By Evan Wheeler (@evmw) and Stuart Campo (@stucampo) from Monrovia, Liberia

Screen Shot 2014-11-26 at 6.44.18 PM

Screen Shot 2014-11-26 at 6.43.35 PMYou can learn an incredible amount about a person in just 160 characters.

In the three weeks since UNICEF launched U-report in Liberia, a total of 7,000 SMS — or roughly 1,120,000 characters — have been received through the platform. These messages have come from just over 600 young people across the country, and have allowed us to learn an incredible amount about who they are, what they care about, and what they expect from U-report Liberia as a platform for youth engagement and social change.

Getting started, small small

Source:  Source:


In the U-report registration process, we ask for some basic demographic details: name/nickname, age, gender, and county of origin.

Based on responses from U-reporters as of today, we know that:

→ 62% of registered U-reporters in Liberia are male and 38% are female

→ 75% of registered U-reporters are under the age of 30

→ The majority of U-reporters so far have come from Montserrado (Monrovia) and Lofa counties — the two areas where face-to-face outreach efforts have taken place so far.

Source: how the number on the map (390) is lower than the total number of U-reporters registered? That’s because not everyone so far has successfully completed the registration flow. Just 390 of 621 U-reporters — or just 62% — have successfully completed all of the registration questions, which in itself says quite a lot about the clarity (or confusion) in the registration questions.

Perhaps the biggest challenge for U-reporters in completing the registration process is spelling. The names of Liberia’s counties are not auto-corrected appropriately by most phones. And, while there are some algorithms for phonetic matching (comparing two texts based on how they sound rather than the composition of their characters), it can be difficult to make sense of location names — and U-reporters can be quite creative in attempts to type the “official” spellings of places that are rarely spoken or seen in print. Newspaper circulation is limited in Liberia and, other than certain grade-school textbooks, there aren’t many other print materials referring to specific locations in Liberia. So, its easy to be unfamiliar with the “official” spellings of administrative boundaries like the counties — especially when major cities within a county are cited more often in everyday language.

For U-report, we collect common misspellings for each county so that we can interpret them correctly. Unfortunately, in order for U-report to learn, our users may have to endure the frustrating experience of trying to get their attempt “close enough” to be matched. We quickly learned that “MONSTERrado” sounds a lot like “Montserrado”, and is far more fun to write.

A U-reporter tries to submit the name of their County (“Montserrado”). This persistence is not seen in many cases, and often users give up after 1 or 2 failed attempts, leaving them unassociated with a geographic label/group.
A U-reporter tries to submit the name of their County (“Montserrado”). This persistence is not seen in many cases,
and often users give up after 1 or 2 failed attempts, leaving them unassociated with a geographic label/group.

The same thing happens with trigger words, like “join”, which we use to initiate the registration flow — but not entirely for the same reasons. We do see some “misspellings” (or misinterpretations) of “join”, such as “joint” or “j0in” or “john”, that can be attributed to word-of-mouth distortion or simple misspellings. The bigger problem is that our instructions are unclear. Or, as we discovered on our big launch day, we hadn’t thought through our sales pitch and made sure it actually made sense. The first flyers and posters we printed included the call to action: “text JOIN to 8737.” Instead of receiving only the word “join,” many U-reporters send in more than “join” and often the entire phrase.

As we see through the registration process, tailoring the language of each message is essential to getting higher rates of ‘successful’ response. For this reason, we’ve placed the entire process of content generation in the hands of the people who know best: the U-reporters themselves.

Polls by youth, for youth

One of the best ways of building interest and trust is to generate polls directly with young people.  In an earlier post, Chris wrote about our early stage design sessions with a group of adolescents in Monrovia. This type of co-creation has proven key to “hitting the mark” with poll questions, and is something that remains foundational to the growth and national scaling of U-report. Every character matters–especially when you are dealing with a richly diverse group of users with varying levels of ‘mobile literacy’ and exposure to SMS communication.

In the first poll, we asked a series of questions about Ebola awareness. We learned that other people (face-to-face communication) remain the most trusted source of information about the virus,and that no school is the thing that bothers young people the most about the current situation.


Seeing that youth considered “no school” to be the most bothersome thing about Ebola, the U-report team decided to delve further into this subject–asking a series of questions about the importance of education for youth across Liberia. Providing feedback from previous polls is key, so the first question in this series mentioned the main result of focus from the earlier poll and then asked a follow-up questions.


The results of this poll shed more light on why “no school” was such a “bother” to U-reporters. Education is important because it “prepares you for tomorrow” (67%) and helps you “gain knowledge (30%). On the question of whether or not there is a problem if someone is a high school drop-out, U-reporters replied with a resounding yes (81%).

Although we had not originally included a follow-up question on high school drop-outs, the large number of respondents who expressed concern about this made the U-report team curious. So we added a fourth question to the poll: Wat is da problem wit drop-outs? The responses were illuminating, and speak for themselves:

Messages from three different U-reporters in response to the question “Wat is da problem wit drop-outs?”.
Messages from three different U-reporters in response to the question “Wat is da problem wit drop-outs?”.

Unsolicited insights

Although the majority of the interactions and conversations generated through U-report have structure and relate to national or sub-national polls, some of the most interesting and insightful messages coming through are those we haven’t asked for.

A total of around 400 unsolicited messages have been received from the world’s newest U-reporters in the past three weeks. These messages run across a wide spectrum, and serve as a window into a community that UNICEF and our partners have worked with for quite some time, but are now getting to know in an entirely different way than usual.

We’re planning a deep-dive analysis of all the unsolicited messages received thus far in the weeks ahead, so for now we’ll whet your appetite with a few of our favorite messages and what we’re learning from every 160 character insight receive from the world’s newest U-reporters:

U-reporters are excited to celebrate their communities.


Screen Shot 2014-11-26 at 6.46.47 PMU-reporters are hungry for information, about a wide range of topics.

Screen Shot 2014-11-26 at 6.47.00 PMU-reporters also are reflecting on some big questions


Screen Shot 2014-11-26 at 6.47.11 PMU-reporters are dedicated, and eager to participate.


Screen Shot 2014-11-26 at 6.47.19 PMU-reporters have high expectations about the frequency of messages coming from the system.  (this message came in just 12 hours after the previous poll)


Screen Shot 2014-11-26 at 6.47.29 PM





U-reporters are appreciative.


Screen Shot 2014-11-26 at 6.48.25 PM


Some things, we still don’t understand. But that makes us even more curious, and eager to connect and learn more about the lives of the world’s newest U-reporters.

Screen Shot 2014-11-26 at 6.50.00 PM Screen Shot 2014-11-26 at 6.50.10 PM