Haystack App


Sensor Data and Artificial Intelligence: measure better, predict better!

Research companies have been using classic questionnaires for decades – and with varying success. It doesn’t mean they’re perfect. With everything we know now, and modern resources like mobile that are available, it’s time to change the way we question people.

At Haystack, we’re constantly looking for ways to improve our research methods in The Next Lab. We’re trying to optimize the way we reach consumers, the way we ask them questions and the way we process data. We’ve developed a mobile app that integrates sensor data (location, movement, time, biometrics) and artificial intelligence, so we can gain a lot more information from respondents without asking them questions.

Capturing passive data to profile people

We did a pilot study with 200 participants, who agreed to share their sensor data. That allowed us to find out a lot more about the participants. One simple example: we could see who commutes to work by taking the train, and who uses his bike or his car and how long it took them. Not by asking, but by measuring. This data enabled us to create consumer profiles: people who sport regularly, people who spend twelve hours a day behind their desk, and so on. We can even determine people’s driving style or find out where they do most of their grocery shopping.

Predicting consumers behavior, measuring in the moment

But we can do much more. We can detect and even predict consumer behavior – like where and when people will go to the supermarket. If we have that information, we can ask them questions about their shopping behavior right on the spot. And answers given in the right context, are likely to provide more accurate information.

So if you want to ask people about their breakfast habits, do it in the morning and not at the end of the day. The right context is crucial. As Gary Vaynerchuk puts it: “Content (survey) is King, but context (timing, place…) is God”.

Less questions, more reliable information

By using AI and sensor data we are receiving a lot more information from our respondents without having to ask more questions. And the questions we do ask, are asked at the best possible moment and in the right context. As we’re measuring exact behavior, like sporting frequency or social activity, we’re taking out the factor of social desirability (the tendency to give answers other people want to hear). That makes the information we receive, far more reliable.

AI and sensor data give us the chance to gain more consumer insights and create new business opportunities. Do you want in on these opportunities? Don’t hesitate to reach out to our expert Jasper Scheir.