Evolving PYLON: How Customer Feedback Drives our Product Roadmap

19th August 2015 0 Comments

Last week, DataSift released PYLON 1.5, our third package of enhancements since announcing the PYLON platform for Facebook topic data. The theme of the 1.5 release is incorporating ease-of-use feedback from our early PYLON customers. All of the features in 1.5 were requested by customers when they began to trial PYLON for Facebook Topic Data in May. We prioritized their feedback and turned around a quick release to make PYLON even easier to use.

Small clarification regarding our vocabulary. By PYLON we mean our real-time processing engine for granular, human-generated data, enabling filtering, augmentation, further processing via VEDO. PYLON then provides aggregated and anonymized insights into Human Data. PYLON for Facebook Topic Data is the first data source to be provided by DataSift’s PYLON technology. It provides aggregated and anonymized insights from what people are sharing and engaging in on Facebook.

More important than the timing of our release cycles is our commitment to iterating the product based on customer feedback. Early customers are driving the direction in which we’re building the platform. But how do we triage, prioritize and incorporate their feedback?

It’s two parts science, one part art. Because PYLON supports such a broad range of use cases, our sales team was able to provide several hundred feature requests based on their conversations. Some of these turned out to be features we wouldn’t support for privacy reasons or features which were already supported. Filter those out and put the requests alongside detailed usage data that reflects how easy or difficult various aspects of the platform were to use and you have a product manager’s dreama solid basis for data-driven feature prioritization.

Obviously we’re not robots (though we have some rolling around the office) and data only takes you so far. Talking with customers about their experience using and integrating the PYLON platform is key to improving it. So we’ve spent hundreds of hours with our customers discussing their implementation approaches, working together to analyze test datasets, collaborating on the platform’s roadmap, and talking about how to make the data analysis through PYLON more powerful and easier to use.

PYLON already supports a broad range of use cases, from creating more relevant paid and owned content to understanding brand reputation and identifying new target audiences. It’s through this ongoing collaboration with our innovative customers that we’ll continue to power new insights from the social data ecosystem.

The 1.5 release includes features like:

— Common Target Mapping. For ease of use, we now have a common target for filtering stories and engagements with one rule in DataSift’s unified filtering language, Curated Stream Definition Language (CSDL).

— Timezone Handling. We now allow customers to specify which timezone they would like to use in performing time series analysis to create timescale visualizations.

— Capacity Notifications. PYLON customers can now opt to receive notifications when the data volumes in their accounts reach 50%, 90% and 100% of capacity.

— Tags Endpoint. A new API endpoint allows customers to see which VEDO tags have been applied in their recordings.

As for PYLON for Facebook Topic Data, we’ve added data from additional countries. It is now available from 55 countries listed below:

North America

United States


Austria, Belgium, Denmark, Finland, France, Germany, Iceland, Ireland, Italy, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom

Eastern Europe, Middle East, Africa

Bahrain, Bulgaria, Croatia, Cyprus, Czech Republic, Egypt, Estonia, Ghana, Greece, Hungary, Iraq, Israel, Jordan, Kenya, Kuwait, Latvia, Lebanon, Lithuania, Macedonia, Malta, Mauritius, Morocco, Nigeria, Oman, Palestine, Poland, Qatar, Romania, Saudi Arabia, Serbia, Slovakia, Slovenia, South Africa, Tunisia, Turkey, Ukraine, United Arab Emirates

Written by Jay Krall

Jay Krall is a Technical Product Manager for DataSift based in Reading, UK. He's originally from Chicago and likes hot dogs. Connect with Jay on LinkedIn

Share This