Use behavioural data to drive behavioural change

“The only way to cause behavioral change, is by taking decisions based on what we learn from how people act today,” says Aurore Belfrage, Tech Diplomat, Investor and Sustainable Tech Strategist. This is especially true to solve the larger world problems such as climate change, but can also be applied to regular business behavior. The use of so called “behavioral data” is unfortunately not integrated enough with decision- and policy makers.

Policy makers and captains of industry have a tricky balancing act: how to make a maximum of impact on climate change, whilst causing a minimum of disruption to society or their business. The problem is that they too often still rely on old metrics and data collection techniques instead of using new technologies. All too often, there is a reliance on surveys or news paper articles, instead of investigating data on how people actually behave.

“There is a reluctance by decision makers to adopt new technology for getting underlying data,”explains Aurore, “often this is out of fear that it contradicts earlier data and it would look like their earlier decisions were wrong. Unfortunately this leaves out massive opportunities on making change go faster, as behavioral data might indicate that people are ready for change or adapt quicker to new ways of doing things.”

Using these kinds of data such as search, geo-location or even sentiment analysis, has become a lot easier to collect, but mainly also to process using Artificial Intelligence. Especially search data has become most accessible, is in real time and can often be filtered on various variables such as location and volume.

“Search is data is great as it immediately gives a clear overview of what is literally at the top of people’s minds. You can easily read the type of concerns or solutions they are thinking of, or what questions they are looking to resolve,” continues Aurore. “By investigating those trends we can make predictive models for the near future and see how we can influence behaviour in the next 6-12 or 18months.”

That same type of search data, especially combined with sentiment, can also be used to forecast behavioral change such as purchase behavior or the attitude towards brands. The important thing in such case is to select the right search terms.

“Starting with using search data like for Share of Search is pretty straight forward, but does need a bit of thought,” says Fred Pirenne, Chief Science Officer at My Telescope,”It’s best to think about the search terms as a consumer, and how they would launch the search. Our research has shown that given the trend on the right search term, you can easily predict future purchase behaviour and predict market share trends.”

In order to get started with the Share of Search, one can start as simple as looking at Google trends, or use the free accounts at My Telescope. Here are 4 quick steps to get started:

1. Know what you want to investigate:
   By having a clear idea on what it is you want to know more about, it becomes easier to find select the right search terms for the task.

2. Start broad, but not generic:
   By using a broad start it’s easier to get also responses in search terms or volumes that can surprise you and lead you onto an entirely new insight. It’s important though that if your brand or product carries a generic name like a last name or description, that you not only use that. You will get too much unrelated data from it, giving inferior results

3. Think outside the competition box:
   Very often you will have a fixed set of competitors, but don’t forget that consumers don’t have the same mindset. So it’s important that you think of your competitors as all those who can fulfill your customers' needs. For example, if you sell garden equipment, your competitor might not be a competitive store or garden equipment brand, but just gardening services in general.

4. Trends are more important than absolute numbers:
   It’s easy to get lost in the numbers, but prediction of behavioral change is best explained by looking at the general trends.

Should the data contradict some earlier data that you have received, dare to look at it more in depth and check if you need to make any adjustments.