Proof of Concept Enables Step-change in Behaviour Data Analysis at Unilever


    Life Sciences Customer Services

    Working closely with Unilever’s Port Sunlight R&D innovation team, Tessella funnelled divergent ideas into a solid engineering prototype, delivering proof of concept of a valuable new behaviour engine.

    Tessella understand Unilever very well, and their combination of very high quality engineering and scientific abilities with a willingness to be on site and work locally with Unilever contacts, meant that the team achieved a very tight integration.

    Matt Reed
    Open Innovation Director, Unilever


    With 400 brands spanning home care, personal care, refreshments and food products, Unilever touches many peoples’ everyday lives in many different ways. Unilever aims to make peoples’ daily lives better, and in developing new products or improving core products they have an increasing focus on helping consumers improve their diets and lifestyles.

    Throughout its history Unilever has developed innovative new products, spending over €1 billion each year in cutting-edge research and development to explore new thinking and techniques; this research, combined with local market knowledge, positions them well to respond to consumers’ changing needs.

    At the end of 2013, a strategic group based at Unilever’s Port Sunlight R&D facility initiated a study with behavioural scientists to create an innovative approach of looking at people’s behaviour. Whereas, historically, a questionnaire-led approach has been more common, the intention here was to actively track participants’ everyday activity levels using a wristband and a mobile app, and to discover if an intervention would change their behaviour for the better.

    This was a great opportunity to fully field test the technical platform that Unilever and Tessella had been developing over the previous three years.

    The Challenges

    In 2010, Unilever’s open innovation team were looking to create a deployable technical platform that would take a holistic approach to monitoring and understanding the four pillars of activity, sleep, mood and nutrition. While other applications were looking at each of these areas in isolation, nothing was available that considered all four areas and their interdependencies.

    Drawing on understandings from quantitative and qualitative research, the plan was to structure and engineer a new prototype behaviour engine. The core of the challenge was to enable a packet of data to be created and collected that could capture an ‘atom of behaviour’, recording where and what had happened – without compromising the privacy of the person who had performed that behaviour.

    Tessella was chosen for the project because of their ability to deliver from end to end, from exploring and developing ideas through to building a working system using an agile process of monthly releases.

    “Tessella’s robust processes are particularly appropriate for this type of project,” said Unilever’s Open Innovation Director, Matt Reed. “Tessella were involved from the beginning in regular meetings to define and develop the principles of operation of the project. Subsequently, with monthly software releases, we were able to review and develop as the project continued, to keep on track and to rapidly get to the essential elements to make the engine work.”

    “Tessella understand Unilever very well, and their combination of very high quality engineering and scientific abilities with a willingness to be on site and work locally with Unilever contacts, meant that the team achieved a very tight integration. Their understanding of the cultural aspects of Unilever meant that the team’s results were accelerated.”

    The Field Test

    Moving forward to 2013, and the first real live test of this prototype engine looked at daily activities, and whether an intervention could encourage people to be more active.

    Working with a local academic partner, data was wirelessly collected as participants went about their usual day. Using the information collected, prompts, hints and tips could be sent out on a daily basis to encourage healthier behaviour.

    Brian Newby, the Unilever Research Scientist heading up this study, said that one of the main benefits of working with Tessella on this project was “[their] innate understanding of the challenge we were facing, and their interest in, and passion for, solving that challenge from the Unilever perspective. Tessella’s willingness to work as part of a team and the rapport generated was instrumental in a successful outcome.”

    As a technical trial of the system, the results were more successful than had been anticipated. Communications between the system and the participants worked extremely well, data uploading and reporting were technically sound and proactive feedback was well received by participants.

    The purpose of the trial had been to test if the infrastructure actually worked in a live situation, and in terms of the daily reporting capability and subject compliance in typical use conditions of a behavioural intervention study, the system was successfully proved.

    Next Steps

    The pilot study has proved the efficacy of the prototype behaviour engine, and having been tested at this level gives Unilever a platform to build on for future trials. Pleased with the proof of concept that Tessella have delivered, Matt Reed added:

    “Unilever are very experienced in running trials, with data collected in batch mode that is subsequently analysed before another collection is run. With this new proven system Tessella have enabled a stepchange, giving future studies the ability to collect and analyse data on the fly in a continuous process that incorporates response and feedback throughout.

    In addition, the foundation work done here will allow more valuable future comparison of data sets, since having set the rules and form of data collection, the more the system is used the more value will reside in the accumulated data.”


    • To determine a data packet that could be sent and received in a compact manner, devised from the huge range of human behaviour data available.
    • To collect information and deliver insight on consumers’ behaviour without compromising on individuals’ privacy.
    • To field test the new technical platform and prove the efficacy of the new data collection engine.


    • Robust data collection and analysis platform with documented rules on data collection established proven standard for future research studies.
    • Data can be continuously collected and interpreted on the fly, shortening time to market for new products.
    • Positive reception by trial participants confirms feasibility of future trials.