Thursday, September 20, 2018

The Effect of Panel Bias on Website Audience Measurement

by Kabir (writer), , March 17, 2015

Read the following guide to understand what panel bias is and how it affects large websites as well as international traffic.

Audience measurement often demonstrates drastically different findings than other Internet measurement tools such as web analytics. Factors such as weighting – which entails giving some consumer demographics more importance than others based on the difficulty of reaching them for market research insight – gives rise to panel bias which impacts on figures.

For example, web analytics could tell you that there are 100% unique cookies on browsers per individual PC, while your audience measurement service says there are just 100 people at-work.

This guide by respected UK audience measurement specialist Kantar Media will explain the impact that panel bias has on large websites as well as international traffic.

Risk of Panel Bias for Large Websites

While you might think that large websites will not be affected by panel bias because they tend to attract a broad demographic of consumers that is representative of the complete market they cater to, this is not necessarily the case.

For example, 30% to 35% of the Internet population accesses goes online at work, and tend to spend more than two times longer online than when they are at home. Additionally, the at-work audience has grown significantly over recent years and they now represent a significant portion of online commerce spend i.e. they shop a lot. A Comscore figure demonstrates that the at-work audience in the US contributes as much as 60% of total online spend, so they are a coveted demographic to tap into for insight into what drives they buying decisions. That said, this demographic often represents only 10% or less of any audience measurement panel because they are notoriously hard to reach (companies tend to frown on employees taking part in market research since this represents a loss of productivity). This 10% is also more representative of small businesses rather than large corporates.

What all of this boils down to is that when measuring unique visitors to their website, most companies will be affected by the inherent bias of audience measurement panels, regardless of their size, because certain demographics will always be harder to tap into than others and their opinions will therefore count for more than the demographics that can easily be reached.

Risk of Panel Bias for International Traffic on Websites

International panels (consumer demographics or segments) tend to be much smaller and more fragmented, which affects audience measurement figures. Some audience measurement firms even ignore at-work international visitors to websites because of this, but the result is that websites looking to measure the true figure of unique visitor traffic are missing a valuable portion of their audience, despite the fact that international consumers make up only a small portion of traffic in most cases. This blind-sight can become a glaring problem as a website grows, especially if it stocks products that have the potential to appeal to the international market.

As an example, the hard to reach at-work demographic in Japan represent at least 50% of their population, so they would be a valuable international audience to cater to in order to gain their valuable feedback.


While the natural audience measurement bias that occurs in this field can be frustrating since it delivers very different figures to tools such as web analytics, audience measurement services are none-the-less highly effective at providing the right steer for key strategic business decisions. Rather than getting frustrated and bogged down by differing statistics, Kantar Media advises companies that every tool has its use. Correctly applied, the audience measurement bias can be leveraged rather than viewed as an obstruction.

About the Writer

Kabir is a writer for BrooWaha. For more information, visit the writer's website.
Want to write articles too? Sign up & become a writer!

0 comments on The Effect of Panel Bias on Website Audience Measurement

Add A Comment!

Click here to signup or login.

Rate This Article

Your vote matters to us