SIRS is using Codeit to add value to existing client projects and enhance new business opportunities.
18 October 2018
As part of its research toolkit, SIRS (http://rdicorp.com/sirs/) offers its clients a full coding solution. However, human coding – SIRS’ legacy coding solution - is both relatively time-consuming and costly, depending on the length and complexity of the text being analysed. Some clients are collecting a very high volume of verbatim responses - a single wave of a study could yield 5,000 to 10,000+ comments. The significant time and costs involved in such major coding efforts using traditional, human coding has meant that many research clients were leaving their valuable verbatim responses on the table.
The need for a new solution to coding became very evident when a SIRS client commissioned a project to code open-ended responses generated by their very large - and strategically important - customer survey. The client was keen to surface the broad themes and topics that were expressed within the verbatim responses. The first quarter’s survey generated in the order of 5,000 comments, which SIRS coded using human coders. Although human coding achieved a rate of ~42 responses per hour, the overall coding process (including consistency and quality checks) demanded great precision and high levels of management time. Additionally, the sheer breadth of the coding task generated a codeframe that was highly extended, fragmented and too detailed to generate sufficient meaningful and actionable insight for the client.
Unsurprisingly, the client declined to commission the project as a tracking study, quoting both time and costs. This prompted SIRS to search for a 21st-century coding solution that would not only deliver significantly greater ROI to its clients, but that would also open up a potentially significant new revenue stream.
Motivated by their original client seeking to undertake another major coding exercise (with even greater volumes of verbatim data and much larger bases than the original project), SIRS reviewed all the available coding solutions on the market. Clearly, an alternative to human coding was required. The result of the solutions review was a decision to implement Codeit, primarily because of its combination of in-built coding flexibility and seamless blending of powerful and leading-edge technology tools. As Jazmine Sargeant, Director of Data Management of SIRS puts it: “Codeit allows us the flexibility to fully customise our coding solution so that we can offer solutions that align with client timelines and budgets”
SIRS coding model now utilises Codeit at its heart, as shown in the operational process model below:
In implementing Codeit, SIRS believed that the solution could offer some critical operational advantages:
- Codebook management tools streamline the codeframe by merging codes together
- Machine learning tests the code frame and immediately commences auto-coding
- A human coder-oriented interface handles uncoded responses delivering dramatic improvements in productivity
- Keyboard based code search enables coders to work rapidly and more smoothly in contract to drag and drop approaches
As it has turned out, Codeit has significantly exceeded SIRS’ initial expectations. There have been clear and dramatic operational benefits driven by Codeit’s unique blending of key technology and UI elements. As Codeit’s AI and machine learning processed and analysed ever more data, the successful automated coding rate rose to an unprecedented 80%. Even the small amount of residual coding took far less time to complete, thanks to Codeit’s enhanced coding UI.
On successful completion of the coding project, SIRS calculated that Codeit had delivered an extraordinary 300% overall improvement in coding speed over their previous methods, with additional significant enhancements in client satisfaction and project profitability.
SIRS Coding Manager, Tim Linneman describes Codeit’s impact as “allowing us to utilize a combination of Text Analytics and AI Machine Learning to code large data sets more efficiently. With Text Analytics and AI Coding solutions we are able to train the AI by manually coding portions of the data and then letting the machine code by following the examples. This new coding process is very beneficial to our clients because we’re able to turn around results in about half the time of manual coding. And because we are spending less active hours on the study, we can pass along cost savings to our clients. “
SIRS and Digital Taxonomy will continue to work in close partnership, further refining and enhancing Codeit to deliver even greater performance enhancements.
Back to Blog