The purpose of this guide is to look at why there are sometimes discrepancies between the figures that SimilarWeb shows for a given URL and the figures that the owners of that URL see from direct measurement data tools such as Google Analytics.
Direct Measurement vs SimilarWeb Data
Why Does My Direct Measurement Data Show Something Different to SimilarWeb’s Data?
There are three main reasons why the data you get from direct measurement tools such as Google Analytics may differ from the data shown in the SimilarWeb Platform.
- Different Methodologies
- The Human Touch
Most businesses use direct measurement tools, such as Google Analytics, to analyze traffic to their own domains. While the technology is usually similar from one tool to the next, the data often varies. This is because of different methodologies used to calculate sessions, session times, and other standard metrics. For example, some methodologies deduplicate visitors which is a method used by SimilarWeb.
The Human Touch
Trackers and analytics codes can be implemented on websites in different ways. Many tools rely on an imperfect human element. This can result in under-reporting pages where a code has not been properly installed or forgotten all together on newly published pages. It can also lead to over-reporting when brands add trackers to other widgets, apps, plugins and email clients, incorporating off-site results along with other figures.
When comparing numbers shown by SimilarWeb to numbers from a direct measurement tool, you need to make sure that your parameters are consistent.
- Are you comparing the same time frame?
- Are you comparing the same devices? E.g., desktop, mobile or both
- Are you comparing the same metrics? E.g., unique visitors vs visits
SimilarWeb uses a consistent methodology across all of our estimations, measuring all sites in a consistent unified manner.
Using SimilarWeb and Direct Measurement Data Together
Rather than competing against one another, direct measurement data complements SimilarWeb helping to calibrate for bias and move from sample to estimation by transforming data sources into intelligent estimations across all sites.