This paper discusses harnessing omnichannel customer feedback data using deep learning, and integrated with NPS survey data to deliver superior customer experience (CX). While there is lot of work done and tool availability in this field for English, there is a lack of algorithms and research done for local languages, such as Bahasa Indonesian. We describe our Quasi Recurrent Neural Network (QRNN) based approach that we used on customer omnichannel feedback for topic modelling, priority prediction and sentiment classification. We further demonstrate the use of our work for multiple clients in Indonesia, in order to set up an advanced CX monitoring and improvement programme (CXM). We showcase our holistic approach to customer experience, integrating omnichannel customer feedback data with NLP-based predictions. We also showcase two cases where we deployed our intelligent CXM tool and set up intelligent conversation monitoring systems to predict customers at the verge of churn-out, beforehand.
When you have the Thunbergs of the world to the Trumps, and many in between, how do you find a common ground for sustainability to win?
Accurate recruitment has been an age-old problem in the world of quantitative research. Think about the last seven days: How many hours did you spend watching Netflix? What is the total amount of money that you spent buying groceries, both offline and online? How many days ago did you last open an app that you do not open every day? The human mind falters in recalling the details of actions taken today, let alone in the last week or month. Nonetheless, traditional recruitment still relies heavily on the user's claim of her category and product usage, purchase history, etc., which leads to inaccurate targeting. An error in recruitment can lead to a much higher gap between the derived insights and the truth. This problem is accentuated even further for mobile-first brands where 'micro-segments' are based on not one but multiple parameters, like purchase history, wallet size, dormancy, product category, etc. In cases where the client's database is used to connect with users accurately (through email/SMS/call), the outcome ends up being inefficient in terms of scale, investments and/or timelines. Can technology really help in accurately recruiting users on the basis of all these parameters with zero margin of error, and complement the insights collected through claimed research at scale, as well as in a cost-efficient manner? This paper demonstrates how a mobile-first brand and its research partner came together to solve the micro-segment recruitment of its app users, in order to solve the core problem of user retention.
This paper outlines the challenges of conducting qualitative research in the aftermath of a national tragedy. As the circumstances in which the research was conducted was traumatic for all involved, intense emotion work was required, especially in the recruitment of participants and moderation of discussions. The paper demonstrates that conducting research on subjects that are socially and culturally sensitive in nature can be emotionally demanding for a researcher, and may have a powerful personal impact on them as individuals. We bring to light how emotional skills become central in such studies and are as important as the intellectual and cognitive skills utilised - if not more. Qualitative researchers need to be resilient in not only managing the emotions of participants in a study, but also at managing the emotions of other stakeholders along with their own emotions.
This time, it will not be the AI winters we saw in the 1970s and 1990s, but a human winter, where the very relevance of human effort is being called into question. There is an irony inherent in this- developments in AI started to gain a strong momentum when an approach, modelled on the workings of the human brain, was adopted. This is now expected to make us obsolete! Further, as the discourse about the need for us to change and adapt to the new reality continues to gain momentum, I want to make the case that it is not so much by changing, but by staying true to what makes us human, that we will ride the AI wave; in fact, the success or failure of AI will depend on how human we remain.
90% of people in Asia want brands to do something about the issues they care about. But what do they care about? And what exactly do they want brands to do? Let us tell you more.
Google and Nielsen came together to develop a way to better understand how consumers go about researching and buying mobile phones. With the internet in the palms of their hands, consumers interact a lot with the medium before considering and/or evaluating or purchasing any product. We wanted to capture this interaction and how this influences the path to purchase for the category. The goal was to reduce reliance on claim-based research techniques, especially owing to the complexity of the digital touch-points to mobile phone purchases, and move to observation-based research.
In this paper, we will seek to outline:- The specific problem facing us in Japan;- Why it was important to adopt a new method to derive answers;- The results of our research and what new insights we garnered;- How we went about maximizing the impact of our insights;- The bottom-line impact to Japanese business.
Getting the right pricing strategy is very critical for CPG companies - more so in the alcoholic beverages (AlcoBev) category in India, because AlcoBev is a highly regulated industry subject to high taxes coupled with restrictions on distribution and mass media communications. This renders it very price sensitive and highly competitive. At Diageo India, we have developed an integrated brand pricing framework that encompasses the following dimensions: (i) consumers' current sensitivity to brand price; (ii) behavioural equity measure (total $ value that is the sum of functional and emotional benefits provided to the consumer by the brand, over and above price); and (iii) drivers of equity. This holistic framework enables maneuvering pricing in the market for the short-term while getting the right balance between price and total brand value, as perceived by consumers in the longer run and also driving actionability in terms of building up equity.
This paper looks at AI-powered voice surveys as an alternative survey format for commercial research. Specifically, the paper is looking at whether this type of alternative format can be more effective at capturing a broader more inclusive spectrum of respondents, including disempowered members of society. To explore the effectiveness of the voice survey format, this research was conducted in an experimental two-by-two design, with the traditional online survey as the comparison. This paper documents the challenges of utilizing an AI-powered voice survey for research and determines what some of the benefits are of this alternative survey format.
Bad ads create bad experiences for consumers and negatively impact your brand. It is critical to know what creative element or pattern can help drive brand performance and direct response. However, creative A/B tests or lab tests with facial expression analysis are expensive and not scalable. In this research, we used Google Cloud APIs to compute thousands of features from large creative samples to conduct creative meta-analysis. We added those features from machine learning and computer vision with human encoding elements.
There is an interesting management quote: Do not expect the output to change if the inputs remain the same. This is true for organisational transformation as well as transformation of the insights and analytics industry. This paper examines this aphorism in the context of volume forecasting for ecommerce launches, looked at in the specific context of a vitamin and mineral supplement launch.