CASE STUDY This case study was revised by Google for the Google Analytics Certified Partner program application.

Advanced Integrations & Reporting: Moderating User Generated Content in Real Time and Assisting Executive Decisions. With over ​100M sessions per month​, Taringa is the second largest social network within Latin America and Spain. Their main goal is to encourage people to create great content and share ideas and knowledge. Since they are publishers, post moderation is a key factor in their business model

Technical Ability Taringa moderators​ were having trouble prioritizing content to moderate as they didn’t have any real time viewer feedback in their own platform. Since their ​moderation rate was constant,​ it was important for them to know what new content was more popular in order to moderate it first. By using the ​Real-Time Google Analytics API​ and combining it with Google Sheets​ and​ Scripts​, we created a compact and easy to use solution that aided Taringa moderators in their arduous task by prioritizing on the fly the new content that was being created by users (Fig. 1).

About Taringa ❏ Taringa

is the​ second largest social network​ within Latin America and Spain driven by user generated content ❏ Buenos Aires, Argentina ❏ www.hotgo.tv

​About Digodat ❏ Digodat

is a Google Partner y Google Analytics Certified Partner. It’s team provides refreshing insight into client´s data y finds solutions that generate value. ❏ Latin America: Buenos Aires, Argentina. ❏ Europe:: Lisbon, Portugal ❏ www.digodat.com

Goals ❏ Provide

a tool for the website administrators to prioritize content in need of moderation. ❏ Aid Taringa’s executives in makign strategic decisions.

Approach ❏ Build

custom dashboards to visualize what content is in greater need of moderation combining the Google Analytics API with Google Sheets and Scripts ❏ Integratiog

Results ❏A

process that consumed hours of manual labor every day is now automatic.

Figure 1:​ New content real-time feed sorted by popularity.

The first iteration of the tool extracted the data using the Google Analytics API, removed all the posts URLs that were already

❏ Better

curation of the website’s content which enhances brand image

moderated by comparing this list with a moderated posts list exported from Taringa servers and showed the information in a Google spreadsheet. The live component came from a 30-minute refresh trigger. After a few days of testing, we realized that some long tail posts were getting considerable traffic over time, yet predictably they never showed up first in the real-time feed as their popularity grew slowly but steadily. These had to be regarded as first priority posts too, so we created a Google Sheet that ran with an own customized version of a 3rd party plug-in (Google Analytics Magic) that allowed us to set up and run multiple queries to obtain and produce the page ranks we needed (i.e. most viewed pages in general, most viewed pages in Desktop view in Argentina, most viewed pages in Mobile in Argentina, and so on - there were 35 different queries). Finally, we came up with a​ unique score of traffic relevance by weighing posts’ pageviews with their average eCPMs​ (Fig.2).

Figure 2:​ Final version of scoring tool to assign moderation priority to posts in real time.

Analytical Ability Taringa directors are specially interested in​ data overviews ​as they are extremely useful in presenting an overall picture of the site’s engagement, reach, popularity, etcetera. When we started working with them they told us they were looking for ways of accessing summarized and aggregated data variations in a more condensed

way, so that they wouldn’t have to navigate through many menus and compare data from different screens. These concerns led us to elaborate and ​prepare a simple but powerful dashboard using, again, Google Sheets, Scripts and the GA API.​ The idea was to take into account the main metrics and dimensions that Taringa execs considered most important so that they would be able to quickly answer things like ‘why did our unique visitors go up/down?’, ‘did something relevant happen in a specific country or source?’, ‘is today’s variation normal?’, ‘does this happen every year?’, and so on (Fig. 3).

Figure 3​: Customized metric variation dashboard.

The dashboard panels ​bring together different relative and absolute variations organized ​by countries most relevant to Taringa, and also broken down by source type. On top of the numbers, there’s an automatic color grading​ that quickly indicates the​ sentiment and intensity ​of that variation based on the semantics of the visualized metric. To put variations in context, the client can scroll down from day variations and access greater period comparisons (week and year). It was also necessary to show fluctuations in terms of device category. It was also important to see the analytics data through device category, so there are tabs in each panel which allow that segmentation. Moreover, to correctly weigh relative variation there are the corresponding absolute grids.

Business Impact Content moderation in Taringa is fundamental not only in terms of protecting the community and its participants, but also with regards to complying to different policies and terms required by other services such as Search Engines and Ad Networks. Penalties or suspensions in these matters would result in a critical blow to the

client’s business model, so the use of real time analytics data to help content filtering and sanitizing remains an utter priority to them. The custom dashboard was a huge time saver for Taringa headquarters because what previously demanded them hours of diving into the analytics reports and making their own manual summaries became an automatic streamlined process with a clear and didactic interface. They were able to quickly and daily have actionable insights and use their time on taking adequate measures rather than coming up with the right data.

Find more case studies and information on Digital Analytcis at ​www.digodat.com​ ​• ​[email protected] This document is property of Digodat S.A.