Rentrak Work Values

Daily Duties at Rentrak:

I mostly worked on the middle-layer; the back-end received content views from a variety of TV providers (such as AT&T U-Verse and Dish), in record-based or XML files, and loaded summarized data into a central Postgres DB. My team’s job was to build HTML, Excel, and PDF reporting based on that DB. We used an object-oriented reporting system, with page objects for layout, built from re-usable self-rendering component objects, and report objects for querying the database, built from re-usable SQL generation code and post-query Perl calculation logic. Generally speaking, pages used the same layout for HTML, Excel, and PDF, although we had the ability to customize the layout per format when dictated by business requirements (e.g., including graphs and charts on HTML and PDF but not (unfortunately) Excel). HTML was generated directly using Mason, and Excel and PDF were generated using Mason-generated XML formats (in-house and XSL/FO, respectively) and a post-processing step. Our network customers usually preferred to print our generated Excel files to give to advertisers and ad agencies, so a lot of attention was paid in the Excel generation to balancing producing a good appearance when the files were printed against allowing the networks to use the numbers for custom calculations on their own computers. SQL generation and calculations were driven by the metrics on a report and the filtering/aggregation level of the report and its underlying summary table; for example, we had a module that took a summary table, filter set, and desired aggregation level and determined how to aggregate or average the average audience numbers on the table to get the number on the report. For example, the network/day summary table might be used to calculate average audience for a single network and day, for a single network and a whole week, or for a set of networks and a single day; the calculation framework would automatically generate SQL to take the average audience value off the table in the first case, to average it across days in the second, and to sum it across networks in the third. Our system provided three major benefits to our customers. Because our data was based on vastly more data points than Nielsen’s, we were able to provide ratings to many smaller low-rated cable networks more cheaply than Nielsen could. Additionally, because of the larger sample size, our ratings were more stable than Nielsen’s were, for very small-ratings networks; for example, one network switched to our data after discovering that they’d lost a large fraction of their ratings because Nielsen removed one (only one) of that network’s viewers from its household sample. Finally, with our larger data we were able to define smaller and more precise demographics; for example, using our data CNBC was able to start selling $100k+ HH Income rating points to ad agencies, which for their demographics they were able to make more money off than they could for Nielsen’s demographics.

What they like about Rentrak:

You give high priority to hiring firms with prestige. Market reputation and prestige of a firm may be determined through on-line research, and conversations with individuals working in your target industry. Note that firms "in the public eye" receive a great deal more critical publicity and attention than those operating with more anonymity. So as you assess your employment opportunities, be sure to seek out the positive aspects of the firm's reputation in addition to noting negative publicity the company may have received.

Tags

Postgres, SQL, PostgreSQL, Linux, Ruby, Rails, Python, Perl, Catalyst, Moose, Software Engineer, Web Application, Web Services

Skills

Perl, MVC, Model-View-Controller, Object-Oriented Programming, Web Development, HTML, CSS, Postgres, SQL, git

These are some of the questions we asked our climbers about their experiences with Rentrak: