The Power of IVR in Predicting Election Outcomes

Rumbi Mtsambiwa and Jennifer Henderson, Viamo

Viamo conducted a national poll in Zimbabwe to gather pre-election opinions via Interactive Voice Response (IVR) between July 26th – 29th, 2018. We partnered with Kubatana, a citizen advocacy organisation, to design the questions and publish the results to a large network of citizens.

Our opinion survey showed that the majority of people (47%) believed that Emmerson Mnangagwa would win, with a confidence interval of 95% and a 5% margin of error. Other estimates included a pre-election in-person survey conducted by The Mass Public Opinion Institute, on behalf of Afrobarometer and funded by the EU. This survey found that 43% of Zimbabweans believed Mnangagwa would win the election, with a margin of error of +/-2.0%. The final election results show that Mnangagwa won with 50.7% of the vote. These results show that Viamo’s survey was accurate, and cost-effective as it was conducted in under two weeks and for less than $15,000.

The methodology for the Viamo poll is known as RDD (Random Digit Dialing). A research paper by the Centre for Global Development in 2015 concluded that in countries with 73% or higher mobile penetration, RDD can generate a representative sample. According to recent statistics, Zimbabwe is at 85% mobile penetration.

Pakistan ‘Get Out the Vote’ Campaign Reaches 43 Million

Abdullah Saqib and Jennifer Henderson, Viamo

The UNDP and the Election Commission of Pakistan sought to mobilize citizens to participate in the electoral process and cast their vote in the 2018 elections. They partnered with Viamo to run Pakistan’s largest mobile engagement campaign ever, reaching approximately 26 million people through calls, 16.5 million through SMS and 0.9 million through social media (Facebook and Twitter) nationwide between July 21st – 24th.

This campaign was organized within two weeks, in collaboration with the relevant government authority. There were 105.96 million registered voters, and we reached over 40% of them, more than 43 Million people in total.

Viamo representatives meeting with the Chief Election Commissioner of Pakistan.

The speed and scale of this project did not take away from the quality. To ensure maximum impact, there was a location-specific strategy used, based on the principle that each population segment should be reached in the relevant medium. SMS messages were sent to urban populations with higher literacy rates, calls were sent to rural populations with lower literacy rates and social media targeted the urban and semi-urban population. A range of tactics were utilized to maximize call pick-up rates, including variations in shortcode/longcode usage, a minimum of two call re-attempts for each subscriber, conducting outreach at preferred times, and the active involvement of all four Mobile Network Operators in the country.