported or based on shared data sets. It’s very easy to doubt the implied correlation or causality, which might actually have just been coincidence. For Starbucks, there was 4 weeks of exposure...
PREVIOUS POST
Buying Guide to Social Listening Tools Each year I survey social listening tools seeking to find a better mousetrap. This year, with the able assistance of Rachel Hong, we’ve plunged into the world of machine learning, word proximity algorithms and blathering sales pitches to yield five observations. They are Black Boxes. Nobody will show you the math and nobody makes claims of accuracy or access to historical data that can be independently verified. While particularly frustrating to math guys, this posture obscures competitive differences rather than illuminates them. No vendor has a toolset even closely resembling the NSA or the accuracy and AI possessed by intelligence agency data mining tools. Most sales reps can’t answer the second question, “How does it work?” The Same with Different Monikers. Some call it “natural language processing” others prefer “machine learning” still others call it “textual analysis” or “semantic analysis” but in the end it’s all the same. Each vendor has a word proximity algorithm that assigns a positive, negative and neutral value to words. Some claim to have databases of idiomatic terms and others claim to enable brands to add their own idiomatic dictionary or industry jargon to the search parameters. Basically these tools score sentiment based on the words closest to the target keywords. A great example is “Brand X is the sh*t.” Within 5 words of the brand name a negative term appears. All the tools score this as negative sentiment. Yet anyone who lives in our culture knows that it is the highest possible praise from a significant youth demographic segment. The good news is that they all err the same way. The bad news is you don’t get a full, clear picture of what’s being said or meant. They Don’t Reveal Insight. Most tools return bell curve results. They show smallish numbers of positive sentiment and smallish numbers of negative sentiment but display a huge amount of neutral (non-score able) sentiment. This is the real limitation of these tools, they don’t really tell you what’s going on. They suggest a direction that becomes the starting point for live, human analysts digging into the posts, tweets, comments and chat items to find out what’s really being said about your brand, who’s talking and to apply a human filter to the words, attitudes or sentiment they discover. Influencers are a Chimera. Many tools pretend to reveal key influencers in the social arena. Most count the volume of posts or the number of friends or followers or both. On the basis of volume, they designate individuals as “influencers.” But we all know that volume and influence are two totally different concepts. The mouth reaching the most ears or blabbing more often may not be the most persuasive and visa versa. In many cases influencers are bloggers who have a volume of either followers or posts. But in most cases there is no real way to understand how influence is defined, counted, tracked or how it changes over time. The ability to identify genuine influencers is an illusion perpetrated by Klout, Kred and others. But it’s as phony as a three-dollar bill. Since none of these tools can legitimately track downstream activity nor can they penetrate some of the closed systems (e.g. Facebook) where influence is accumulated and wielded. And since no one will share their “secret” formula or have a credible third party verify their methodology, there is no reason to trust or believe them. But, as we used to say about the Nielsen Ratings, they are the only or best tools we have. So to make optimal use of these evolving social listening tools and to decide which works best for your brand set your buying priorities in terms of needs and usage. If the math and the mechanisms are roughly the same, determine which one works best on these 3 criteria. Staff Resources. Do you have analysts dedicated to this social listening and sentiment mining? If so, ease of use is a secondary consideration. A more complex filtering tool might work best. If not, and civilians will have to make use of the tool, ease of use is the primary consideration. Dashboard/Graphics. Many of these tools tout their graphics packages, which produce a wide array of colorful charts and graphs that can easily become a dashboard. If you have designers, artists or PowerPoint wizards on your team, this doesn’t matter. If you don’t this can save you time and give you a very professional look regardless of data quality. Price. If you have a serious investment...

Danny Flamberg

I am a veteran marketing consultant working with leading and emerging brands

The Typepad Team

Recent Comments