March 18, 2020
Imagine asking 1,000 marketers two questions:
1. “Do you want an NPS in line with Amazon and Netflix?” Odds are 1,000 of them would say yes
2. “What research methodology would enable this?” Odds are 0 of them would reply machine learning and traditional qualitative research
However, the story of Colt Technology (a network & communications provider, operating in 212 cities, an annual turnover of $2b) could change responses to question two in the future.
12 months ago Colt’s NPS was 41. The telecom sector’s benchmark is 15 (B2B Satmetrix) – clearly, Colt were doing many things right. But Colt are a daring business and wanted to aim higher – their mission is to become the most customer-oriented business in the industry. To do this, they strive to be continuously progressive in how they speak to their customers and also how they look to improve their Customer Experience (CX).
However, this is no small task. Colt is one of the world’s largest network & communications providers, dealing with thousands of customers per day – plus, depending on where the customer is in their relationship with Colt will dictate which part of the business they’re engaging with – either, Sales, Delivery, Network Operations or Customer Services.
Qualitative research could help Colt better understand how to deliver against this complexity vs. yet more quantitative research in what is a data-rich organization. However, Colt’s data wealth provided a hurdle for making qualitative research impactful – qualitative research was heavily undervalued.
This made the research challenge twofold:
This twofold challenge was answered in three parts:
1. Technology: Machine Learning
Every quarter, Colt’s NPS surveys generate 100,000 open-ended responses. Fred Reicheld – founder of NPS – says open-ended responses are the most valuable – but underused – element of the NPS methodology. But where to begin mining such a vast resource of text-based information?
Clearly there was too much information to manually interpret. Therefore, Northstar built Colt’s very own machine learning algorithm called CAST – which mines Colt’s NPS open-ended responses and helps the business understand the following:
2. Tradition: Old School Interviews
However, CAST wouldn’t be able to achieve any of the above, without the valuable input from traditional qualitative interviews. Across our many conversations with their customers, our researchers were able to:
3. Telling: Culturally Relevant Insight Communication
Effectively communicating qualitative insight was pivotal in allowing Colt to optimize their customer experience. This was challenging as we had to make information stand-out in a data-driven business.
The answer was to develop an insight portal that would fit vs. Colt’s internal communication culture, developing the UX experience they wanted (i.e. optimum pathways to content, content hierarchs), and providing a multi-media experience (audio clips, summary reports & concise consolidated reports).
In short: embrace technology, but never underestimate the power of traditional methods. And never forget – effective communication is half the insight professional’s battle. However, successfully combining the three can allow a business to meet its objectives, no matter how lofty.