As promised, here is the analysis of the social media campaign used to help win Grand Rapids-based nonprofit Kids’ Food Basket a much-needed delivery truck from Toyota’s “100 Cars for Good” campaign.
To show all the public relations majors out there that “Management by Objective” isn’t just an esoteric concept you memorize in a PR 200 class and subsequently forget, I’ve framed the analysis of the campaign in terms of the “RACE” acronym (Research, Action, Communication, Evaluation).
KFB has rapidly become a favorite charity in West Michigan since its founding ten years ago. As a result of a decade of good work to feed school kids nutritious meals, they have a solid base of support not only in financial donors but almost more importantly in volunteers. They somehow manage to coordinate over 150 volunteers PER DAY who assemble the “sack suppers” that go out to local schools whose populations are made up of a majority of students who qualify for low-income food assistance.
As the contest required mobilizing sheer numbers of people to vote, social media was identified as an ideal platform to launch the campaign from: it’s inexpensive, widely-used by the publics that are advocates for KFB, and it affords the unique opportunity to change up strategies and tactics on the fly while the campaign was being executed.
The primary platforms used to execute the campaign were Facebook and Twitter given their broad adoption in the US. Fortunately KFB has a number of followers on Facebook which made them easy to target with updates throughout the day, which was broadened by the personal networks of all of the individuals associated with the campaign which was considerable.
For Twitter, in addition to the scheduled and conversational tweets, we identified lists of relevant individuals and organizations with large networks of followers in the hope that they would retweet our message to their followers. Relevance was determined by geographic region, but also context – like culinary arts and hunger. We used a variety of tools, including WeFollow (a Twitter directory) but Google searches and other analytics tools like Klout. It’s important to note that having a lot of followers does not in any way correlate with how influential a Twitter user is – and in the new world we live in, influence is far more important than sheer numbers:
- High profile chefs – particularly those involved with childhood nutrition (Jamie Oliver was our ideal RT because he’s directly involved in all of the issues that KFB addresses).
- Users in West Michigan / Grand Rapids identifed as influential.
- Individuals and organizations involved with hunger.
The full lists are available here if you’re interested (ooh spreadsheets – fun!).
Another variable we considered was the current environment and popular/topical users on Twitter at the moment so we attempted to incorporate the element of timeliness into the campaign.
Of course, no social media plan would be complete without incorporating the traditional media – so their social media presences were included as target publics.
The day of the contest, armed with the data we had collected, everyone on the teams fanned out and contacted as many of the contacts as possible using personalized appeals. One major concern I had with our approach that I hoped to counter with individualized appeals was being blacklisted/blocked/suspended for being labeled by too many people as a spammer. Fortunately we were able to avoid that pitfall.
- Ads: Scheduled to run from the moment the contest opened, to the moment before it closed (08/04/2011 6:00am – 08/04/2011 11:59pm), we identified several possible strategies and could have run multiple concurrent campaigns.
- Posts: Staff members, committee members and advocates fanned out and began posting messages as well as updates (including screen captures of the voting results throughout the day – which were only available for viewing at the moment of voting).
Twitter: Staff members, committee members and advocates fanned out and began tweeting messages and RT-ing each others’ messages throughout the day beginning as soon as the contest opened. Several used tools like Tweetdeck to schedule tweets to appear throughout the day in addition to the organic conversation that took place.
YouTube: Interns with Johnson Controls had created a series of three brief cute/humorous videos appealing for votes which were promoted both in advance of the campaign but most heavily the day of.
- Over $250.00 of unsolicited donations came in yesterday from as far-reaching places as Oklahoma, Hawaii, Pennsylvania, New York, Louisiana, Washington State, Ontario (international!).
- KFB gained 568 new Facebook followers that day.
- KFB picked up over 30 Twitter followers that day.
- Adrienne Wallace personally netted 45 Twitter followers and 14 friend requests from local community members on Facebook.
To help the social media component of the campaign, I used Bit.ly to create a link to the 100 Cars for Good page (bit.ly/votekfb)
- It received a total of 411 clicks the day of.
- The vast majority of the traffic (55 percent) was direct uses of the URL (Email Clients, IM, AIR Apps, and Direct), with 19 percent from Facebook, and 12 percent from Twitter.
- Although 90 percent of the visitors came from the US – ten percent came from outside the US, which is fantastic!
What’s really cool about bit.ly is that it catalogs ALL analytics for a particular URL regardless of whether or not it was the version you used, or the version others used. So one can see how much traffic they drove to a particular site compared to all other users of that same URL (as cached by Bit.ly).
This comes in handy because Toyota won’t publish data on the total number of votes for any of the winners – so we can use the 46,238 total clicks that bit.ly drove to the Cars for Good page to see how Kids Food Basket compared to other charities.
The 100 Cars for Good campaign kicked off with the first day of voting on May 9, 2011 (which recorded 1,250 clicks of the URL). There were two MONSTER days – June 1 saw 10,235 clicks of the URL (Comfort Zone Camp was the winner that day), and June 13 saw 11,486 clicks (SPCA Tampa Bay was the winner). If we remove those two outliers, it looks the median clicks per day is 266 – so Kids Food Basket drove almost double that amount of traffic using bit.ly.
Toyota graciously gave each organization $250 worth of Facebook Advertising credits to use to promote themselves during the contest. This was a good example of how insignificant “impressions” or “hits” are when one is gauging the effectiveness of a campaign.
- 1,207,260 total impressions
- 41,164 “social impressions” – defined as “The number of times your ad was shown with social context (i.e. with information about a viewer’s friend(s) who connected with your Page, Event, or App.”
- Out of that 1.2 million total impressions, only 260 actual clicks (10 of which were “social” clicks) resulted.
- For Kids Food Basket’s Facebook page, these are the latest analytics:
- 665 “Active Users” since last week
- 181 “Likes” since last week
- 504 “Wall Posts or Comments” since last week
- 1,221 “Visits” since last week
Twitter: As mentioned, in addition to constantly tweeting generic reminders and appeals to vote (virtually all of which were unique to decrease the liklihood that they would be written off as spam), we made appeals to the following individuals with sizable contingents of followers. Most were targeted based on their proxmity to Michigan (or possible interests in Grand Rapids/Michigan), involvement with childhood hunger, the culinary arts, nonprofits, or other related issues – but some were just for fun:
- Celebrities: @charliesheen, @ebertchicago, @MP_Gosselaar, @chefguyfieri, @IMKristenBell, @daxshepard1, @bflay, @pauladeen, @noreservations, @jamieoliver,
- Organizations: @sharestrength, @WeCanEndPoverty, @FeedingAmerica, @AkshayaPatraUSA, @ONECampaign
- Local Media: @grpress, @WOODTV, @wzzm13, and @FOX17
It’s very difficult to measure how much total traffic (in tweets and retweets) was generated as a result of the campaign throughout the day because it’s so difficult to search Twitter’s archives. What is more important, however, is the action generated as a result of that flurry of tweeting – which can be measured by the bit.ly analytics (ie how many people actually clicked on the link to go vote).
Of all of the “celebrities” we targeted, we were mostly unsuccessful with one major exception – Andrew Zimmern (Chef, Author, and host of Bizarre Foods on the Travel Channel) retweeted our appeal. That was a big win because he has over 234,000 followers and is very active and engaged.
YouTube: The videos uploaded to YouTube generated 355 views which is very positive given the attention span of the average person and the very limited timespan in which they were intended to promote the KFB campaign. The reality of content like this is that it takes time to build viewership – and it’s critical to YouTube’s recommendation algorithm that the videos are properly tagged and indexed.
I can say definitively that while the social media campaign was certainly helpful in generating traffic and votes, the analytics seem to show that the social media campaign was only responsible for a small portion of the total number of votes. If Toyota were to make the total number of votes and analytics available, we could do a better comparison.
What REALLY won the contest was the hard work of the KFB employees and volunteers over the past ten years which earned them the trust and respect of large networks of friends and associates who were willing to take action on August 4th to vote. There is no silver bullet when it comes to social media, despite what many snake oil peddlers will tell you – it’s all about relationships and good communication which can only happen over time.
Anecdotally, some supporters of KFB were unable to vote because one needed to not only have a Facebook profile to vote – but they also had to “like” the “100 Cars for Good” app which may have scared off some privacy-minded individuals given the sketchy array of options managed by Facebook. I received emails inquiring about other ways to vote, but unfortunately Facebook was the only way.
If we had more time, it’s likely we could have done a better campaign by adding a major component: promoting the link that would allow users to sign up and be reminded the day of the vote. As a result of the nature of the “100 Cars for Good” app – prying the actual URLs from it was tedious and difficult – so I didn’t uncover that this was an option until the day before the campaign. Had we been able to incorporate this component – we wouldn’t have had to rely on trying to reach people within a timeframe of a matter of hours (it’s important for us digerati to remember that not everyone logs in to social media daily like we do). This was an important lesson – make sure you have ALL of your assets allocated up front so you know what you’re working with.