Our data comes from real-world internet users who’ve chosen to complete our surveys. This people-centric approach gives us a holistic view of consumer behavior across a range of categories that other approaches struggle to match. It also means there’s a degree of subjectivity to our data, as respondents will sometimes interpret questions and answer options differently.
While our questions are designed to be as clear and easy to answer as possible, it’s important to consider the following:
- Respondents’ level of interest in a given subject may differ
For example, when looking at “Personal Interests”, not all respondents who've claimed to have an interest in “vegan food” are necessarily following a strict vegan diet; some of them may just try to eat vegan a few days a week, or some may just enjoy vegan dishes alongside meat-based ones. The same applies to followers of a given sport or team; not all followers will be avid supporters.
- Respondents may interpret answer options differently
For instance, when asked about types of accounts followed on social media, respondents have the option to select “vloggers”. However, not everyone may agree on who is and isn’t a “vlogger”, especially if they follow individuals who are well known for other reasons. Although we take steps to ensure respondents complete questions accurately and consistently (such as by providing examples when relevant), there will be occasions where the lines between one category and another are blurred, just as they are in the real world.
- Response styles differ between countries
Country-level cultural factors often play a role in the way people understand and answer survey questions, particularly those on attitudes and interests. For example, we see that respondents in Japan favor a conservative response pattern. In other words, they tend to be less comfortable disclosing information about their own attitudes and opinions. This aligns with cultural norms in the country, which value a reserved and indirect style of communication. It’s also reflected in other areas such as social media, which tends to be less popular in Japan than in other markets. This, along with with the country’s relatively old age structure, means we tend to see lower percentages for attitudes and some behaviors in Japan than we do in other markets. To a lesser extent, we see this type of conservatism in South Korea too, as well as in some Western/Northern European markets such as Germany, the Netherlands and Belgium. Meanwhile, we see a more positive and enthusiastic response style in several Latin American markets and parts of Asia Pacific, particularly Brazil and India. In these markets, respondents are more likely to place themselves on the agreement end of agree-disagree scales, select affirmative answers and pick multiple options they identify with from lists.
Our respondents are only human - like the rest of us, their memories aren't perfect. To help them answer accurately, we prefer asking about shorter timescales rather than longer ones. However, regardless of the timescale used, most respondents will only select those options they've meaningfully engaged with over the specified period; they won't, for instance, select websites they've accidentally clicked on, or only visited very briefly. In this respect, survey data naturally filters out some of the noise that can be associated with other approaches. For more detail on comparing our survey data with other sources, see here.