Have you ever conducted innovation research online and found yourself in a situation where you don’t entirely trust what consumer feedback is telling you? Many of us have had to deal with overstated interest in a new product. Uncovering both rational and emotional needs is vital for new product development (NPD) strategies to accurately size the unmet need or opportunity. While traditional qual techniques can uncover emotions, the results can’t easily be scaled. Alternatively, quant research can deliver stated emotions, but lack depth.

Together with Johnson & Johnson, we developed a new hybrid Qual-Quant-AI online research approach for early-stage NPD research. By using a voice analytics tool, we can analyze ‘how’ people communicate their needs, attitudes and interest, to better uncover emotions for more effective innovation strategies. The voice analytics AI tool we selected, audEERING, can detect emotions from voice. In an online survey with voice input a hypothetical ‘Smart Health Band’ was evaluated by consumers. We 1) analyzed the content of responses to look at what people said and 2) using the voice AI tool analyzed how they said it and which emotions were present.

The content analysis - the what was said - showed high interest in the concept. However, the voice AI delivered additional, unexpected results: while still high, the implicit or subconscious interest was not as high as the content analysis had suggested. Analyzing the emotions delivered through voice showed a more realistic level of interest in the product.

Especially for new products it is essential to get accurate estimates for the product’s market potential that are not overstated. The additional voice element helped Johnson & Johnson calibrate the interest in the product to show how much of it was genuine, and in consequence make better informed choices about future strategy. In addition, the emotional analysis uncovered differences in gender reactions. These segmentation insights could prove valuable for future marketing and communications strategies.

Relevance & Research Question:

The benefits of using VR prototypes for higher-cost categories in innovation and design projects are well known – logistics, greater modification flexibility, lower cost, virtual in-store and competitive choice scenarios.

With VR technology becoming more sophisticated and user-friendly, how can market research further benefit from it?

We will present a case study on a new fridge-freezer concept with Electrolux and a leading VR specialist company.

Methods & Data:

54 VR explorations across three countries - Germany, Italy, Sweden - accompanied by a total of 9 in-depth focus groups.

Results:

Across all age groups the VR part worked very well, both as a stand-alone and in combination with focus groups. Following outputs were key benefits:

Usability: Detailed feedback was gathered – particularly in comparison to fragile 3-D renderings.

Immersive: With prototypes seeming so real, we received more spontaneous and little post-rationalizing feedback.

Curiosity: A thrilling tech factor led to higher engagement.

Involvement: Fridges – a potentially lower interest category – enjoyed higher levels of excitement.

Playfulness: Tasks were treated more like games.

Flexibility & Speed: 3D renderings could be changed from one fieldwork session to the next.

Democracy: VR created equal conditions for prototype and comparison device.

Realistic Environment: Simulating a realistic shop floor atmosphere including competitive products was made possible.

Focus Groups + VR:

Higher participant focus due to the VR experience– everyone was highly engaged.

Higher attention to detail due to the amount of time spent with the prototypes during the VR experience!

There was nevertheless a strong creative dynamic present in the F2F groups.

Research moderation expertise is needed to manage “digital excitement” –managing overexcitement that leads them to jumping from one design aspect to the next.

Added Value:

The VR approach was more cost-efficient, more environmentally friendly, offered higher flexibility and in the end enabled deeper and more valuable insights – particularly with regards to usability!

Benefits of VR continued after the research: designers were more open to implementing the design changes as they felt less "emotionally attached" to the stimulus (compared to physical models). Our results were thus received more openly and applied without hesitation.

Relevance & Research Question: For large companies it has become increasingly difficult to measure marketing effectiveness across a multitude of media types, offline- and online channels and campaigns. It has also become more important to have a consolidated view of all marketing and non-marketing activities. MediaMarktSaturn and Neustar set out to answer the question how much incremental value is generated by 120 marketing activities of MediaMarktSaturn, many of which are happening simultaneously.

Methods & Data: The solution uses MMM (Marketing-Mix Modeling) and MTA (Multi-Touch Attribution) statistical models to analyze the data. It integrates a huge variety of data types from sales volume, spends and online tracking data to survey-based funnel KPIs, weather data and store traffic counts. To account for the complexity of the data, a hierarchical Bayesian approach is used.

Results: The effects of marketing activities on brand KPIs, web and store traffic and online and offline sales are being analyzed coherently. It can be shown that different campaign types and media touchpoints influence these KPIs in different ways, opening the opportunity to optimize on specific campaign goals. Survey-based market research results can proxy long-term brand health in the process. The measured efficiency of media types supports a balanced media mix, moving towards the use of online channels, however, it also shows that media types like out-of-home and radio still play a strong part.

Added Value: The study proves that it is possible to measure the incrementality of offline and online marketing activities in a non-experimental environment. It allows MediaMarktSaturn to compare all marketing activities on the basis of ROI and optimize its media budget accordingly. The introduction of survey-based brand kpis provides the basis for steering multiple outcomes. Last but not least the study supplies further evidence of the utility of hierarchical Bayesian methods in tackling imperfect and highly differentiated data.