MobileActive.org

Intro: Why Mobile Polling?

Polling via SMS can be a unique way to engage current supporters and attract new audiences. Polls can ask any number of questions, from opinions about an organization to views on a controversial issue. However, perhaps the most valuable aspect of polling isn’t the feedback that organizations receive directly from a poll, but rather the relationships with constituents and growing mobile support base that polls can help build.

Organizations engage in mobile polling for two reasons:

• to generate a list of mobile numbers to use for future communications and engagement
• to get an informal sense of constituent views for use on an organization's web site, for generating media coverage, and learn more about a particular segment of its constituency.

Benefits of Using Mobiles for Polling

• Quick response: People have their mobiles on, can carry them with them most of the time. Hence, a high percentage of participants respond within the first 10 minutes of launching the survey.

• Demographics: the mobile phone is the most ubiquitous technology with the broadest demographic reach. It is also one of the few ways to reach a younger market.

• Inclusion: Anyone with a mobile can join in the survey, especially if you provide a free number. There are more people with mobile phones than Internet access, so mobile surveys may lessen a digital divide inherent in surveys on the Web.

• Immediacy: A respondent can be reached at exactly the moment that his/her opinion has been formed. This is useful for constituent satisfaction surveys where you want an immediate response. For example, you might want to ask: How long did you have to wait to receive a response; or: on a scale, how satisfied were you with the experience, etc.

• Additional information: Other data can be collected on the participant, including GPS location and activity with the devise* (*With prior participant permission).

• Mobile list generation: Mobile polls, when widely distributed, have the potential to generate a list of mobile phone numbers. You can ask individuals participating in the survey whether they want to continue to receive information from you (about the survey results, for example, and other information from your organization). Since mobile communication is strictly opt-in in most parts of the world, mobile polling allows you to increase the list of individuals who agree to be contacted by you (via mobile or through other means).

How Does Mobile Polling Work?

There are three ways in which mobile phones are used for polling:

1. SMS Polls. For SMS polls you will typically set up a poll with a vendor or a do-it-yourself software (see Desktop SMS Tools that asks a question via sms with a pre-determined set of choices to text back. The polling software records the respondent's answer via sms and allows you to view, download, and analyze responses.

Respondents are driven to the survey via web widgets, Web sites, fliers or other media (advertisements, billboards, radio announcements, TV spots, etc). You can ask respondents to forward the poll information via SMS as well.

2. Mobile website (WAP) polls. Alternatively, in environments where respondents can access the Internet via their mobile devices, you can set up a simple web-based survey that is customized to be viewed on a mobile phone (that has the ability to access the web).

Using the built-in WAP or WHML browsers on a mobile handset, you can ask a user’s opinions on any range of topics. By adding a “sniffer” that determines which phone a person accessing the survey is using, the server can feed pages to the user that are optimized for their phone’s particular browser. There are a number of vendors that provide this service around the world. Below are some screenshots from some of them.

3. Voice (calling mobile numbers). The most expensive option (and prohibitive in many countries)are phone surveys that call users on their mobile numbers. This is a method used primarily by commercial polling firms who need statistically valid data.

Case Study: Media Focus on Africa

Prior to the Kenyan elections in December of 2007, the Media Focus on Africa Foundation (MFAF) asked its constituents some tough questions.

Should politicians accused of corruption be prevented from vying for political seats? Is tribal identity more dominant than the identity of being a Kenyan? Can voting still deliver credible results after the chaotic party nominations and bribery?

The questions were advertised on television, radio shows, and newspaper advertisements. Thousands of Kenyans responded to the polls via SMS on their mobile phones, helping to bring issues of voting and civic participation into the national conversation.

How did they do it?

The poll project was part of MFAF’s Election Assistance Campaign, which sought to promote civic participation and discussion of political issues prior to the December 2007 Kenyan elections. The polls were produced in collaboration with Butterfly Works.

MFAF advertised the polls on a talk show, a youth program, a popular drama series, the Internet, and on twice-weekly full-page ads in a national newspaper. The group also sent out SMS messages with the current week’s question to people who had responded to the poll the previous week. According to Linda de Kooning, a media consultant at MFAF, the objective of the polling campaign was to make the television programs more interactive and use the responses as input for upcoming shows.

People would send SMS messages in response to the questions asked on the various media. The group asked users to give letter responses (like A or B) because MFAF only had one short code, and would sometimes ask multiple questions during the course of a television or radio show. This allowed for “A” to mean yes for the first questions and “C” to mean yes for the second. Because the answers people sent were inconsistent (some people would respond with “yes” instead of letters, and others would send wordier responses), MFOA had manually filter and sort the responses.

People who responded to the polling questions would then receive a text message in response from MFAF that contained information about the polling topic. “This was to make them really think about the issue,” said Linda. “For me this was the most important part of the whole campaign.”

One problem that the campaign ran into had to do with the design of the questions. Linda noted that they ran into "editorial interference" -- media outlets wanting to change the questions content, and resulting in "middle ground questions." "Sometimes I thought, what are we really polling now?" said Linda.

What did it cost?

MFAF originally tried to use do-it-yourself software and a modem to send and receive the polling messages, but found that their modem wasn’t able to receive large amounts of SMS at one time. They eventually used a polling platform and shortcode from Kenyan mobile vendor Interactive Media Services. Linda said a shortcode number in Kenya costs about 100 euros each month for use on one of the networks.

Poll participants had to pay for the price of the text message (about 5 KSH, or $.07 US) to respond to the poll. The group found that if the campaign covered the cost of the text message it tended to get a lot of spam responses. However, the group decided that asking participants to pay for the cost of the SMS tended to exclude poor people, who might not respond because of the cost.

What was the impact/results?

Each poll received between three and four thousand responses per question, with most responses coming from newspaper advertising and the viewers of the drama series. The results of the polls were published in newspapers, on the MFAF website, and in some talk shows. Overall, the group estimates that it had about 30,000 unique users respond to the poll.

The goal of the campaign was both to engage participants in the questions, to gauge (albeit fairly unscientifically) public opinion in Kenya, and to build a list of phone numbers to continue mobile campaigns in the future. “What we intend to do is build up a relationship with all the people who participated and try to provide them with more accurate and better information,” Linda said.

In order to measure the demographics of the poll respondents, MFAF sent follow-up messages asking respondents demographic information such as gender, age, and location. It received about 1,000 responses, and found that respondents were split roughly equally between men and women and most people who responded tended to be between the ages of 18 and 30.

Key Questions You Should Ask Yourself

• What are your goals – what are you trying to accomplish with the poll? Do you want to grow a mobile list, engage an audience, get feedback on your programs or services? Be clear about what your strategic objectives are for the poll.

• Who is your audience? If you design an SMS poll, does your audience have and routinely use text messaging? An audience that doesn’t use SMS is obviously unlikely to respond to polling requests. Consider what audiences are more likely to have mobile access and use text messaging – likely younger audiences – and realize that many of the responses are likely to come from that demographic. If you design a WAP poll, consider whether your audience can access the Internet on the phone and what possible barriers are (costs for data access, handsets, etc.)

• What is your poll topic? Ideal topics are timely, controversial, or at least interesting enough to elicit a response. Boring or stale questions are unlikely to attract responses.

• How will you word the question and responses? The question should be short and to the point – remember that you have a limit of 160 characters. Ideal questions are clear and concise, so avoid confusing wording or double negatives. The question should contain the expected answers and not be open ended. YES and NO are standard responses for opting into or out of a mobile list and recognized as such by mobile polling platforms. For polling purposes, consider using Y or N or 'yay' or 'nay' as alternative responses to polling questions.

• Logistically speaking, it’s best to not have more than three possible responses so that people don't get confused by too many options. If this is an ongoing campaign, consider changing the question about once a week.

• Will you follow-up with people who responded to the survey? After someone sends a poll response, consider sending them a follow-up message requesting that they join your mobile campaign. You could also consider following up with a request for demographic data that will make the poll results more meaningful. Be sure to ask whether you can follow-up (many carriers have opt-in rules for continued communication).

• How will your constituency know about the poll? Studies have shown that polls advertised on television – where respondents are likely to be sitting with their phone at their side – have higher response rates. Radio, print, and web advertisements are also an option. SMS polls have also been effective at events such as concerts. For example, an SMS poll at the Live Earth concerts had a response rate of 20% of the 18-34 age group at the concert. On average, 63% of responses were from within the first hour of the survey announcement.

• How will you distribute your poll? Polls can be distributed two ways. The question can either be marketed directly through various media or users can text a short code, which then responds with the poll question. Consider which method would work better for your mobile campaign. Although the second method requires an extra SMS, it does allow for users to keep the message in their phone so they won’t forget the question or responses.

A Mobile Polling Campaign in 10 Simple Steps

1. Decide if you will hire a mobile vendor or go the do-it-yourself method. If you hire a mobile vendor, they will help you setting up the poll and bulk SMS. They may charge a monthly fee. For vendors focused on the NGO sector, you might want to check the MobileActive vendor directory.

2. If using a vendor, they will provide you with a short code and its platform that lets you track poll results.

3. Design your polling question and answers carefully.

4. Test the poll with some users to be sure it is logical and makes sense to them. Revise and test again.

5. Craft a mobile marketing strategy to advertise the poll. Be clear about your goals, audience, messaging, and integration with other media.

6. Integrate the poll with your website and current campaigns. Consider posting the poll questions on your website.

7. Set up your web site and test it.

8. Share the results. Post them on your website or other media that you used to advertise the poll.

9. Follow-up with people who responded to the poll. You could consider sending each user a message thanking them for participating and asking if they want to be notified of other campaigns or receive SMS notifications. You can also ask people to text in their email address in order to continue with web-based follow-up or fundraising. Remember to follow the opt-in rules mandated by the Mobile Marketing Association.

10. Evaluate results. How many responses did you receive? What percentage of people chose to opt-in to your mobile campaign after responding to the poll? Is there anything you can deduce about the demographic that replied?

11. Repeat.

Costs

The easiest way to conduct an SMS poll is to work with a mobile vendor. Fees will include the vendor fee, a setup fee, and the actual cost of the SMS. Most vendors will also provide you an online toolset to analyze data and measure responses. If you use do-it-yourself software, consider the costs of sending and receiving SMS messages for your country.

Additional costs include marketing and advertising the survey in a variety of media.

Also consider intangible costs such as staffing costs to set up the survey, test it, refine it, and then, upon completion, analyze and publicize the results.

A Checklist for Best Practices for Mobile Polling

• Make sure you have very clear campaign objectives.
• Market the campaign effectively.
• Integrate the campaign with your website and other media.
• Follow-up with people who responded to the poll and recruit them to join your mobile campaign.
• Write polling questions with concise and effective wording.
• Use questions that are controversial or interesting.
• Share the results of the polls with your constituency.

Additional Resources

SMS Polling: A Methodological Review