Connecting Vulnerable Populations to Cessation Resources Through Digital Paid Media: NCI’s Smokefree.gov Paid Media Campaign Produces Notable Results
In 2003, the Tobacco Control Research Branch of the National Cancer Institute (NCI) launched the Smokefree.gov initiative (SFGI), a large mobile health smoking cessation program. The Smokefree.gov Initiative (SFGI) has since evolved into numerous mobile-optimized websites, smartphone applications, text message programs, and social media channels.
With such diverse resources, ensuring that these tools reach current smokers, including smokers who represent vulnerable populations (e.g., pregnant women, teens), is a significant challenge. One means to address this challenge is via paid promotional campaigns.
Paid media campaigns have the benefit of leveraging multiple channels in order to reach target audiences, and as such, can leverage both broad media (e.g., a website) as well as highly tailored communications delivered in a targeted way (e.g., a landing page for a website tailored for one specific audience). Given that current smokers in America represent both types of audiences, paid media campaigns have enormous potential to be particularly effective in reaching these groups.
In the fall of 2015, we worked with NCI to plan a national paid media campaign to connect specific vulnerable populations (i.e., target audiences) to SFGI’s smoking cessation resources that could help them quit smoking for good. Each audience, whether general – like male smokers – or niche – like pregnant smokers – required a personalized approach.
Step 1: Data Collection
We took a six step approach to campaign development. First, we began by compiling consumer and media consumption data to learn more about each target audience group – male smokers, female smokers, pregnant smokers, LGBT smokers, teen smokers, and Spanish-speaking Hispanic smokers. We then filtered data for each group based on points known to be relevant to smoking, such as socioeconomic status, habits, and location.
Step 2: Audience Profiles
Once we compiled our data, we created targeted audience profiles with overarching attributes and indices of likelihood of different behaviors for each group. We touched on numerous topics – education, income, profession, habits, belief systems, relationship status, likes/dislikes, media consumption, music preferences, and frequent/favorite activities, to name a few. If a group had data supporting an index of over 100 for any particular item – whether it was not completing high school or really enjoying karaoke– we knew that data point was especially relevant to that particular group.
Step 3: Audience Personas
Third, we used our target audience profiles to create a “persona” representing each audience. Our personas were “real,” individual members of a target audience group. Each one had a name, photograph, and detailed identity and worldview. We even created a personal-behavioral context for each persona exploring his/her physical, emotional, and cogitative attributes. These personas facilitated our understanding of the perceptions and behaviors of each group, and enabled us to proceed with our media planning and targeted content creation with that single person in mind, rather than relying on a spattering of data points about a group.
Step 4: Media Plan
Using our personas, we developed a comprehensive media plan to reach each target audience through paid media. We employed granular targeting methods to reach each of our vulnerable populations effectively, as some were more niche populations, like pregnant smokers, which are extremely difficult to target. We then refined the elements of the media plan by evaluating multiple metrics including geography, time of year, and individual online behavior and media consumption patterns. While each group had different characteristics, all groups indexed high for time online, video consumption, and social media use, supporting our decision to focus on digital media. Digital media is especially effective in operationalizing the development of persona-based campaign materials and for reaching smokers from both general and vulnerable populations due to the high exposure to online material via daily habits and media consumption patterns. Additionally, use of digital media in paid media campaigns has a very high return on investment – smokers from both general and niche populations access digital channels with relatively equal frequency, and we can easily and inexpensively manipulate digital channels to provide highly tailored content to different audiences.
Step 5: Content Creation
Next, we created unique content for each target audience. In addition to audience-specific copy and design in ads, images, search terms, videos, posts, and marketing emails, we also created a specific campaign landing page for each target audience on Smokefree.gov. The content on the male landing page did not match the content on the female landing page. The content on the teen landing page looked nothing like the content on the Spanish-speaking landing page. We created each page in a way that not only followed a visual asset map that would appeal to each group aesthetically, but we also provided different resources on each landing page depending on what would be most useful to a particular audience based on engagement data and usage rates. The articles on the landing page for pregnant smokers, for example, were different than the articles on the landing page for LGBT smokers. While every page featured an app, a text message program, top articles, and social media content, we uniquely tailored each page to the audience most likely to visit the page because they engaged with a piece of the paid campaign that targeted that particular audience.
Step 6: Optimization
A key benefit of executing a primarily digital campaign was our ability to optimize performance in real time through continuous monitoring within individual ad platforms to ensure satisfactory pacing, spend, and performance for each audience within each platform. We frequently adjusted digital bids to optimize budgets and performance, and we moved individual mediums’ budgets toward the most cost-efficient placements and refreshed creative content throughout the campaign to help prevent fatigue. Typically, we might expect a decline in click through rates once a campaign has been active for a few consecutive months, but our ongoing optimization and creative refreshes prevented variability during time periods in which we would typically expect to see a decline in engagement.
For example, after our native advertising with Outbrain (a piece of our planned media mix from our initial media plan) proved to be a successful and cost-efficient medium during the first three months of the campaign, we added Yahoo native advertising to our media mix in March to see if it would perform just as well. Ultimately, Yahoo native ads turned out to be the most efficient driver of web traffic for the entire campaign – cost per session was only $0.12 and the ads drove almost 372,000 sessions.
The campaign, which ran from December 28, 2015, through August 4, 2016, confirmed the value of the additional steps we took in building targeted user profiles and audience personas. It helped us identify improved strategies for reaching our target audiences, both in outreach methods and in the content we created for each group. Using these strategies, nearly every aspect of the paid media campaign performed at or above benchmark rates for ad engagement, cost efficiency, and web engagement, suggesting that the paid media elements we chose were both cost-efficient and effective in reaching our target audiences.
Web traffic exceeded our goals for the campaign time period – we initially set an overall target of 2.5 million sessions during the campaign period, but we ultimately drove more than 3 million. Our digital paid media elements directly drove more than 1.1 million of those web sessions, and our web traffic increased by 140 percent year over year. We also saw significant increases in both organic and direct web traffic, with uplifts at 49 percent and 60 percent, respectively. We can partly attribute these upticks to increased awareness from paid media, and we also acknowledge that the campaign’s existence creates a “halo effect” that naturally drives more traffic to the website, even when it’s not directly attributable.
Our paid social media ads drove nearly 23,000 engagements, and despite having fewer posts during the campaign period compared to the same timeframe the previous year, the Smokefree US Facebook page saw a 109 percent increase in shares, a 102 percent increase in likes, a 95 percent increase in comments, and a 50 percent increase in engagement per post.
We also saw over 33,000 adult SmokefreeTXT signups during the paid media campaign (a 45 percent year-over-year increase for adult sign ups during the same time period the previous year as well as 1,336 new teen SmokefreeTXT signups.
Our app focused Facebook and Display ads alone resulted in more than 124,000 clicks to the iOS and Android app stores, resulting in 23,554 total downloads of our QuitGuide and quitSTART apps during the paid media campaign period.
Last, but perhaps most notable, the campaign consistently gained in cost efficiency throughout the duration of the campaign. The overall campaign average cost per session was extremely low at $1.00 – very strong considering the media mix and overall scale of the campaign.
The experience with this campaign and its success provides a few key takeaways for anyone considering a digital paid media campaign for healthcare intervention and/or targeting niche populations. If you have to prioritize funding, Search, Google Display, and native advertising were the three most cost-efficient mediums when it came to driving sessions, and opt-in email marketing was a huge driver of engaged web users that also enabled us to more thoroughly highlight specific tools and resources. The campaign results also highlight the importance of maintaining a diverse media mix with channels that are geared toward the media consumption habits of specific audiences – our decisions were heavily based on the data-driven market research that helped us create our audience profiles and personas, and it helped us reach audiences more effectively than ever before while driving substantial engagement.
Our evaluation of this campaign confirms the value of building targeted audience profiles and personas in identifying improved strategies for reaching special populations in ways that are both cost-efficient and effective.