Labs

Anti-spyware

We infect a standard PC with a real-world selection of current spyware by exposing it to an unprotected broadband Internet connection for some unsafe surfing. To this we add a selection from our archive of rogue diallers, insidious toolbars and commercial keyloggers. This new testing methodology mirrors the ever-changing landscape that anti-spyware vendors must deal with, but within a tightly controlled environment.

Many other anti-spyware benchmarks base their results on files scanned and threats detected then removed. However, we no longer feel this to be valid because of the spurious claims made by anti-spyware vendors. Instead, we look at what's actually detected from our line-up, what's successfully removed and then blocked when we try to reinstall. The speed of scanning and the impact upon system resources is also recorded. See the results in the feature table opposite.

We test every product in its default configuration, making changes only where prompted to do so. Definition and program updates are made before scanning if available, and threats are deleted or quarantined afterwards depending on what the product recommends.

How we work out the ratings

At the bottom of each review are six ratings: Accuracy, Speed, Ease of use, Features, Value for money and Overall. Accuracy is based on the products' average scores in the tests listed above, while Speed is a reflection of how long it takes to scan the complete system.

Our Ease-of-use rating is determined by a hands-on evaluation of factors such as installation and initial configuration, automatic definition and program updating, scan scheduling, impact of scanning/protection on system resources, information available regarding threats found, plus the quality of help files and support.

We reward features that take the idea of spyware protection further than just the basics. For example, by adding process guards, host file immunisation, browser hijack setting restore, system rollback and click-trail cleaning.

Our Value-for-money rating is based on a combination of all the above scores and the cost of the actual product, while the Overall score is simply an average of all the scores. The Labs Winner award goes to the product with the highest overall score, with our Recommended award going to the runner-up.