Within Facebook lies a community of people who enjoy creating and responding to polls, which, for this case study, are short questions with a limited number of responses. There are hundreds of polling groups on just about every imaginable subject, along with the general use groups. Based on a survey of the groups, there are at least 200,000 regular poll users. There are also regular Facebook users who ask questions within their own feed. The possible user base for quality polling could range into the millions.
Facebook currently has a limited polling feature that only works in groups, and only allows for multiple choice questions. Results cannot be downloaded or shared.
The new Polling Library will expand the types of polls, allow users to create polls directly in their feed, and share them in a library. Marketers can add polls to their paid ads for more effective user feedback.
The original idea for the polling library came from two users stories:
A user on Facebook initially asks for book recommendations in a post. They receive hundreds of recommendations in the comments. Some comments contain long lists. Some only have titles without authors, or authors without titles. The user is grateful for the incredible quantity of data, but ultimately frustrated. They ask if there is a way to consolidate the information without having to go through each post.
A marketer sends out a poll to an extensive mailing list looking for a 1-5 star rating on the quality of a new product. The email includes a link to a popular polling site, and the poll is short – only 2 questions. Unfortunately, there is a very low response rate. The marketer reaches out to colleagues on the mailing list for feedback. They report that either the link looked suspicious, and people are learning more and more not to trust unsolicited emails. Also, because people receive so many emails, many were never opened or went directly to a spam filter. The marketer is looking for a more direct way to reach their clients.
In combining these stories, the initial concept was for an app that could send a polling question in a multiple of formats to any platform. This brings the question to the responder’s preferred and trusted platform – even a text message. Then the data is collected and available for download as a spreadsheet or chart.
Before building a complex app, I decided to create a proof of concept feature. With some polling features and a massive user base, it made sense to limit the scope to Facebook. This provided an existing design template and user flow.
As this is a new feature, my research focused on user interest, product viability, and how it compares to current polling options. To determine if users were interested, I focused on the two user stories. For the recommendations question, I surveyed Facebook users about how they request, respond to, and organize recommendations. I interviewed people running ads on Facebook to see if polling would be of interest.
For product viability, I interviewed a professional poll creator who specializes in the poll creation software Qualtrics. I also investigated the design patterns of Facebook to see how a polling library would fit into the existing structure. And finally, I compared the feature to some of the most popular easy-to-use polling products – Facebook’s current polling feature, Google Forms and the free version of Qualtrics.
The general consensus was that while most people are looking for recommendations, they do not have a satisfactory way of organizing or keeping that information. The most common way to process a recommendation was for the person to immediately look it up, and hope to remember it. Half the responders were happy with their current system for organizing recommendations.
The feature would appeal to a specific kind of user, more than the general public. Facebook has a number of polling groups that might have interest. A search for “poll” groups shows about 60 groups and 200k members.
I interviewed three different companies that post ads on Facebook: a realtor, a visual artist, and a performance group. All three expressed interest in adding polls to their paid ads.
The realtor expressed the most interest, explaining how she could use it to make connections with potential customers in a localized market.
Reviewing Qualtrics and Google Forms helped determine the types of poll formats that could be added to the feature. After interviewing the Qualtrics expert, we determined the best formats are short answer, star rating, scale, and the current format – multiple choice. In keeping with the current format, new formats will also allow three questions within the main question. For example, when looking for a book recommendation, the person can include spaces for author, title, and why someone likes the book.
The biggest disadvantage of using a third party polling platform is that the user has to find a link and decide to click on it. By keeping the poll in Facebook, the user doesn’t have to leave their comfort zone or take extra steps. However, unlike 3rd party platforms, for the polling feature to remain effective, the number of questions are specifically limited in quantity and format.
The feature should have a familiar feel and flow so new users will adjust to it quickly. This proved to be a challenge because Facebook has many different features, some hidden or only available if someone knew where to look. Some features, like Marketplace, had two completely different points of entry – by creating a new post, or through a dedicated layout. For the feature, I created both flows to determine if they are both viable options.
This task flow demonstrates how someone can create a poll for book recommendations. The poll will appear on their feed. Then anyone seeing the poll can respond to it and see their answer in the comments.
Any type of poll can be created and responded to through the Facebook Feed. Polls can also be accessed directly through the Poll Library interface.
Within Facebook, a member of a group can create a poll in a post. The options are limited to multiple choice, either with one or multiple answers. I’m using the design patterns as a starting point for the new feature so that it fits in with the rest of the website.
The layout of the Poll Library is based on Facebook’s existing Marketplace interface, which allows anyone to create an item for sale and post both in the Marketplace and their feed. This is a proven design, although as research will show, not as easy to use as creating a post.
Testing both paths was critical in determining the best user paths and the overall viability of the feature. I created two tests in Maze.co where participants could walk through the paths, first creating a poll then taking the same poll. Participants were at least casual users of Facebook.
Overwhelmingly, the preferred path was creating a poll by creating a post. While there was not a high level of success – 34% – those who did felt that the path matched Facebook’s environment. 100% of them agreed that the survey looked as they expected. The main point of failure was more of a technical issue than design. Due to how Maze shrinks a layout, the click area to create the poll was too small on some screens.
The path to create a poll directly through the poll library failed. The point of entry is on the far left side of the screen, and users expected it to be in the middle. This was designed after the Marketplace interface. In a post interview test, the in-person participant couldn’t navigate the Marketplace either.
A future design of the poll feature will require re-evaluating the home page. Is there a better existing layout to emulate, or would the feature require something new? With some small adjustments to creating a poll as part of a post or advertisement, this path could be a successful feature.
Overall, the idea of creating polls through Facebook is viable, but limited. There is much more benefit for marketers than the general public. If the feature existed, it could still be an option for both.
Something to consider is that Facebook is overloaded with features. There are approximately 30 features ranging from selling goods to finding a blood bank. Most of the features require 2 or 3 clicks to find, and it seems very few are used. When adding a new feature, it would need to be heavily promoted at first. Otherwise, it would blend into the background.
While Facebook is a highly used platform, it may not be the best for a feature like this. Linkedin and Reddit also have polling communities, and could prove more successful. Platforms with more personalized interfaces like Instagram and Twitter would also be interesting options to explore.
Ultimately, I would like to return to the original concept – an app that sends the poll to any platform. Bring the polling to the people, and democratize the information.