This was originally published at the Rednod blog by Alistair Croll, Sarah Severson, and Alex Bowyer. Alistair Croll is the co-author along with Sean Power of Complete Web Monitoring recently published by O'Reilly.
Last week, we helped out our friends at Visible Government with their Beers for Canada campaign. In the end, the campaign raised just over $1,000 in two days; donations will help open government data to citizens and promote transparency in public offices. We learned a lot about what did and didn’t work, and in the interests of transparency, we thought we’d share some of the lessons we learned along the way (and see if we can collect some ideas for next time.)
How it worked
A week before Canada Day (July 1) we built and tested a simple site that encouraged donors to “buy their country a beer” — basically making a donation. We told a few key bloggers and Twitter personalities about it beforehand; then, on June 30, we started talking about it online. We continued to mention it, and amplified what others were saying, until midday on July 2.
From the outset, this was a short-term campaign built around a single day. We did this to give it urgency and purpose. We chose to start talking on June 30 because so many people were out the office (and away from their computers) on the holiday itself. But it’s important to realize the differences between a short-term campaign (minimal upfront work, strong word of mouth, modest goals, and real-time virality through Twitter) and a longer one. The timeframe also meant that most blog coverage only hit on July 1st (and thanks to all the bloggers who covered us!)
What worked? What didn’t? What would we have changed? Here’s a quick list.
While this was our first Twitter campaign, we did manage to get some things right. Here‘s what worked:
- We built analytics into the process. We used bit.ly (to track viral spread), Google Analytics (for goal conversions), Paypal audit accounts (to see donation amounts) and Clicky (for real-time web analytics.) Clicky is essential for short-term campaigns because it provides minute-by-minute visitor information, whereas most analytics tools only show traffic daily.
- We made the action obvious. We had one simple goal for people to accomplish on the donation site: donate. We even broke it into three different tiers (beer, pitcher, and round) to make it straightforward.
- We didn’t build it all ourselves. We used Paypal for donations; while it has its issues, it’s also a well-known and trusted brand, and we seem respectable by association. We also used free services like Google Groups and Clicky. This means we didn’t need to code too much.
- We set up tracking with hashtags and keyword searches. This meant we could watch the activity online and amplify it or respond to questions.
- We had plenty of ways for people to reach us. We had links to the Visible Government website, and generated enquiries there. We also linked to the Google Group discussion, which added new members and triggered conversations.
- We had a great cause. The simple fact is that without a decent motive, you won’t have much success. People felt they were doing their civic duty by mentioning us, which helped spread. If your cause isn’t just, people will feel icky promoting it.
- We tested it a lot. Even though we didn’t find every mistake, the launch was surprisingly smooth because we verified it properly and used real infrastructure (from our friends at Syntenic.)
- We had a simple, catchy message. “Buy your country a beer” was strangely patriotic, and people liked it. Made To Stick is the bible for clear, simple messages. Early on in the design process, we were tempted to overload the message–something like, “Buy your country a beer and promote open interactions between federal government and Canadian citizens.” That wouldn’t have worked because it wasn’t simple. But “buy your country a beer” is intriguing. Remember that the tagline’s purpose is to provoke interest. Once you’ve got someone’s attention you can do things with it.
- Set up Reddit, Digg, and other social news aggregators. We put badges on the Beers For Canada website encouraging people to Digg us and promote us on other social news aggregators. This made it easy for people to support us and spread the word.
- We set the right kinds of goals up front. How do you know you won if you don’t know where the finish line is? One of the first things we did was set goals for the campaign. We wanted to see donations, of course, but we also wanted to see unique visits to the Beers for Canada site and how many went further to the Visible Government site. When we started we had no idea how the campaign would do so we focused less on numbers (500$ or 5,000 site visits) and more on what we wanted to achieve (visibility and engagement.)
- We used calendar meetings to remind promoters. This was a neat trick. When we asked people to mention us online, we sent them a calendar invite as a reminder. This way we knew when they’d do it, and since most of the people we asked had an iPhone or a Blackberry, they could do it from wherever they were–particularly important on a holiday (though as you’ll see below, in hindsight we could have spread those out more over a longer period of time.)
What did we learn?
Here are some of the lessons we’ve learned, and the things we’d have done differently.
Beforehand, in the planning phase:
- A short timeframe limits others’ ability to build online context about you. When you’re running a fundraiser, people want context. It’s a catch-22: If you do something quick and spontaneous, you’ll build excitement and mystery, but you won’t have the time to inform bloggers and the press about what you’re doing far enough in advance for them to provide details and perspectives. If you tell bloggers too soon, you lose the excitement.
- Plan out your whole message before you send the first tweet.We carefully crafted website copy but didn’t think enough about who would tweet what, when. In a real-time campaign, your copywriting isn’t done when you publish the site. It’s constant, and it needs to be planned.
- Schedule things, and have a single coordinator for the life of the campaign. At noon on June 30th, one of us put out our first tweet–and forgot to use the bit.ly URL that would track the spread of the campaign. This would have been avoided by having an initial schedule, and then having a single person adjust that schedule as things progressed and feedback came back from the analytics tools and the campaign. You simply can’t assume that ’someone’ will do it.
- Be transparent and obvious. Make sure the people affiliated with the campaign are clearly identified. I was personally thanking a lot of our supporters but my connection to either the campaign or Visible Government was not clear since it was coming from my personal account. Not only does this keep your campaign transparent it help you build you reputation and social capital making it more likely you will get those people back for a donation. One possibility would have been to temporarily change our avatars to include a visual cue–like the Visible Government maple leaf–for all those officially behind the campaign.
- Have a clear call to action. The website was pretty blunt about donations. We set it up, then told the world. What we quickly realized was that the Tweets themselves–not just the website–needed to be clear what we were asking people to do. Were we asking people just to tell their friends? To donate money? To watch the hashtag? To visit the site and learn more? In Twitter’s 140 characters, there’s only room for one call to action. You need to tell people what to do and make it easy for them to do it.
- Facebook is for slow burn, Twitter is for ADD. Twitter’s like speed dating: you see something, and quickly decide if you want more. By contrast, Facebook favors a groundswell of support: as more and more of your friends like something, you do too. The duration of your campaign affects which social networks you’ll rely on. We shouldn’t have wasted time on Facebook for a campaign of this duration.
- Define analytics goals better. We didn’t take the time to implement goal funnels within the system, which was a shame. What’s more, referral URLs are useless in a world where many Twitter users rely on Tweetdeck, Seesmic Desktop, or the Twitter client on their Blackberry or iPhone. To address this, we should have segmented shortened URLs using Google’s URL builder to inject metadata into the shortened URLs so we’d get a better idea of visitor source.
During the campaign:
- Personal claims of action work best. Megabloggers like Tim O’Reilly, Om Malik, Austin Hill, Michael Geist, Tara Hunt, Mathew Ingram and others generated a ton of traffic and awareness. But the messages that generated the most donations–rather than just visits–were those where the RT testified to an action. Someone who said “I just bought a round - you should too” generated far more actual donations than someone who just said, “check this out”.
- Have an FAQ–and update it. We drafted an initial FAQ that had lots of information in it, as well as links to Visible Government. We were able to direct people here if they had questions. But we were missing certain pieces of information (for example, why donations weren’t tax deductible) and took too long to respond to questions and update the FAQ.
- Vary the message. Tweets about hashtag visualizations showing campaign growth, mentioning who was blogging about us, and retweeting others all kept the dialogue going, but they were done ad hoc and should have been better planned.
- You only get one chance to make an impression. We live in an information-starved world. People will only click on a link once unless they think there’s new news. So if your first message says, “check this out,” they will. If after that you say, “donate to this cause” they’re less likely to: they’ve already seen it. Only when there’s new information–”50 people have bought their country a beer”–will the audience consider revisiting things.
- Make the site interactive. If we’d provided people with somewhere to comment or share their thoughts–or even to suggest how the donations should be used–we’d have had more raw material for the campaign and could play back these comments to the online community that was discussing it. This also gives people a reason to check back and see how the discussion is progressing. Again, with a 36-hour campaign, this may be a lot more effort than you’re willing to expend, but we might have been able to use a Subreddit or some other already-built system.
- Spread your messages over time. Lots of people agreed to help spread the message, but it happened all at once and the initial message quickly lost traction. It would have been far more effective to have one person mention us, then let the second person tell the world all the great things that happened after the first mention, and so on. By firing all of our guns at once, we didn’t let the message “snowball” and build on existing momentum. A campaign like this needs lots of ’seeds’ to get the message out.
- Give donors a way to tell others automatically. We made it possible for people to tweet the site from a link on the site. But we should have had an option, selected by default, that made a tweet saying, “I just bought the country a beer and you can too.” This should have included a different shortened URL or analytics link, so we could differentiate first-visit traffic from viral donor traffic.
- Respond in person. You can’t plan for everything so make sure you are ready to answer any questions both publicly and promptly. Also, thank people for their donations — but respect their privacy; if you can thank them through direct messages, great. If they made a sizeable donation, you can acknowledge it by saying, “someone just donated $100″ (or in our case, “someone just bought the country a round.”) Don’t single out donors publicly as they may not want the attention.
- Keep people updated. If you’re tracking donations, tell people about the progress. Celebrate big donations or interesting blogs. The more you can show people that others are doing things, the more engaged they’ll be. Appeal to their inner lemming. We could have build a dashboard for statistics (donations, reddit ranking, retweet count, page views, etc.) We did discuss the amount of transparency we wanted (which is ironic for a transparent government initiative.) The real dilemma here is that you need to wait until the news is newsworthy. If we’d said, “hey, we have a total of $14 donated!” people would have discounted the success of the campaign.
After the campaign:
- Have a next step. There’s a lot of positive sentiment about Visible Government now. We have some great ideas for how to use the money, including the forthcoming Code for Canada contest and an initiative to get computer science students to develop transparency applications. It’d be great if we had this ready to discuss when the campaign ended, because it would allow us to continue and amplify the engagement that the campaign generated. Plus, it’d let people feel good about what they’ve done. In other words, every campaign is part of a bigger picture of long-term connection with donors, markets, and audiences.
Even though we didn’t focus on the numbers too much this time around, we still set some goals so we’d know what we were measuring. Not only did this give us a measure of success it helped evaluate the experience as a whole and focus us to come up with these lessons. We could clearly look at graphs and numbers and say “Yup. Nobody talked about us for over 4 hours,” and then wonder why.
- Viral spread versus megablogger attention. This campaign was promoted almost entirely on Twitter and using our personal and professional networks to spread the word. We were fortunate enough to have some really influential people blog and tweet about it. But we didn’t see the viral growth among others’ networks that we’d have liked.
- Conversion funnels and donations. Though tens of thousands of people read the tweets (these people have over a million followers collectively!), we only saw 1,642total visits, but that translated to about $1,000 in donations. Conversion rates were less than 0.2%, which we attribute in part to the passive message we used at first. In other words, the tone of the campaign emphasized attention (”visit this page”) over conversion (”please donate”).
- Attention generated. Our bounce rate — the number of people who saw one page, then left — was only 51%, which is great: over 25% of visitors wanted to learn more about the campaign. What’s more, Visible Government saw a huge spike in attention. Compared to the previous week traffic spiked by 300%! We also have several conversations with the press underway as a result of the campaign.
In the end this was a quick-and-dirty campaign that raised some well-deserved money and got good visibility on a national scale. Along the way, we learned a lot about campaigning in a digital world, particularly one based on real-time word of mouth.
Now we want to hear from you. What’s worked for you before? What else should we consider for next time? What did we do wrong?
[Disclosure: Rednod’s Alistair Croll is on the board of directors of Visible Government]