Zinc Research president was interviewed for an article published today by American market researcher Tom H. C. Anderson – who, like Singh, is also a well-known social networking evangelist. The complete article follows:
With November 11th around the corner, and the 2012 election just a year away, those interested in the intersection of research, social media and politics should find today’s post interesting and timely. Recently when speaking at an MRIAconference in Canada I was fortunate enough to meet Brian Singh, Managing Director of ZINC Research. I found we share similar backgrounds and passions, including leveraging our marketing research and social media savvy to supportpolitical campaigns.
Beyond political polling so often mentioned in the media, those of us involved in Next Gen Market Research understand better than most how traditional research techniques such as segmentation can be combined with new techniques in social media to go beyond measurement and truly impact and influence behavior.
Today I talk to Brian about his 2010 experience helping underdog Naheed Nenshi move from just 8% support in September (well behind leaders, Alderman Ric Mclver with 42% support and Barb Higgins with 28% support), to become Mayor of Calgary by gaining 40% of the vote primarily by leveraging social media to engage voters. No small feat, especially in the conservative Calgary, Nenshi became the first Muslim mayor of a major Canadian city.
Tom: You and I have a lot in common, we’re both researchers, we both have an academic background in Economics, and both have an interest in the intersection of politics and social media.
I know you recently worked on a rather big Canadian campaign. Can you describe for us a bit how you got involved?
Brian: Yes, it was Naheed Nenshi’s successful run to become Calgary’s Mayor. It was in the National news – he was the first Muslim Mayor in any city in North America. I would like to add that his faith was actually incidental – he really was the best candidate suited for the needs of Calgary. I am actually involved in another campaign right now for a party leadership race.
How did I get involved? Well, I knew these people. In Naheed’s case, we went back a ways and met through a mutual friend and stayed in touch, given our respective interest in politics. He knew that I did public opinion research, but he was also well aware of the work I was doing in social media research. The point of my involvement was oddly enough during the 2010 Olympics when we were in line at Japa Dog in Vancouver, when he heard that the previous Mayor, was not running for a fourth term. Once he got off the phone, he turned to me and said I wonder if I should run for Mayor – you are on the team right? At that point, I had never been involved in a political campaign, and I did have trepidations as in our business we live and die by our objectivity. However, Naheed is a force of nature and does not take no for an answer and eventually the discussion came to being part of the team and him emphasizing that I knew the stuff as well, if not better, than most and this was our opportunity to put my money where my mouth was. For me, it was a wonderful experiment in proving that the advice that I provided to clients actually does work! Thank goodness it did.
Tom: I think it’s amazing that there aren’t more researchers who are asked to give advice. On the political campaigns I have been asked to manage social media for here in the US, we sometimes let the marketing research take a back seat, but segmentation and positioning certainly are key. Still there often isn’t much time and money to do traditional research so we focus a bit more on Just In Time social media marketing. Was this the case for you in your work as well or did you also help via research?
Brian: In our case, we didn’t have the money initially but we did have the time. Naheed had a previous run at office, but given his role in a civic engagement process, was mentally prepared for the campaign. With other team members initiating actions on Facebook and Twitter and amongst Calgary bloggers, it was very much a grass roots campaign where people were observing and engaging candidates online. Thus there was a wealth of data available to us in terms of who was following Naheed as well as how the network was being built. I will add there were members of this team that were highly talented in executing marketing campaigns and I will always attest that the candidate did know what he wanted to do and the team that he had in mind and all of us were allowed to work independently to do what we had to do to help him fulfill his objective of becoming Mayor. There was a lot of high price talent who volunteered their time who were very in-synch with the business concept of a campaign.
In terms of polling and research, I would say that we were fortunate that I sat on 8 previous tracking waves of social media use and adoption in Canada. We also had a segmentation scheme and we had cross tabulated the segmentation scheme by political party. One of the interesting things that we noted was that one of the candidates fell into a particular camp that didn’t seem to be engaged online and that became part of the initial thrust of the campaign to own the online space and to hope that the network effects offered by social media and other forms of in-person engagement kicked in.
The final element was that as part of my campaign contribution, I included a number of questions on one of our omnibus surveys that proved quite useful in understanding the dynamics of the campaign and where Naheed stood from a brand position relative to other candidates. We were also able to determine a lot of efficiencies which were valuable to the strategy team as well as shaping and scheduling campaign tactics that helped our team stay ahead of all of the competitors and enabled us to build a brand that resonated with the Public and to deliver a successful result on election day.
Tom: The “high priced talent” you mentioned, I assume you’re alluding partly to political consultants. How savvy do you think they are in leveraging social media. I assume it’s similar to what I’ve seen on this side of the border and across many industries including advertising. A lot of consultants are quick to bandy the lingo about, but have relatively little experience in actually driving traffic and engagement in a meaningful way?
Brian: No. The high priced talent I mentioned were not political consultants. In fact, there were only a few members of the team that had campaigning experience and limited at that! The talent I talk about comes from traditional areas of business – one of Naheed’s handlers was a former eBay executive who had moved back to Calgary. Another was a lawyer, in the midst of maternity leave, was responding to the volumes of emails coming in on a daily basis. Yes, a well-established lawyer answering e-mails voluntarily and for free! Another was a well-known and well-connected real estate agent who had run for political office and is a well-known community organizer. A final example was another woman who ran a consulting practice that helped wealthy donors make informed decisions about charitable giving. Thus, this high-price talent came in a non-traditional form. It came from people who cared deeply about the city, but had their business, social and political capital vested within the community.
Next, all of us were involved in start ups at some point and were well-known as self-starters. Thus, it was easy for us to work with and around each other. The campaign had a similar feel to a start-up experience – everybody working towards a common goal of success, knowing their role, determined and what needed to be done to get Naheed over the line.
Your second point about consultants was absolutely right. While there are those who have the political experience, and know how to engage potential voters in a meaningful way, it is only those with experience in traditional media and/or in the performing arts and had a voice in traditional media, who have been successful and that we would entrust to a campaign. If we did a canvas across the team, all of us in some way or another had had experience over extended time in traditional media. The lead strategist, had a background in acting and theatre, Naheed was a champion debater and had a regular editorial column in the newspaper and was regularly on radio across Canada. Others had been involved in television and I myself did almost three years of radio in university. The reality of it is that we all grasped what social media offered as a platform for us to share our voice and engage the public and it was embedded into the DNA of the campaign (to which I can attest our online engagement specialist/”brand manager” did a an excellent job). Social media was simply the most affordable and the most effective means for us to do this, given the resource constraints that we initially faced.
That said, we did leverage social media but did come from a traditional background of sharing our voice, engaging people and trying to drive traffic towards the ideas presented under Naheed’s platform. In the end, we knew what was critical to getting Naheed over the line – it was his presence in traditional media. In the last month of the campaign we launched a radio campaign, but more importantly, we started to spend money on newspaper and billboard advertising. In fact, in the last week of the campaign, at the key point voting decision, we knew it was imperative to get Naheed on television. There are many who came up to me after he was elected and said that the television spot – that the lead strategist cleverly engineered for the Saturday before the election – was pivotal in confirming their willingness to vote for Naheed in the end.
Tom: I find that even though the Obama campaign garnered a lot of awareness among politicians for social media, they’re still pretty much in the dark. I’ve told some of the politicians here in the US I know to start to build their social media brand and followers way ahead of the campaign, before they actually need them, and then to continue after they win or lose as the case may be. But it always ends up being a last minute thing it seems. Was this the case for you as well?
Brian: Obama is an interesting case. He all but abandoned his social media base immediately after getting into office. He seemed to recoup some ground, but there was a momentum element there that was lost. In Naheed and the campaign team’s case, we were all actively involved in social media and online and were literate in both forms of media – both traditional and social media. I think one things that we always forget are those people who grasp the environment of social media have come from some sort of background where they have been actors or on radio but they are used to an audience. I think that element of performance is very important because those who get it stay committed to it. With the wealth of knowledge we were garnering from all the campaign metrics, there was one thing that was evident – that while they may not be delving deep into matters of policy, they love the opportunity to be engaged.
This was always part of Naheed’s vision of how he was going to be leading from the front when he got into office and the effects were immediate. He stayed on Twitter and has remained tweeting ever since and his followers have grown and his network effects of those followers have kicked in tremendously on a personal front. Within our municipal government, the process of engagement is being completely reengineered to actually include a substantive social media component. In fact, they are trying to infuse social media into every City department and it’s really working. At one particular public consultation 1,600 people showed up to express their concern about a roadway – that is unheard of! I think that staying committed to the platform and realizing that it’s not about the platform but it’s really about being open, willing to communicate and stay engaged and that is good for business – that is the key to this.
We had a saying amongst the team that we were shocked how well things went. Almost everything that we tried worked. Due to the real-time nature of social media, if things didn’t go so well we learned very quickly to assess metrics to determine the return on engagement. So, if something was not working we were able to quickly repair it. One of the strategists (who was the mastermind of the brand and the online engagement and marketing campaign) was able to monitor every aspect of the website as well as the Facebook and YouTube presence. One thing that he was able to observe was aspects of engagement to various elements of the YouTube/video campaigns. Much of what we learned was counter to the notion of keeping it short. What the metrics indicated was that 3-5 minutes was the optimal time to keep people’s level of engagement peaked within this timeframe. This led to our team developing certain video forms for longer periods of time.
Other aspects that worked really well included the understanding that sometimes certain things have to be broken out and treated separately – hence a number of micro sites cropped up alongside the traditional website. This was useful in providing another avenue for people to provide focused thought and the feedback received here was analysed for sentiment and key learnings from that were taken forward into the campaign. One of my roles was straight up sentiment analyses and to be quite frank it was much easier on a city scale than a national one. I was able to quickly assess things using simple tools within Google as well as just straightforward observations and text analysis on newspaper discussion forms and within Twitter.
The last element was that we were able to deliver multiple touches with the Public and one of the things we did within the polling was to work out the efficiencies as to how do people become aware of candidates, how do they research candidates and how do they decide upon candidates. This was vital in terms of understanding what forms of media work together. We did segmentation based on the candidates themselves but also across all ridings (i.e., voting districts) and this allowed us a unique opportunity to tailor our activities in specific ridings. In the end, Naheed won the majority of the ridings quite comfortably.
Tom: What are some of the things that worked really well for you? And what were some things that didn’t work as well as expected?
Brian: It is hard to say that things didn’t work as expected. As I said before, we were shocked at how well things did work. What was really unexpected was the public response. We had no idea that once people were engaged, they started to take on a lot of these actions themselves independent to the campaign. This was amazing to us. There was graffiti, there was a rap video, there was a purple dawn where the volunteers created an event on their own and outside of our campaign plan. This was the most amazing thing in terms of the unexpected. I think the elements of things that don’t work well and this goes back to scale – the city is a well defined entity and most civic politics are local – I recently got involved with a candidate who was running for a national party and had low levels of awareness. The dynamics of national campaigns in Canada are very different from the US and the timelines are very short and there is not enough time to get the work done to raise awareness.
The other thing too is when you look at the elements of the needs of campaigns now, where it is a data literacy pentagon where the various points are – engagement of voter identification, polling and writing intelligence, web analytics, traditional media monitoring and social media monitoring and analytics, that the better parties have the voter identification piece well sorted out. They already have their core vote established. Needless to say, our candidate didn’t have those resources to do that and the machinery of another party really run ruckshot over the campaign. I like to think that given the starting point for that candidate to where they ended up doing reasonably well – but the reality was that to actually have everything working together there needs to be a number of pre-existing conditions to work and Naheed’s drive and knowledge and saleability while a lot of work, made that campaign incredibly easy.
At present I am actually working on another campaign – for another candidate – and that is a whole different strategy. It is interesting to see that some people feel they can do a Nenshi but the truth is that parties and candidates are becoming much more savvy with the media and it really does come back down to the research and commitment to building a strong political brand.
Tom: It truly is all about marketing. Getting into the details a little more, how if at all did you manage ROI of social media in the campaign?
Brian: The ROI of the social media campaign was pretty simple – Naheed won!
Seriously… if we had to breakdown the individual components, measuring the ROI of the social media element was primarily done using straightforward metrics that were provided via Facebook in terms of engagement, the types of people who were engaging, Google analytics in terms of the buzz around the campaign through the various sources identified within, as well as all of the online advertising that we were using to redirect people to the website and the social action elements of Facebook. Twitter was effectively used to redirect traffic back to the website as well as coverage in various forms across blogs and traditional media. We also assessed a number of links that were posted – links from the campaigns website and content posted on social media within the discussion forums of traditional media (i.e. comments at television, radio and newspaper websites).
However, to establish a dollar amount as an ROI would be tricky – the eternal quest for perfect action-reaction causality.
What was easier to determine was a return on engagement. Given the amount of time and effort and money that we put behind certain elements, it was easy to assess what type and quality of impact we were making and this was used to fine tune every element of the campaign. Yes, everything that we could measure, was measured.
We got to the final week of the campaign with $260,000. This is a ridiculously low number compared to jurisdictions within the US. Hence, one of the reasons why I said it is unlikely that this will ever be replicated. But if we were to break down all of the effort that was put behind us from all involved being active within social media, we could postulate, based on how much money was spent, excluding man hours, campaigning, organizing volunteers and all of our time dedicated to the campaign strictly from a cash perspective of approximately an ROI ratio of $4 to every $1 spent – i.e. for every $1 spent on social media elements (including paid advertising, we received $4 in campaign contributions. If ROI is considered from an election perspective, it would be less than $3 per vote.
This is simply a proxy. If we had to factor in all the time and energy that we put behind us online and into the social media elements, I would guess that it would be in the ballpark of 2:1. Yet, that is positive and this emphasizes why campaigns are notoriously staffed with volunteers. But, as was evident from the tools that we had available to us, it was the primary means that we used for engagement, thus, the end result only emphasized the positive nature of the ROI.
Tom: Here in the US I find a lot of politicians seem to be more willing to invest in traditional media. It’s hard for me to understand why lawn signs for instance is such a thrill for politicians other than I assume some sort of ego shot while driving around town. How were decisions about how to allocate budget between various traditional media and online media made?
Brian: Don’t discount the lawn signs, Tom! They were very important to the campaign!
They fulfill a number of roles including visibility, but more importantly, it’s a badge of support. They are like the status update on Facebook, but in your front yard and to neighbours and strangers. For any politician, this is valuable.
I guess we looked at it from this perspective, which was different than the expectation that supporters would have lawn signs. Naheed earned that right.
Signs feed into the “herding effects” that you hope to stimulate via a political campaign. This is no longer a simple popularity contest – a campaign is effectively a “public-collective” marriage proposal. Yes, asking all of the constituents within a jurisdiction to check off one candidate to agree that they are going to feel comfortable that this person is going to be their voice, listen to and represent them in some form of government. Thus, the desired herd effect is how you get a majority, if not the largest number of supporters coalescing around you.
Tom, everything matters in an election. We had everything broken down by the various stages of a campaign from awareness, to finding out more about a candidate, to the point of how they decide upon a candidate and we did lot of statistical analysis to grasp the efficiencies of certain media, but we also look to the spatial resolution on this on a ward by ward basis. We knew that signs played a role. Hence, larger campaign signs were strategically located based on key demographics that we had flagged within the data, but also in terms of where his favourable response was.
This type of analysis also lent itself to investments in location of the campaign headquarters and how much we should be paying for certain forms of advertising, given that we also assessed, based on the stage of the campaign, what media would contribute in lift based on their strength within factor groupings of various forms of engagement. This was used as a decision-making tool for what types of investments we made and we looked at it also towards return on engagement.
Besides our signs, which provided concrete visibility that Naheed was a candidate in the Mayoral race, much of our early investment (based on how we saw the various stages of the campaign emerging) was on online engagement and advertising. This was ratcheted down as we felt comfortable that the online presence was firmly established and our supporters had become advocates on behalf of Naheed. At that point, we decreased the budget for online advertising and shifted over to the more traditional means. What was the percentage in terms of traditional versus online? It was gauged based on what we were capable of doing with the available money. We had the research and metrics from the forms of engagement we had seen, and knowing where the traffic was emerging, calibrated by other forms of research, including our polling, helped assess the benefit of certain forms of media which helped guide our decisions.
Tom: One of the many concerns here is the question of what proportion of the voters are actually online/using social media? Did you get good numbers on that? I assume older voters were a concern?
Brian: It was pretty simple to get good numbers on who were the potential voters using online/social media. Facebook has some excellent metrics and we triangulated that our demographic criteria, social media data and voter identification. We had a sense of what kind of numbers it would take to win based on potential voter turnouts and we had worked out a number of scenarios around that. Based on this strategic planning, we knew that there were more than enough people active in social media to connect with, and hopefully deliver, a winning result.
Oddly enough, older voters were not an immediate concern. Here’s why. We actually considered the segmentation that I had discussed earlier and grasped that we had to deal with the hyper engaged. The hyper engaged are a very interesting group as they cut across a large demographic – i.e. they can range from a first time voter at 18 all the way to somebody in their 70’s. The fact that they had an online presence and were willing to discuss the campaign was the most important thing to us.
The next phases of the campaign was how do we actually start to build awareness and get people to “infect” others with the idea of voting for Naheed. This is where connecting with older voters became a concern. One of the things that we saw within the data was that Naheed had a number of committed voters. It’s no big surprise that if your parents voted, you would likely vote as well. We also believed that such families discuss politics at home. So if a child was becoming interested in politics, parents would consider their child’s interest and reconsider how they would vote to actually be part of the dialogue with their child. It was our goal to be part of the discussion at the Thanksgiving dinner table (in Canada this is the second Monday in October and turned out to be exactly one week before the election). This is one of the approaches we used to connect with older voters. We also used traditional advertising – including billboards, television and radio to connect with older demographics that we felt (based on our research) that held potential in becoming Naheed’s supporters.
The reality of it now is that most of the initial dialogue is all done online as campaigns are ramping up. It is important that anybody, regardless of the potential voter base, have a presence online to demonstrate that they are committed to being accessible. This does not go unnoticed as there was much discussion concerning our team’s accessibility within the media. Was this a younger older thing? No. Returns showed that we cut across many demographics and Naheed won 10 of the 14 districts. What we demonstrated was that we used the tools at our disposal to connect with the diversity of Calgary’s voters.
Tom: What are your thoughts on social media impact by reaching influencers online, and having them take the message and carry that to others who are not online?
Brian: This is exactly how the campaign evolved for us. We knew that there was a segment amongst social media users who were highly influential, and the influence moved from these highly engaged users to the business users and those who are true socialisers, through to those who are pretty much using Facebook only as a social media platform. This group is becoming more aware of their influence through some of the evolving online monitoring tools – like tweetreach and Klout – and are becoming savvier about what they say and who they are willing to support. They also tend to be a lot more politically “catholic.” I don’t use that term in a religious sense. I mean that they are more interested in politics from a pure politics perspective than to be absolutely partisan.
This is an interesting bunch as they run the full political spectrum from left to right, and in some regards have voted across party lines many times. One of the funnier moments of our campaign was when we were at a cabinet meeting and we did a roundtable on how many parties we had voted for (definitely not a US phenomena!) and we had all voted for at least 3 to 4 different parties over our lives and in some cases some even voted for candidates in all parties.
Thus, we were these people in some regards. We were engaged online. We believed in the platform. And we wanted good people to do well in politics. We had a sense of who we wanted to connect with and also what kinds of messages we wanted to pick up. We knew that these are the people who were more interested in policy. But at some point (closer to election day) policy was not an issue in the campaign – it was the belief that Naheed had a better policy platform than the others. It was our goal to ensure that the influences, discussions and engagements we had seeded moved from a point of content to belief to the permission to vote for Naheed.
Given we understood this particular flow of moving from the influencers to the business community to those people actively using the platform for social purposes. We felt that by the time we were ready to move to traditional media, the hyper-engaged influencers had done a spectacular job and that the dialogue online was purely of support and demonstration of the thoroughness of thinking, but more importantly they had bonded with the associated brands of Naheed and the campaign itself.
Tom: I assume it’s somewhat similar in Canada as here in the US. We can get access to voter records including party, name and snail mail address. Did you try to do any data merge with email or social media presence to connect with voters of the respective parties?
Brian: This data is available and given the limited resources that we had, this was not formerly undertaken but it can be done. It does require a tremendous effort to do it, but there are proxies that we deployed to mimic this. This was critical to the campaign from a performance measurement and resource deployment perspective. It is not secret that parties in both Canada and the United States use voter identification as a critical tool to try to harness as much support before Election Day. Is this corroborated with social media? In some regards yes, but there is a certain amount of geo-targeting that social media offers; in this case it was the city of Calgary.
However, when it came down to the additional elements of watching where our volunteers went, where they were coming from and what they felt comfortable discussing based on the elements of the platform, there were enough cues to start that triangulation process between using various online and traditional media with voter records. In fact, we built a voter prediction model based on voter behaviour, demographics and a host of residential data that was available from the City, that we were plugging our numbers into and a month before indicated that Naheed was going to win this with 40% of the vote. We were shocked at this initial result, given that at this point Naheed was polling at 16% amongst committed voters. In the end it proved to be absolutely accurate. Thus, it does indicate that if we are able to refine this triangulation process, it becomes a powerful planning tool.
I would caution that there are other factors at play that need to go into a data merge. There are elements of commitment that are very difficult to quantify in terms of the psychological dimension they bring to voters on Election Day.
However, establishing a platform to facilitate such triangulation is a definite step in the right direction. It really is a data-driven process where things have to be analyzed daily, if not hourly in critical periods, to work off of benchmark and tracking data against the overall strategy to ensure that you are using the right type of information at the right time to help drive key campaign decisions. An example of this is at a certain point parties who have a voter identification system with committed voters need to grasp exactly what sort of networking effect do these people have through various forms of word of mouth and social media, but also at what point do the psychological effects of saying that you are in on a positive trajectory kick in, to give the amplification effects that lift a candidate or party beyond the line and above expectations.
We saw this in the recent Canadian Federal Election, where the left of centre, New Democrat Party over exceeded their expectations in Quebec. In some regards this was solely based on the strength of character of the Party’s leader. But at some point, something turned within this population where it became a wave of support for the Party. All pollsters and political pundits in Canada did not foresee this and it does indicate that every campaign needs to have a team with data literacy. When I mean data literacy I mean to be familiar with different forms of data and their context and meaning to the decision-making process. This is almost akin to bringing together a group of individuals with such diverse backgrounds that are able to grasp each other’s literacies to yield a comprehensive model of what a campaign strategy looks like and what calibration tools they will require at various phases of the campaign, to ensure they do not miss subtle shifts that could work to their advantage on Election Day.
Tom: To what degree did you use paid social media advertising vs earned? i.e. we had some success using Facebook local ads as well as retargeting campaigns by using cookies on our blogs etc. and then showing them banners elsewhere on the web.
Brian: This is a very tricky question, Tom. There was an article I wrote that was published in the local press prior to my commitment to the campaign that discussed the role of social media and what it offered. Every form of advertising within social media should be about redirection to repeat, reinforce and amplify the core messages of a campaign. However, it can’t be seen as trying to buy the vote. Thus, the message was equally important in terms of what was presented in paid social media advertising.
The earned elements were fairly straightforward based on some of the things we looked at in terms of tweets, page views, unique visitors above and beyond what we saw was coming through a simple click through on paid social media ads. I would not get into the details we used around some of these forms of identification – we worked with what we had and the online team did an excellent job of gathering the data that they used to refocus online content. But the forms of advertising were fairly straightforward and were designed to generate multiples on the paid advertising. Initially, this was likely not the case, but as critical mass evolved, the paid advertising became more sophisticated based on the analysis of the tools offered by the various social media platforms used, as well as other research that was available to us that we provided to the team to recalibrate the campaign.
Thus, yes, the goal was always to get the earned advertising, but at all stages everything was still geared towards having the maximum engagement of the paid social media advertising.
At a city level, this was easy. In the upcoming National election, it is trickier. There are brand elements of party affiliation that are hard wired. However, a well orchestrated and post-election sustainable social media campaign paired with a promise of open government (Gov2.0), could be the decisive factor in many “to close to call” districts.
Tom: Thank you Brian. Sounds like it was a fascinating and exciting campaign, and has given me some ideas.