Summary Score

From SiteRay wiki

Revision as of 19:05, 5 March 2009 by Oliveremberton (Talk | contribs)
(diff) ←Older revision | Current revision (diff) | Newer revision→ (diff)
Jump to: navigation, search

Contents

What are these?

Summary Scores appear at the top of every report. They aim to summarise the overall qualities of the website as briefly as possible, providing a quick comparison between websites.

Summary Scores

Currently, there are up to five Summary Scores available:

How Summary Scores are calculated

Each Summary Score is calculated as a weighted average of other test scores. For example, 5% of a summary score might be the Spelling score, 10% the Keywords test and so on (these are not real figures).

The precise weightings for each test vary slightly over time, as we are continually refining them based on the changing nature of the web (see Why does SiteRay keep changing?). You can always view a complete breakdown of how each summary score is calculated by viewing that Summary Score in full.

Penalties

By itself, this approach can tend to give overly favourable scores in some cases. For example, consider a hypothetical website:

  • Website has 1,000 pages
  • It is impossible to spider the website beyond the first page
  • The first page is well written and spelt, no broken links, proper use of headings, titles, readable text etc.

In this case, a purely weighted score might do something like this:

Test Result Weighting
Spiderability 1/10 5%
Spelling 10/10 5%
Alternative text 9/10 5%
Broken links 10/10 5%
... etc ...

The resulting score (in this example) could be around 9.5/10, which is clearly unfair for a website with such a fundamental problem. Simply increasing the weight of the Spiderability test is not enough, as sites which are spiderable (the majority) would all score much higher for something that should be taken for granted.

Instead, we use a system of penalties. Where a particular score is below a certain threshold, SiteRay applies a penalty to the overall score, proportional to how significant that problem is. Our earlier example now looks more like this:

Test Result Weighting Penalty
Spiderability 1/10 5% -3
Spelling 10/10 5%
Alternative text 9/10 5%
Broken links 10/10 5%
... etc ...

In this example, the new penalty of -3 would reduce the score to around 6.5/10, a much fairer score. Note that all of these figures are example only – the real weights are much more varied and complex.

Each time a penalty is applied, an explanation is listed next to it on the appropriate Summary Score table.