Tethr Blog

Maximizing the value of your data science investments

Data scientists are tired of being bogged down by your organization’s lower-value tasks. They didn’t know they signed up to be your go-to reporting person, your call center database expert and at everyone’s beck and call for even the simplest of questions about data. They got into data science to solve complex problems with machine learning algorithms that can make a huge impact on your business. But instead, in many cases, they end up working on basic blocking-and-tackling that could be addressed by other teams if they had the proper tools.

CEOs and CFOs aren’t satisfied with the returns on their data science investments

Despite the initial excitement of getting on board with enticing data science investments, the expectations businesses have are often not being met. One reason that companies aren’t getting the value they hoped is that resources are too often deployed on low-value activities like data cleaning, data structuring and dashboard building.  

Consider this in a conversational analytics setting. In order to surface meaningful business insights from conversational data, companies must convert speech to text using transcription, correct transcription errors to generate accurate data, build machine learning categories to train machines to find insights within the text data, tune machine learning categories to eliminate false positive and negative hits and build dashboards and reports for the business to be able to act on the data to drive meaningful changes in the business. Only after this work is done can data science teams progress in advanced work like building predictive models and scoring algorithms.

Almost all of these steps–aside from the advanced work at the end–is arguably better done by people in the business.

Let experts of the trade be experts of the trade 

Imagine a business is requesting a new insight for its call data. Maybe it’s a wireless company looking for customer feedback about a new product feature the company has just rolled out.  The business team approaches the data scientists and puts in the request. The scientists work on teaching the machine learning platform to recognize when customers are talking about this new feature and a few days later produces a new machine learning category for the business to track. But the business determines that the category is producing a lot of false positives. They ask the data science team to rebuild the category and tune it to focus on the specific thing they’re looking for. Since the data science team isn’t made up of business people, they’re often guessing at what this would sound like in a customer conversation. Back and forth they go for weeks until they are able to achieve accurate results. It isn’t long before the business team needs new dashboards and reports that track this insight, meaning there are more things that the data science team needs to build.  

And the cycle continues. 

Not only can data scientists become insight bottlenecks for businesses, but they also get frustrated when they’re spending all of their time on basic work rather than the more advanced, engaging work they spent years studying for. No one is happy. Leadership feels that the data science investment the company made isn’t paying off and the data scientists themselves can event start looking for work elsewhere.

Empowering the “Citizen Data Scientist

Conversation analytic tools boast a powerful value proposition: to reveal why customers are contacting you, to provide analytics around customer sentiment and how your agents engage with them and the impact of the customer experience on overall loyalty and other key business outcomes. However, AI-powered conversation analytics aren’t all created equally and many aren’t delivering the return on investment. By providing a limited set of out-of-the-box categories, or worse yet–a blank slate on which to build a solution, business are forced to deploy data science teams to make the most out of failed conversation analytics projects.

Leading companies are instead focused on investing in platforms that can 1) provide a significant jump-start to the process though hundreds of proven, tested machine-learning categories all packaged and ready to deploy against their conversation data, and 2) equip practitioners in the business––the citizen data scientists––to do the blocking and tackling. This focus on packaging and ease-of-use frees up the data science team to work on advanced functionality.

Practitioner listening platform to the rescue 

Tethr is the only customer listening platform built for the everyday user. Tethr allows members across the enterprise to build their own reports and dashboards. Users are able to search all of their call data for concepts and keywords on their own. Rather than relying on the data science team anytime additional data is needed, the everyday practitioner is empowered to do so on their own. This frees up data science teams to delegate out the work that is better done by those closer to the business and customers. In return, they can focus on work that is of higher value to the enterprise as a whole.

Information silos be gone

Leadership, sales reps, marketers, call center managers and product teams can all uncover valuable insights inside the call center data with Tethr’s platform. Think of all the time and money wasted when all departments aren’t able to access or interpret the insights coming out of your customer calls and chats. When it comes to improving customer experience, it is critical to have the tools to extract actionable insights. 

Interested in learning more? Check out this video explaining why data scientists love Tethr.

Share