Using Passive Data to Reveal Actual Consumer Interests
Modelling user interests from Gener8’s passive datasets enables a wider and more accurate understanding of people’s behaviours
Brands invest billions annually in market research to gain deeper insights into consumer behavior and to make better-informed business decisions. However, the valuable data generated by this expensive and time-consuming process can be compromised by survey fraud—essentially, when participants provide dishonest or insincere responses.
At Gener8, through the means of leveraging the passive data from our data feeds, we are able to mitigate response quality issues by using recent user activity to understand whether they have an interest in topics like Formula 1.
So how does this work and what are the benefits of this approach?
Understanding user’s interests through passive data modelling
First, let's explore how we are able to achieve this.
At Gener8 we sit on a vast array of data feeds, including:
- Consumer Browsing, which enables us to understand the search terms made and the websites visited by the user
- Purchase Intelligence, which allows us to understand which brands and items the user has purchased for hundreds of merchants
- App Usage, which helps us to see the apps that were used by the user
By analysing each data feed with a pre-built intent trigger model for an interest like Formula 1, we can start to identify the users who have an interest in this topic.
A data snapshot from the Consumer Browsing dataset, focused on F1-related search engine keywords, helps to visualise how this concept works in practice.
By interacting with dashboard filters above, you can better understand the search intent landscape for F1-related sub-topics. For example, by filtering by ‘Drivers’ you can see that Lewis Hamilton drives the largest share of traffic, with sub-intents like his net worth and the upcoming move to Ferrari being the largest drivers of searches. Or by filtering by ‘Legend Drivers’, you can see that Ayrton Senna is driving the most interest, which is most likely pushed by the newly-released Netflix miniseries on him.
But importantly, whilst this can be super interesting for insights purposes, by focusing your filtering attention on the ‘User ID’ column, you can also begin to understand how we can infer whether an individual user is interested in Formula 1.
When developing passive data modelling-based classification for our users we combine the search intent signals with those of web browsing activity, purchases and app usage. What's more we score the user activity for each data source based on behaviour frequency and depth (e.g. action and/ or purchase intent search terms) to extend our confidence.
Following this process enabled us to significantly expand the Gener8 Audience Segmentation Framework, incorporating an ever-growing range of Interest and Purchase Intent columns through which we’re building an exponentially enhanced understanding of our users through which we can deliver greater value to our clients.
What are the benefits of passive data modelling?
At its core, passive data user interest modelling enables us to have a ‘validated activity’ view of user interest and purchase behaviours, which aren’t impacted by the spectre of survey fraud. Each of these provides an additional perspective for understanding the user, enabling cross-analysis across multiple dimensions.
For example, after identifying users who are interested in Formula 1, we can analyse their streaming service subscription preferences from responses to the Gener8 Snapshot survey.
By comparing those interested in Formula 1 against the wider Gener8 Snapshot survey responses, we identified that these fans indexed highest in their subscription rates to Sky and Now. This aligns well, as these are the only two streaming service options for viewers wanting to watch Formula 1 races live in the UK.
We can also compare the Formula 1 interest group with responses to key demographic questions.
For example, by looking at car ownership we observed that those interested in Formula 1 tend to own a car, and index higher than the wider respondent population for ownership (i118). This could be in part due to the fact that 85% of the Formula 1 interest users are male (i141), and men more likely to drive in the UK than women.
Furthermore, by dissecting the Formula 1 user interest group by generation, we identified that it indexes highly amongst Millennials, which is in line with wider industry research which highlighted that the average Formula 1 fan is 32 years old.
By having Psychographic user scoring process across our user base, we are also able to:
- Validate survey respondent answers. Combining passive data scoring alongside quant survey responses, we can identify the delta between what people say they do vs. what they actually do.
- Identify the right respondents to survey, fast. By categorising users by their passive behaviours we are able to find the right user cohorts quickly, even if they haven’t provided a survey answer to a question on that interest/ behaviour in the past.
- Keep up with evolving interests. People’s preferences are always evolving, and a survey ‘moment in time’ snapshot of interests can quickly become out of date. Our passive data modelling rules run on a daily basis, enabling us to find and focus analysis on users with a strong active interest.
How can I access this data?
We utilised Gener8's Psychographics alongside Gener8 Snapshot survey responses and Demographic datasets to uncover insights from those with an interest in Formula 1.
Gener8 Labs’ complete data and insights solution empowers media and marketing businesses to find actionable consumer and market insights, using our unique, consented, first party panel data sets that are all connected around one user ID.
Discover how you can power your decisions and gain a competitive edge from our behavioural truth set by contacting us today!
