Bhaveshi Agarwal – Fractal

Could you please tell us a bit about yourself and your background?

I did my triple-major undergraduate degree in psychology, sociology, and economics. I didn’t want to let go of either psychology or economics and that’s when I came across behavioural economics, which was a perfect combination of my interests. That’s why I did my Master’s in Behavioural Economics (“BE”) from The London School of Economics. 

After my graduation, I did a three-month fellowship with the NITI-BIU for behavioural insights for public policy, and also an internship with a behavioural start-up in India called BEAST (Behavioural Economics and Science Team). 

In January I started working as a full-time Behavioural Architect at Fractal Analytics and I have been working with tech clients. Whilst my previous experiences provided a smooth transition between academics and the corporate world, I have enjoyed my current role the most since it’s a full-time position so I have felt more in touch with my projects and the team.

Can you tell us more about Fractal as a company and your role within it?

Fractal comprises of three wings – artificial intelligence, data engineering, and design. They try to create an integrated approach for the clients using these three fields. 

As part of the design team, I work with designers to bring about an integrated behavioural science and design approach. Specifically within my role, we try to bring in behavioural insights using primary and secondary research to help clients understand user behaviour and how they can leverage these behavioural insights to adapt their business strategies further.

That sounds so interesting! Could you tell us about any interesting projects that you have worked on?

One of the companies that we are currently working with is very data-driven. One of the most interesting aspects of this project is that we are collaborating with the data analytics team to understand and track user behaviour using numbers and data metrics that the client has already established. So there is no need for primary research! Whilst they have been able to see patterns and trends, they have not been able to pinpoint exactly why these patterns are emerging. That is where we are able to bring in our expertise, show them the value of behavioural science in their current work, and inform them how the focus can be shifted to infer user behaviour without even needing them to carry out any additional research.

What are the main methods of data collection or experimentation that Fractal uses to incorporate BE into their work?

User research can be both quantitative and qualitative. You may be sending out surveys, running user interviews, or possibly setting up an experiment and collecting data to run an analysis. 

One experiment methodology that we use is called ‘ethnolab’, which is a patented method developed by Final Mile, which was acquired by Fractal. Within this, you present a series of scenarios to the user and ask them why the person in the scenario made a certain decision. By inferring why the hypothetical person made a decision, they would automatically think about what they would do themselves in the given situation. Because they are not aware that they are thinking about themselves, this reduces the social desirability bias.

Some clients come in with very tight deadlines so user research may not be possible. In these cases, we rely on secondary research and our expertise to give them something of value. We do tell the clients that this is our best attempt at evaluating the behaviour, however, we cannot pinpoint exactly what is happening. Sometimes the client will be presented with this initially, like what they see, and may then ask us to do user research afterwards. It can be nice to start with secondary research because user research can be time-consuming and very expensive. 

Could you tell us about some tangible changes that you have observed after making use of behavioural interventions in past projects?

During my fellowship with the NITI-BIU, we were trying to look at the current policies by mapping out the entire journey of how a policy is made and implemented. We tried to figure out at which stage could behavioural economics be used to bring about behavioural change. One of the major things I learned at that point was that we underestimate how much change just educating people and giving them the necessary information can make. 

One of the projects’ aim was to see how we could encourage people to conserve more water. People are very aware that conserving water is important but they are unaware of how to do it. They don’t realise that very small things such as turning the tap off when it is leaking can cause an instrumental change, especially if many people across the country are making an effort. The government is currently attempting to change this by producing more information pamphlets which inform the public of small changes they could make to conserve water, which they wouldn’t even have to go out of their routine to do.

Of course, there is a lot more to be done but this is a great first step, and it informs people that they can make a change as individuals.

Could you tell us about some challenges that you have faced whilst trying to incorporate BE into your work?

One of the biggest challenges is that when you try to give clients very heavy BE insights, they don’t know what to do with them. This is when we have to refine our approach by not only providing them with digestible insights but also giving them actionable business strategies. 

Another challenge is that many clients have heard about some BE concepts through mainstream media but they have a very different and simplistic understanding of them. In reality, these concepts are a lot more complex. You have to help these clients unlearn what they know and let them re-learn what you are trying to convey. You do this by feeding them bite-sized chunks. 

Do you have any advice for people who are trying to incorporate BE into their business?

Start small and don’t try to bring about a big change. Small changes can add up to a big change. A lot of the time it is trial and error because something that may have worked for another company may not work for your own, and that is okay! Not all users are the same so one size does not fit all. 

Throughout your career, have specific points about using BE stood out as being particularly enjoyable or surprising to you?

Whenever I learn more about BE, the world starts making more sense. I really enjoy that feeling so whenever I am reading or doing secondary research I tend to get carried away. One of the most interesting things is when you are working with companies whose products you may be personally using. You need to draw a line between viewing the product as a user versus a researcher. You are only one type of user amongst many different types of users. If you research only from your own point of view, you are limiting yourself.

Have you faced any other unexpected difficulties whilst trying to use BE in your work?

One instance was when I was doing my fellowship with the NITI-BIU. Because it was in the public sector and on a large scale, we wouldn’t see immediate results. This can get bothersome after a while as you are doing all this work and you know that it may cause a change after a while but it is not always tangible. You have to keep the faith that what you are doing will add up to something in the future.

When it comes to the private sector, one of the biggest difficulties was knowing how to account for different levels of expertise. You keep revising what you are presenting to the client because they often don’t want high-level insights, they want insights that they can implement easily and immediately. A lot of the time you come up with suggestions that the company may have already tried and tested. Sometimes you don’t have access to information about exactly how they tried and tested it, and whether it is the same as the method that you are suggesting. Sometimes we try to navigate through this by talking about what they have already tried prior to starting the research but sometimes research is missed out or they may not be willing to share the information. In these cases, you either have to try to refine what you are suggesting or suggest something different altogether. 

Are there any final thoughts that you wanted to share with our readers?

One piece of advice would be that the scope of BE is growing because more and more companies are realising its value. So don’t only look for mainstream BE roles. Many companies have roles which may not be labelled as BE roles but may still be very BE-centric. Have an open mind because the scope is increasing but you are the one who can define that scope. Some examples of these may be market research roles, data analyst roles, or behaviour change roles in the public sector.