Creating a Tech Start-Up Research Ecosystem for True Insights

Author: Marie Wolfe, Unilever July 15, 2015


Although the market research industry was founded over ninety years ago, there still remains a large opportunity for methods that are able to capture the gap between what consumers SAY they do and what they ACTUALLY do. Unilever and weseethrough partnered to tackle this long standing challenge by leveraging the latest advances in wearable technology and an innovative video indexing platform. Going beyond what others have done with Google Glass, weseethrough programmed advanced applications to automatically trigger interactions with consumers through the glasses during the research. The resulting videos and responses were then processed through weseethrough’s cutting edge video indexing platform, finally enabling video as a scalable research technique. Over the last year, Unilever has conducted eight varied pilot applications of the methodology across 20 countries. In each pilot, the method uncovered breakthrough insights that we not previously discoverable from explicit response alone or more basic observational technology. The creation of this methodology was made possible through a new ecosystem approach. In this ecosystem approach, Unilever Consumer & Market Insights (CMI) provided “research coaches”, mentoring and experimental pilot opportunities to help the start-up bridge into market research and thrive in a large corporate organization.



Traditional research fails to capture the gap between what people SAY they do versus what they actually DO. However, understanding this gap is critical to help Unilever truly understand the consumers it serves. In fact, recent neuroscientists have discovered a molecule in the brain that actively makes humans forget unimportant details of their day to protect them from going insane (N. Hadziselimovic, 2014). This likely means that the brain is actively blocking consumers from recalling key details of their routines and marketing interactions.

Traditional research methodologies rely on memory, leading to large error gaps, especially in Unilever’s consumer packaged goods categories. For example, consumers struggle to fully recall meals consumed and ingredients used in the last week, but this information is critical for Unilever to truly comprehend consumers’ habits and needs. It is unlikely that these details can ever be fully and accurately recounted through explicit questioning.

Prior attempts to understand this gap have fallen into three main categories:

1. In home field observer – Placing a researcher in home to observe the consumer’s habits. Downsides include a potential to create biases in consumer behavior, high cost, and difficulty comparing the results across respondents.

2. Diary Logging – Although closer to the moment of use, the methodology is more invasive and labor intensive for consumers. Due to the structured nature of the questions, it does not leave room to discover habits and behaviors outside of the dimensions already set in the diary.

3. Camera Placement – Though closer to true observational research, camera placement is laborious to set up, has limited mobility and coverage in the household. In prior studies, there was no efficient way to process the massive amount of video hours coming out of constantly streaming cameras. Static mounted cameras were also unable to capture the details needed for many of Unilever’s categories, such as seeing the exact ingredients used during meal preparation.

The new class of research created between Unilever Consumer & Market Insights (CMI) and weseethrough leverages advances in technology to overcome past barriers in four key ways:

1. Creating technology at scale that captures true consumer behaviors non-invasively with high quality footage

2. Creating capabilities to analyze large quantities of video data and other sensor data efficiently

3. Going far beyond basic technical applications done by others by directly integrating the start-up’s engineers into the research process

4. Forming a sheltered support system and learning environment for start-ups within the classic research and corporate environment so the start-ups can thrive

The discovery, design and execution of this new class of research was made possible by an innovative Unilever CMI ecosystem approach to support technology start-ups who are bridging into market research and competing against large, established companies. Unilever CMI is likely the only market research client creating such a robust ecosystem that helps start-ups build their own market research capabilities to disrupt the industry.

Over the last year, Unilever has conducted pilots of eight different business applications of the methodology across 20 countries. In each pilot, the method uncovered breakthrough insights and understanding that were not previously discoverable through explicit response or basic observational technology.



A year ago, the industry was beginning to play with the application of wearables to understand consumers’ behaviors, but the applications felt superficial and more for bragging purposes than leading to new, actionable insights.

Given that nothing existed to meet Unilever’s “SAY vs DO” brief, Unilever CMI sought out talent from the technology industry to build the future of research together. Alex Foster, now CEO of weseethrough, was then the CEO of Glass Fit, a virtual running partner on Google Glass and introduced to Unilever as a talented entrepreneur with no research experience. Unilever briefed him on the challenge and a week later his team had a full dashboard programmed, a new name and logo, and weseethrough’s journey into market research began.



Unilever strongly believes that it cannot wait for the industry or technology companies to come to CMI with the future of consumer research. As one of the largest buyers of market research in the world and a company obsessed with innovation, Unilever CMI feels that it is its responsibility to craft the future with talented forward-looking thinkers in partnership. Unilever designed the insight ecosystem to support entrepreneurs with a great technology and idea, but who may not have experience in holistic study design, storytelling, and research best practices. Unilever invests significant resource into “research coaches” for the start-ups. With the right coaching and mentoring, Unilever CMI has found that start-ups are eager to learn these areas with a startling energy and speed. With this positive mindset, the start-ups are quickly able to build up their skills to compete against long time practitioners.

The accelerated pilots are also treated as beta experiments, where the emphasis is on trying new, challenging business questions. The pilots are conducted in an environment that is focused on learning, experimenting and iterating, rather than a traditional client and agency service model. Additionally, the CMI managers leading the pilots provided overall coaching and mentoring to the start-up throughout the process. In return, they learned valuable lessons from weseethrough about how to think like a start-up to tackle age old challenges.



Unilever has applied weseethrough’s talents against a wide range of business questions. Those business questions fall into five buckets, which are applicable across all of Unilever’s brands and categories:
1. Understand how consumers really use new product prototypes in a natural home setting

2. Accurately track media consumption habits

3. Capture true habits, experiences and complex routines

4. View the true shopper experience in-store and how they interact with point of sale materials

5. Map the full consumer journey from in store to in home consumption

The team leveraged weseethrough to tackle these business questions in a way never previously possible by combining two technologies:

  • Advanced Google Glass Programming – The study used the glasses not just to record, but push questions through the glass at key times that are triggered by pioneering programming
  • Advanced Video Analytics – An innovative video indexing platform allowed analysis of the captured video and responses at a scale, speed and cost not previously possible

Methodology Structure:

1. Respondents were recruited following Unilever’s standard recruitment criteria.

2. Local field agents delivered a pair of Google Glass and any other required technology or stimulus (proximity beacons, product prototypes, etc.). The field agents were trained to speak with and ensure consumers felt comfortable with the glasses and placed any proximity beacons correctly.

3. Respondents were asked to wear the glasses as they performed their normal day to day routine. Depending on the objectives, video was captured constantly or only triggered during key moments to preserve battery life.

4. Where appropriate, custom questions were visually displayed on the glass and read out loud through the bone conductive speakers in the local language. weseethrough’s technology is able to trigger questions and reminders to consumers automatically by beacon proximity, geolocation, audio recognition, visual recognition (identifies the consumers have come in contact with a brand, object, product, or piece of advertising) and on a timed basis (i.e. reminders to speak out loud every five minutes, etc.). Where Wi-Fi allows, weseethrough can watch a live feed of the video from the consumers’ glasses and push questions live from a remote location.

5. The respondents answered the questions verbally, which was captured by the glass. Their answers were automatically tagged for ease of retrieval later.

6. Later in the week, the respondents participated in a virtual interview to test their memory of the subject being tested and review key parts of the footage for further questioning.

7. Realizing the massive amount of video coming from the studies, weseethrough invented a cutting edge video indexing platform during the pilots. The dashboard is capable of processing a large amount of video with advanced tagging, translation and clipping. This is done at a very fast rate and at a fraction of the cost of other video competitors, finally enabling video as a scalable research technique.



For a company that has spent millions of euros over the years on traditional, in-depth habits and experience studies, the breadth of new learnings coming out of these pilots was transformative. The respondents quickly forgot they were wearing the Google Glass, similar to what occurs when wearing a pair of ordinary eye glasses, allowing the capture of organic moments.

While each study had its own unique breakthrough insights, this methodology consistently delivered new learning in five buckets:

1. Consumers consistently underestimated time spent – Whether recalling time spent on media consumption or house cleaning, consumers consistently and vastly underestimated the amount of time they spent. Through the new technology, Unilever captured that the actual time spent was at least 40% higher than claimed 70% of the time. Some estimates varied as much of as 350% off of a solid base number.

2. Product use and amount recall is highly inaccurate – Respondents’ recall about the products and the amounts of each product that they used were highly inaccurate. For example, consumers in Brazil claimed to use “a little bit” of a product, but the video footage captured them splashing it around the entire house in large quantities. Similarly, in meal preparation, consumers claimed to use “a bit of aromatic” while pouring half of the bottle into the dish. Some of the products
consumer were recorded using were fully forgotten in the follow up interview, even when directly questioned.

3. Unilever Sustainable Living Plan opportunities are abundant – This approach revealed numerous new opportunities for Unilever to help drive sustainability. For example, the amount of water used around the house during the five hour cleaning routine in Brazil was shocking given the water crisis. This learning spurred the team to further develop and support water conservation advertising that is now on air in Brazil. New insights around product waste and food safety were
also uncovered during the studies.

4. Unobtrusive nature of the glasses leads to new empathy – The footage through the eye-glass perspective and the high quality audio capture of consumers speaking to themselves, others and instinctual audio cues (sighs, yawns, etc.) brought a new depth of understanding to the consumer experience. The feelings of loneliness, frustration or boredom are rarely captured when an interviewer is present or when the respondent is answering an explicit questionnaire. However, these feelings were clearly captured in the video footage and audio capture of the consumers.

5. Memory recall is weak, even when footage clearly indicates something was “noted” – Time and again, footage clearly proved that the consumer looked directly at something, yet had no memory of it afterwards. Even more extreme, consumers also failed to recall certain actions that they physically completed during their routine. The dynamic video and audio captures prove to be a crucial component of understanding consumers.

Overall, this research innovation has transformed Unilever’s research methodologies across a range of business questions. Unilever is now in the process of scaling the methodology globally across categories and regions. It is considered an important next wave in the future of research: a true breakthrough to finally understand the opportunities in the gap between what consumers SAY versus what they actually DO at scale.