Monthly Archives: August 2014

5 things you can learn from Twitter’s new analytics

twitter dash 4

Whether you’re a brand curious about how your tweets are faring or an individual longing to gaze into your social-media navel, Twitter’s opening of its analytics dashboard this week is sure to excite you. Twitter first launched Google-esque analytics to its advertisers and verified users back in July, but now, anyone can go to the dashboard site to see how far and wide your tweets go.

For users new to Twitter analytics, the results are surprising:

  • First, a negative: Fewer people are exposed to your tweets than you might think. A Twitter account with about 10,000 followers will, on average, have each tweet seen by only 300 to 500 people. This intuitively makes sense, since obviously not everyone is using Twitter at all hours of the day, and your 9 a.m. missive about coffee is buried in the stream by the time another user logs in an hour later. Still, Twitter reach is less than expected.
  • Second, a positive: “Engagement rates” are through the roof. Twitter defines “engagement” as a user taking any action on your tweet, such as clicking on a link, favoriting it, retweeting it, or replying. While the standard interaction rate on normal banner ads is about 0.07%, Twitter interaction rates hover around 2.5%. This appears to be a decline from a few years ago (in 2011, Twitter boasted on its blog that brand interaction rates were 3-5%), but 2.5% is still amazing — users exposed to your tweet are basically 36 times more likely to take an action than if they saw a digital banner ad.
  • Third, click-through rates are also very high. Using my personal account @benkunz as an example, from Aug. 1-28 I sent out 109 tweets, of which 34 had links. The tweets with links generated a 1.70% click-through rate based on 330 clicks against 19,380 actual impressions. More than two-thirds of my Twitter interactions were clicks, and that CTR is 24 times higher than the average banner ad response rate. The caveat here, of course, is I’m not a brand trying to sell you anything, just a human trying to be interesting, but that response rate bodes well for any brand with an authentic, meaningful content stream on Twitter.
  • Fourth, if you want your tweets shared: News tied to the zeitgeist of Twitter chatter works. In our non-scientific-but-super-insightful study of what tweets have the highest interaction rates, missives related to current events usually have 2x or 3x the average Twitter interaction rate. In the past few weeks, our tweets about Matt Damon doing something innovative with the ice bucket challenge, Starbucks redesigning its stores, or Robin Williams being photographed as a mime in Central Park in 1975 had much higher interaction rates that our supposedly smart asides on life, liberty and the pursuit of advertising. News, on Twitter, sells.
  • Fifth, timing matters. We found the highest interaction rates on Wednesdays, Fridays and Saturdays, and the lowest on Sundays and Mondays. Apparently Twitter users back away from social media on Sundays and are too buried in work on Mondays to spend much time interacting.

Wisely, or perhaps sadly, Twitter only allows you to see analytics on your own  tweets. As of now, you can’t view the Twitter activity of Pepsi or Robert Scoble to parse how big brands or social-media celebrities are faring with their own Twitter messages. But if you want to examine your own tweets to see what makes people respond, the dashboard is an excellent resource.

Of course, the other option is you could continue to just be yourself online and share real, authentic insights and news not worrying about what other people think. Brands and humans, we’ll leave that up to you.

Why can’t advertisers get personalization right?

girl heart eye

 

Deep near the bottom of today’s NYTimes story on Facebook — “How Facebook Sold You Krill Oil” — a marketing manager for Reckitt Benckiser, a company that sells fish-oil pills, says “Facebook is a fantastic tool for doing personalized marketing at scale.” The NYT case study goes on to explain how the advertiser in question is able to pinpoint-target different female demographics, users of other fish-oil products, and even “lookalike” profiles of individuals with similar interests. The Facebook campaign worked and fish-oil sales went up.

Yet all this is outbound targeting from the marketer’s perspective, similar to really good archers being able to fire arrows into the chests of different consumers with different needs in a crowd. It’s not personalization.

True one-to-one personalization was described by Don Peppers in the 1990s as an iterative process in which consumers are identified, differentiated on both their financial value to a company and need from a company, interacted with, and then given customized services or communications. The most crucial step was to make all of this a feedback loop, a “learning relationship” in which a marketer grows ever more sophisticated about anticipating individual needs over time.

It is this anticipation of needs that creates loyalty, Peppers said, because once a consumer has trained one organization to anticipate her wants, she would face switching costs going somewhere else. The classic example of marketers who actually pull this off are local coffee shops who see you walk in and immediately warm your specific version of coffee and breakfast muffin without you asking. You skip the line, you save 3 minutes, you’re individually recognized, and so you don’t want to go anywhere else. One-to-one personalization, evolved as a learning relationship, becomes almost like a strong marriage — where all the prior history of learning enhances the bond between consumer and brand.

Now, let’s revisit advertising.

Today, I checked my ads for Facebook, and saw this: retargeting ads from a hotel chain I searched two weeks ago; a technology course apparently inspired by an online friend who works in Web services; a tiny chip-wafer thing I can attached to objects around the house to “never lose them again”; and an ad for a local plumber. Personalization score, from 1 to 100: About a 5. At one point, I was interested in the hotel.

Twitter was even worse. I got two ads for a Red Bull music festival; a local Applebee’s ad for a frozen fruit drink; and a promotion for small-business insurance. Personalization score: 3. Maybe, just maybe, we’ll go out tonight to a restaurant, but I don’t like frozen fruit drinks.

All of these ads express a knowledge of my personal needs about as sophisticated as the  direct mail lists that trigger our Pottery Barn catalogs. I am a target with some general color descriptors: So, go ahead, fire your arrow into me.

Product-focus creates a personalization failure

The problem that personalization faces is most marketing engines base it from the marketer’s point of view (since the guy with the ad budget is calling the shots). This creates a data collection model centered on a product, which inherently has vast gaps. A home services company will build a CRM system to include lots of information about an individual prospect’s interactions and a customer’s service records, but that data is only germane to the product. Jane Smith may have a forecast lifetime value of $20,000 in utility bills over 10 years … but the data systems don’t recognize that she is a mother with three daughters interested in mountain climbing and kayaking. The utility frankly doesn’t care.

All of this is driven by economics: individual companies want to store and analyze only data related to their product sales; social networks want to release only the data needed to target a product to a customer; and a truly customer-centric personalization campaign would require coordinating millions of potential product offers, likely from competing brands who have little incentive to sell services outside their own scope.

In simple terms, because the varied needs of a customer would require unified data and services that cut across brands, until an ecosystem of brands has an incentive to share data and revenue, personalization will not happen. What this gives us is ads on Facebook for fish-oil pills based on a rudimentary understanding that you are friends with someone who takes vitamins, but no anticipatory personalization that informs you of where to take your wife to dinner on the anniversary of your first college date.

The customer loses out

What would it take for true personalization to arrive? A few companies may come close. Amazon and Walmart, which house millions of product SKUs, have incentive to use data to anticipate your needs and the service offerings to potentially benefit from a vast range in sales. Twitter, which conceivably could parse your real conversations to build better profiles of your mind than simple Facebook stated interests, could truly personalize its #discover newsfeed to create content germane to your interests.

Facebook might do true personalization, if it could somehow depress ad offers that had nothing to do with your interests. Unfortunately, Facebook wants the billions from the unexpected plumber ad or computer technology course offering that sprays you despite your disinterest. Most Facebook ads are priced on a cost per click, rewarding Facebook if a user clicks on the ad, but not worrying about the 99.97% of users exposed who don’t want to respond. The economic model focuses on what advertisers catch, and ignores the adverse impact of what they spill.

Television is in the same boat. The typical U.S. consumer receives 6,600 spots per month, based on 4 hours and 34 minutes of TV viewing per day. You may want 3 of those products, putting your “response rate” in the 0.05% range, about the same as responses to online banner ads. If TV truly personalized ads, it might have to forego 99.95% of its advertising inventory — and lose billions of dollars a year by removing all that unwanted communications bloat.

Truth is, personalization will never arrive until a giant platform is able to match transparent data on individual needs with a vast consortium of products and services that can be personalized in promotions without friction from marketers demanding that their product appear next in line. Marketing would have to become a true marketplace, and product-makers would have to cede their product focus to a willingness to give consumers what they really want — even if that means the brand down the street wins instead of them.

Which is why, tomorrow, you’ll see more unpersonalized ads on Facebook for fish oil. Today’s economic incentives decree that unpersonalized ads are the way to go.

Photo credit: Ruurmo