Monday Lightning Talks 13.45-14.15 (Bridewell Hall)
Dropping down your ignorance Ratio: Campaigning meets KNIME: Rodrigo Barahona, Oxfam Intermon
Ignorance is not bliss when you want to be accountable with your constituency and partners, and mostly when you want to learn from experiences and improve your efficiency in campaigns and advocacy. KNIME and other free tech tools have helped us build reliable systems to evaluate complex campaigns, advancing our capacity to know and learn from evidence on what works and what doesn’t. However, as any other tech-lead process, there’s always the cultural factor to be addressed.
Read Rodrigo’s blog post or watch his talk below!
Lightning Talk Details:
Building on the philosophy underneath Vanity Metrics, at Oxfam Intermon we started developing a system to evaluate campaigns and spike actions. We wanted to improve our knowledge on what works and what doesn’t on mobilizing citizens (mostly through signing petitions or online action) as part of advocacy efforts. The turning point was the fact that after a very successful campaign (more than 100k signatures and a lot of noise on social media, that helped moving forward our advocacy work in Guatemala), we were asked to explain the story of our success… and there were some answers that we just couldn’t provide because we didn’t have reliable information.
We somehow couldn’t explain where the majority of those signatures came from, so we couldn’t establish recommendations for future similar actions. We started implementing a relatively simple Google Analytics Tracking system, and made our customized version on google spreadsheets, so we could keep a track of all the different codes that were being generated in different campaigns. With this process we could track a high percentage of the traffic to our web-petition site, but we were still challenged by the incapacity to track the different sources until the signing. So we were better on knowing where do visits come from, but still not where do signatures come from.
So after more tests, we finally came out with the solution. But this system in reality duplicated our tasks, as we had to do cross analysis of two different web-pages to get the conversion rate. So when we were asked to implement the improved system on an upcoming campaign that had 4 landing pages (x 2 as it was in 2 languages, x 2 as required by the conversion rate analysis), we thought it was going to be impossible, until we came across KNIME, a tool that has helped us to configure periodic automatic tasks of web analysis that reduce what could be an excessively time consuming work. Results were magnificent in terms of learning and accuracy of the system.
Two final considerations: the human factor is a factor; launching this system requires the implication of very different profiles on the organization, some of them reluctant to change and to incorporate new behavior needed to keep the system clean and accurate. KNIME as a platform for analysis also has proved to be useful to other many analytic tasks that we undertake in campaigns MEL (such as twitter exploitation).