Building Advocacy at Scale

Never have word-of-mouth and peer-to-peer influence been more ubiquitous or powerful. Billions of individuals now have the means to share their experiences, opinions and ideas – and to organize for action – at scale. They are revealing and gen­erating unprecedented amounts of data about themselves. This has profound implications for Chief Communications Officers and their teams. They must become adept at extracting actionable insights from what many are calling the “Big Data” era.[1]

How does an enterprise best engage individuals, in addition to audiences, publics or segments of populations? If the goal is not merely to shape the opinion, sentiment and perception of those individuals, how do you spur them to action, continuing behavior and advocacy?

One organization that answered those questions is the Obama for America 2012 campaign. Rayid Ghani served as “chief scientist” leading the presidential campaign’s data analytics team.

The following excerpt is from Rayid Ghani’s presentation to the 4th National Summit on Strategic Communications ( on April 23, 2013 in Washington DC.

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

Something we struggled with quite a bit was that it seemed everybody in the campaign was a 20-year-old.

I think the average age was probably 24.  A lot of those people think everything happens in the online world; that digital is everything and nobody takes any action offline.  Well, that’s not true, and especially not true in an election where the most important action is voting. You can’t take that online.  You have to take a physical action.

Voter registration is mostly an offline action.  In most of the states that mattered, you have to print the form, sign it and physically go mail it.  You can’t do a lot of these things online. You should be able to do them online, you can’t.

So a lot of the work that we did was try to figure out how to use digital tools to really affect offline actions?  How do we get people to engage with us online, and not just so that they can “like” us on Facebook (that’s great for them, not very useful for us) or re-tweet something that we’re telling them or start clicking on our Web site and getting our Google analytics numbers up.  That wasn’t the goal.

The goal was to build digital tools because that’s where a lot of younger people spend time. They don’t answer phone calls. They don’t have landlines. They don’t open their doors. They don’t watch live TV.  They don’t open emails. So you can’t get them through any other channel.  They are sitting on a lot of social media channels and they’re used to only taking online action.  So how do you get them to really take offline action?

We didn’t have a way of doing that online, the same way that we were really good at doing field work, where we could get people’s neighbors to go and talk to people and convince them, persuade them to vote and persuade them to vote for us.

So we tried to connect the two worlds.

One of the tools we built was called “Targeted Sharing.” One of the things that we learned from the field work was that the best person to persuade you is not us as an organization. The best people to persuade you are people like you! So we built a large number of neighborhood teams.

We also built an online tool that was based on a Facebook app. The idea was that people would come into our Web site looking at some content. Let’s say the content is designed for persuasion of voter registration and instead of the usual “share this on Facebook” prompt – which is not very effective for most organizations – we asked them to share our content with a specific set of people.

Now, I could give you a list of 500 people and say, “Share this content with the people who live in swing states that influence those who might not be registered to vote.” And you would look at that list and probably go away and not do anything at all.

So, to reduce that barrier, we would individuals for Facebook authorization to allow us to access their friend list and their social graph.  We would look at their friend list and match their friends to our list of voters. Once we had that match, we could do a few different things. We could get their support score and their persuasion score and their turnout score. In other words, we could identify the likelihood that the friend would support us, could be persuaded to supports us the likelihood they would actually vote.

We couldn’t match everyone because the Facebook permissions that we were asking for didn’t allow us to get a lot of data. We would get a friend’s name, location by city, state, and maybe month and day of birth; very simple things. So we couldn’t match somebody to a list of voters because there are lots of John Smith’s in Chicago, Illinois, or example.  So what we would do instead is use Facebook data to build voter scores.

So at some level we could get an idea off your friends, who are the best people to target for voter registration or for persuasion.  And then we would infer your influence on your friends.  So if you have 200 friends you might not interact with 90 of them at all, and another 50 you used to talk to but don’t anymore, and the remaining 60 you may have influence on.

So we would estimate your influence based on the interactions between you and your friends on Facebook and then show you a list of ten friends who are the best targets for you to contact on our behalf and can be influenced by you. We used that for getting people to register to vote, for getting people to persuade friends to vote with us, for turnaround votes, for volunteer recruiting, for event recruiting. For all of those things we could figure out who your best friends are – who the best people are for each specific campaign goal that we have and who can you influence.

That allowed us to increase our reach among young people, and for a campaign, under 35 is young.  That enabled us to get people to take very low-touch actions. You’re not asking for a lot! You’re asking for a couple of clicks to send a message to friends saying, “Hey, you should go and register to vote. Tomorrow’s the deadline, or you should go here and watch this video.”

We reached out to about a million people using that Targeted Sharing tool, and a million is not a large number when you think about the scope of the campaign. But a million people allowed us to access 200 million of their friends, and that reach was really critical.

The million is great, but that million allows you to pick and choose any of the 200 million people you want to target and tell this guy, “Hey, can you please ask ten friends?  And every other day we would have a different kind of “ask.” So when registration was ending in Ohio, for example, we could send a message to people saying, “Here are seven of your friends in Ohio who may not be registered. Can you please ask them to register to vote?”

So, we were able to build digital tools designed to get people to take action offline.  We measured our conversions and our outcomes; not what the tools were doing online, but how well they were able to translate into offline actions.

# # #


[1] From “Building Belief: A New Model For Activating Corporate Character and Authentic Advocacy,” The Arthur W. Page Society,