http://www.thomascook.com/blog/ has a few broken links within the menu structure. Such as:
City Escapes -> Eurostar.
Deals -> Car Hire
From a website speed perspective:
As well as sorting all examples of the above, I would work towards the goal of every webpage loading within 2 seconds on most devices. Currently the website seems to load (from a test server in London) within ~4 seconds.
I would achieve this by doing the following:
Removing redirects from code that’s potentially outdated
Removing legacy files from webpages
Gzip (possibly not applicable)
Using an example page:
Would save 35kb by losslessly compressing the images.
Minifying would save: css (22kb) js (23kb) html (3kb)
There are 13 redirects on this page (mostly external sources) which if applicable could be updated. In total that’s about 100ms of extra page load time just in redirects.
Gzip (if possible) could reduce transfer size by 285kb
Pages such as: https://www.thomascook.com/cruise/ which are currently https however are loading content from an insecure source would be a priority to sort. Due to possible effects on conversion rate (yellow padlock).
There is a large amount of white space below the email signup, this would be removed as to potential cases of users thinking the page has ended, whereas there may be something that could push them to convert.
With a better page load time, an increase in conversion should follow. Until knowing the most visited pages on the website it would be difficult to say what could be done to improve conversion.
I would focus on updating all of the current 301/302 redirects to their actual destination. Due to the fact that this not only helps SEO and provides a better user experience (14ms saved by changing the link).
Would use software to run daily crawls throughout the whole site to find any broken files/links/images. This would sort testing from a “technical” aspect. Would then create a list of different types of pages, running throughout the list to see if there were any broken features in various browsers. (Possibility to automate through screenshot software if there is a budget for it?)
Final test would be to study the analytics and see if there were pages that were converting poorly/lots of users abandoning to see if there were any issues with this page.
Some overall information regarding the website:
370 4xx errors (404/403) (links to these pages only count once, so a broken link could be on every page and only still count once)