Welcome to the Q&A Forum

In my opinion, the 'normal' optimised images would be better because this would allow you to insert keywords in both the image name & alt tag. In both instances, I always recommend that you name the image file name using a keyword. In all instances make the images as small as possible WRT file size, that faster your site loads; the more Google will 'reward' you.

Interesting development which may be of interest to you Ernst: Google admitted just the other day that they "haven't supported rel=next/prev for years." https://searchengineland.com/google-apologizes-for-relnext-prev-mixup-314494 "Should you remove the markup? Probably not. Google has communicated this morning in a video hangout that while it may not use rel=next/prev for search, it can still be used by other search engines and by browsers, among other reasons. So while Google may not use it for search indexing, rel=prev/next can still be useful for users. Specifically some browsers might use those annotations for things like prefetching and accessibility purposes."

It may potentially affect the rankings on: pages without SSL pages linking to pages without SSL At first, not drastically - but you'll find that you'll get more and more behind until you had wished you just embraced HTTPS. The exception to this of course, is if no one who is competing over the same keywords, is fully embracing SSL. If the majority of the query-space's ranking sites are insecure, even though Google frowns upon that - there's not much they can do (they can't just rank no one!) So you need to do some legwork. See if your competitors suffer from the same issue. If they all do, maybe don't be so concerned at this point. If they're all showing signs of fully moving over to HTTPS, be more worried

Thanks Tim for your answer. Totally agree with your opinion from customer experience perspective. But most of all I was wondering how Google sees it. Is it bad if a webshop is closed for one day. Does it affect the ranking in search results for the rest of the days?

As long as all of the URLs stay the same and you aren't building a blog network (a bunch of sites that interlink), I don't think you'd see much of an SEO difference. I haven't seen anyone complain about that particular issue which is a good sign. If/when you follow through, please comment back and let everyone know your experience!

Hi Tom Agree with Martijn that it depends for example, the robots.txt is generally the first port of call for bots as it allows them to understand where you want them to spend their finite time crawling your site. You can aslo give direction to all bots at once or specify a subset. It is generally the best option for blocking pages such as you /cart/ etc were they don't need crawling. The problem with robots.txt is that it doesn't always keep pages from being indexed especially if there are other external sources linking to the pages in question. The meta tag noindex on the other hand can be applied to individual pages and you are actually commanding the robots to NOT Index the relevant page in serps, use this option if you have pages you don't want appearing in Google (or other search engines) but the page may still be relevant for authority or able to acquire links (make sure to use Noindex follow) as you still want the robots to crawl the page. Otherwise use Noindex Nofollow hope that this helps.

WOW, this is an interesting thread. In theory, rel next prev is what Google wants you to go with. In practice, however, I haven't not seen this work as advertised by Google, and end up going to rel canonical in most cases. Here's one way to think about it: Allow non-filtered pagination for top-level categories to be followed, but NOT indexed. Give them their own rel canonicals (self referencing) and ensure the intro content (or any other static content on the page) only shows on the first page (which should rel canonical to / instead of /?page=1). This will ensure your product pages all have a path going to them. Use rel next prev here, which "may" help the main/first page rank better by consolidating ranking signals from the paginated set. For sub-categories and/or filters/facets use rel canonical pointing to the canonical version of that category page. None of this, however, takes care of the crawl budget issue on enterprise eCommerce sites with crawlable filtered URLs. Therefore, I also use the robot.txt file, or nofollow attributes in links to handle this. I don't often use the nofollow robots meta tag for a variety of reasons. Again, in the real world rel next/prev doesn't seem to be working very well. I haven't tried it in awhile, but it just doesn't seem to do what Google says it's supposed to do. This is why you see a lot of sites using rel canonical instead. I think we should think about this in terms of where you are in the architecture instead of trying to fit all pages into a single scheme. For example, some sites may even want to allow the first facet/filter to be indexable - such as "Filter by Brand" - because people are searching for it. If that's the case, I'll often advise they turn that into a real category instead of a filter. Another example, if you sell kitchenware and Pots & Pans is a category but Material is a filter you're missing out on a lot of searches for things like "Copper Pots & Pans" or "Stainless Steel Pans". It all depends on the situation, and has to be evaluated on a case-by-case basis.