Carbon DMP | Scoring Intent data to find Christmas shoppers

Black Friday & going into the holiday season is a key time for consumers showing intent-to-buy but advertisers and publishers need to pick up on these signals throughout the year.

From reading articles of interest online, doing product research, reading reviews, right through to adding to a basket – consumers are always signalling intent. Setting this within the context of multiple devices and channels against a backdrop of regulatory and commercial pressures makes for a big challenge – but it needn’t be. Tapping into scored intent data in a compliant way can uncover key trends, behaviours and signals that help drive profitable outcomes.

Using Clicksco network data – which includes a vast amount of retail, lifestyle, end to end ecommerce, review and price comparison sites – we were able to leverage and score a range of intent signals using our machine learning algorithms to see trends for how early people start Christmas shopping, and what categories & brands trend in the lead up & during Black Friday.

Black November: High Intent with rising volumes

Data is only categorised as Intent when a consumer is in-market to purchase a specific product or service. The key to unlocking intent – beyond the interest – is context and the nature of the web browsing behaviour e.g. informational site, a shopping comparison site, a review site etc.

With Black Friday often signalling the official beginning of the holiday season, intent signals intensify in the lead up, during, and well into December. Taking a selection of data for our “Black Friday/Cyber Monday” category from mid-October right through to the 2nd December we can see a general uptick. This reflects the nature of shopper activity as we approach November, and even quickening growth as we get past Halloween. For publishing networks such as Clicksco this means being ready to capture those consumers as they begin their shopping journeys – from the research stages when intent may be low, right through to the purchase, with intent rising as they move through that journey.

Meanwhile, the average intent across the period remained fairly constant – well above 0.95 – but grew as we moved towards Black Friday itself (see below).

Why & How we score Intent Data

No two visits to a web page are the same yet they are often categorised in the same way. By scoring data it creates more reliable and accurate signals of intent for better performing marketing. The run up to the holiday season is a prime example of context changes with intent that can impact our scoring, and how marketers should use scored intent to influence their audience building & activation.

Scoring Intent creates high value audiences through the ability to profile and predict shopping behaviour that can then be leveraged within a publishers inventory to maximise their yields. Carbon’s algorithms calculate intent scores based on 4 core statistical features of a user’s site visits to give accurate, reliable and actionable insights.


The number of days that a particular user has been active within an intent category across multiple sites, not just the original site, giving more reliable intent signals.


How recent the activity within an intent category has been measured across multiple sites to ensure the recency of intent rather than just the recency of a visit to a site to enable accuracy of intent.


A measure of the regularity of the user within a given intent category, how often they have the intent again across multiple different sites and instances.

Source & Site type

The type of website/publication they are on while being profiled. For instance, an ecommerce site will have a higher intent score than a blog or general interest site.

Fancy a chat?

Find out how you could use scored intent data to fuel ROI

Intent Scores across categories

Using a combination of cookie counts and average intent scores (ranging from 0-1, where 1 would represent the highest signal of intent) we looked at a few key categories of our rich, multi-layer taxonomy for the festive period.

Given that the Clicksco network has a large amount of ‘Consumer Electronics’ related content, review and comparison sites we decided to take a selection of categories in this space to see how average intent scores behave in the run up & during Black Friday.

We can see from the graph above that whilst there’s a bit of fluctuation in intent scores, the general trend is upwards as we move through November indicating the increasing propensity to buy. On Black Friday itself these categories all peaked for volumes for the whole period whilst also averaging nearly 0.95 for average intent, apart from televisions.

Televisions is an interesting standout for both volumes and average intent. Over the whole of November the Clicksco network saw the volumes of people in-market for TVs fall, but average intent scores continued to grow, peaking on Black Friday as we can see below. It’s important for marketers to be aware of these shifts to influence their campaigns – high volumes with lower intent represents top-of-funnel opportunities, which can still be targeted with awareness campaigns to move them through the funnel. However, those with higher intent are the opportunities to drive revenue.

From late October through November there seems to be some fluctuation in travel volumes too with a slight surge in intent levels towards the end of November, perhaps as people consider a winter getaway, or begin planning holidays for 2020.

Intent on having a profitable 2020

So why is this important? Marketers often set campaigns based on certain KPIs – from brand awareness through to encouraging a purchase. Intent data adds a key layer in knowing when an audience is in market, whilst scoring intent data reveals insights into how close to purchase the audience is. Similarly, we’re able to score brand affinity to find audiences more/less affiliated to specific brands.

Above: Example from the Carbon audience builder UI for building audiences with varying levels of intent and brand affinity.

For publishers that means they can see what their audiences are in-market to purchase to optimise their ads to reflect that intent therefore improving the user experience and yield potential. For advertisers this means targeting the right people, in the right places, with content that meets their level of intent to move them through the funnel, or ultimately encourage a purchase.

Interested in knowing more?

Carbon works with publishers, brands and agencies in order to help them grasp the opportunities of using uniquely scored intent data to improve performance. Arrange to chat with us to find out more.


Leave a Reply