It always attracts a lot of attention when Googlers are up on stage and open for questioning. The Ask Me Anything (AMA) with Google Search at the recent SMX West conference was no exception. This panel, moderated by Danny Sullivan, featured two prominent Googlers: Webmaster Trends Analyst Gary Illyes and Webmaster Outreach Specialist Mariya Moeva.
In this post, I’m going to recount the bulk of the questions asked, and the answers from Illyes and/or Moeva. Please note that responses have been summarized rather than directly quoted. Let’s get started!
1. Does Google maintain any metric along the lines of a “Domain Authority” concept?
Gary Illyes: This is something that Google feels does not really work. For example, on blogspot.com, the user blogs created there really shouldn’t inherit the authority of the main domain. All ranking is page-specific.
2. What about subdomains?
Illyes: No such thing as subdomain authority.
Mariya Moeva: Before we get asked, I’ll add that there is no similar signal for folders, either.
3. Can you talk about the update called Fred?
Illyes: We make tons of updates all the time, and this is not something we would have thought to name or announce. It only got a name because of an interaction on Twitter with Barry Schwartz.
Google won’t communicate about what this particular update did, but you can find the things that it targeted if you carefully read all of the Webmaster guidelines. Unfortunately, we can’t say which particular ones.
Moeva: One of the issues with naming updates is that people start attributing all sorts of things to those updates, even though they aren’t actually related.
4. When will Google stop ranking internal search pages? Why does this still work?
Illyes: We do frown on these pages getting indexed, as they are not that useful for users. We do have algos that try to get rid of them, and for that matter not even crawl them, but sometimes these don’t catch everything, and we may have to manually intervene.
Moeva: If you’re trying to get these pages out of the index for your site, you can use the parameter handling feature in Search Console to say, “Don’t look at these pages.”
5. What data does Google use to see the “also try” info in a Knowledge Panel?
Illyes: Maybe we leverage the Knowledge Graph categories to find similar pages, but that’s just a guess.
6. What if your site does not get a Knowledge Panel? How do you get one?
Illyes: There is no way to do that really. Google does consume Wikidata, so that could get you in. Google also uses the CIA World Factbook. Anything that gets in requires multiple sources of data to support the need for it to be in there.
7. How do I control the picture that shows up for my business in the Knowledge Graph?
Illyes: If an incorrect image comes up, report it. Note, though, you need many people to report it for it to get looked at, and they have to come from accounts that have not previously abused the system.
(Illyes then mentioned the example of Stone Temple Consulting, which shows an incorrect image of a Chinese temple, and explained that sometimes the problems are hard to fix, but they are working on it.)
8. Can you bring back the link operator?
9. We heard you say that Baidu uses AMP, but we thought they had their own thing.
Moeva: Baidu started with mobile instant pages (MIP), which they still support, but more recently, they have expanded to support AMP pages as well.
10. Do you plan to expand the Google Posts feature? (Note: This is a feature where people can post information directly into the search results, but it’s used in a very limited way today. You can read more about it here).
Moeva: This feature was used during the US elections, and we are looking at future potential uses. Of course, for this to work it has to be high-quality info. The target groups are usually not really savvy on the SEO side, so the feature is good for the website-challenged. So we don’t know how we’ll expand it, but we’re trying to see what kinds of groups it makes sense for.
11. In Search Console, when we see links from hacked sites, should we disavow them?
Illyes: Do you trust the links? If you don’t, then disavowing them makes sense. Our algorithms do their best to auto-discount these, but these algorithms are written by humans and sometimes subject to error.
12. Should I have my exact keywords in the domain name?
Illyes: Exact match domains (EMDs) are not inherently bad, but some people do things with them that makes them bad. If you see someone else that’s ranking and you think it’s because they have an EMD, our advice is to not worry about it and try to figure out the other reasons it ranks.
13. If we have an EMD, but it is really the company name, do we need to worry about EMD algos?
Moeva: Unless you have really poor content on 50 EMDs simultaneously, don’t worry about it.
Danny Sullivan: If you have an EMD that matches the really good solid content you are already doing, you’re on the right side of history, and you’ll be fine.
14. Is a subdomain or subdirectory better for targeting different countries?
Moeva: Our site has a good post on this. It shows a table for when/how to pick which variant.
15. How does Google deal with the new domain names (such as .SHOP).
Illyes: These are treated the same as any other TLD.
16. Is it true that Google won’t crawl a URL that has more than two folders?
Moeva: There is a limit to the characters in a URL, though.
17. Does folder depth make a difference ranking wise?
18. If we launch a subdomain, and it gets penalized, will the domain be impacted?
Moeva: It depends on the reason why. If a blog got hacked, then that impacts only the subdomain. Generally, we try to be as surgically granular as possible.
19. Does Google access info from Gmail and other sources to personalize results?
Illyes: Yes, if you’re logged in.
Sullivan: This happens only within Gmail (I think).
20. Why did you decide to do a video on how to hire an SEO?
Moeva: Our target was non-savvy people, to help them understand how to approach it.
21. How many sites do you manually review in a week?
Moeva: Every Googler has to do 20 sites before breakfast (joke). Seriously, scale is the main objective. Google is trying to find patterns, more than sites. Ideally, we do this in a way that you find a pattern that scales across many sites.
22. How does it affect our ranking if our responsive mobile site has the same content as the desktop?
Illyes: This is the desired behavior, and there is no duplicate content issue since they are on the same URL. Also, there is no such thing as a duplicate content penalty.
(Author’s note: If you have duplicate content because you are scraping other people’s sites, this can result in your site being hit by Panda, but that is not what Illyes is addressing in his response.)
Having a responsive site will help you, especially in the future. With mobile-first, when it rolls out, responsive sites won’t have to change anything. Other sites may have to do something more to not suffer when that transition happens.
23. What would be harmed by Google being more open and transparent?
Sullivan: To be fair, they are very open on many things.
Moeva: The issue is that we’re trying to get people to focus on the right things. If we start dissecting (and sharing) what we do every day, this would prevent people from focusing on their sites. You should try to focus on the good stuff. Worrying about what Google change happened, when the next one will be, etc., diverts your focus from where it belongs.
Illyes: Both of us teach a class to Nooglers (new Googlers) called the “life of a query.” We know a lot about how it works, and this class is very specific. But, we have to decide if revealing a piece of information will hurt us in the long term. If we share info that causes people to do the wrong things, this can hurt the search results, and we can’t allow that. But we try to be as transparent as possible.
24. Can Google see a bounce on your site, even if not from a search result, and is it a negative result when that happens?
Sullivan: In other words, do you use click-through rate (CTR) as ranking factor?
Illyes: We only use CTR for QA purposes, not direct ranking purposes.
Moeva: We also don’t use things like Chrome to capture data like that.
25. If CTR is not a ranking factor, then why do tests sometimes show that it does influence rankings?
Sullivan: What if you believe that Google is lying about this — that you believe they’re using CTR, even though they say they aren’t? What would you do differently with your site? In principle, nothing. You should be trying to optimize to improve CTR and retain users anyway!
26. What about RankBrain and machine learning? What’s up with that?
Illyes: Not really anything new happening. The team is focused on some other machine learning ideas, and I can’t say what they are or whether they will be applied to search. But we’re always throwing out ideas on how to better understand pages or queries and working to make search better. Machine learning is just a tool that you can apply to different things.
Moeva: A great example is what video to show you next on YouTube; that’s an interesting challenge that is a perfect application for machine learning. We can leverage what people have watched and match them up with choices made by others with similar affinities.
Sullivan: Could you have a machine learning search algorithm system now?
Illyes: We probably wouldn’t want that, as we still look at search pages manually, and machine learning algorithms are freakishly hard to debug. For example, if we see some problem manually, it would be impossible to figure out why the decision was made. If we replaced the traditional algorithms with machine learning, we would have a really hard time improving results, because we would not able to identify where the failures are or what caused them.
Moeva: The algorithms are only as good as the training data that we give them, and that’s a big issue in coming up with good ones.
27. How do you measure voice searches? Can the rest of us get to see that?
Moeva: Voice queries are often super long-tail because of their use of natural language, so it’s hard to decide what to show. (Moeva then asked people to provide feedback on what they would do with it.)
28. Is it possible to highlight quick answers in Search Console separately?
Illyes: We get this request a lot, and we’ve talked to the highest possible person in the search team to see if they can, but we have not persuaded them yet. We’re still fighting for it, but it’s not a simple problem. Please ping us on Twitter and tell us how you would use that data to help us.
29. How do you get featured snippets?
Moeva: Create content that is relevant to the query and structured well. Structure it as an answer to a question.
30. Is WHOIS info used as part of local ranking?
Illyes: No idea.
31. Does the percent of 404 pages on your website impact your ranking?
32. Should we cross-link between product and category pages?
Illyes: Does it help users?
33. Does a large e-commerce site have a natural advantage over a small site?
Illyes: It doesn’t matter.
34. Will traffic from email and social media help your rankings?
Sullivan: In other words, do you use social signals?
Illyes: We did it in the past for a specific feature, but then the social media site turned off the feature, and it was a really bad thing, and for this reason we’re not willing to be dependent on social signals.
35. Is content length a ranking factor?
Illyes: No. The quality and relevance of the content is the key.
36. Is it bad if we have http and https live on a site at the same time?
Illyes: In general, no — but we have seen situations where sites have some differences between the sites, such as they implement hreflang to the http version of the site, but not the https version. That type of thing could be an issue.
37. Do you look primarily at the source code for a page or the document object model (DOM)?
Moeva: We look at the DOM.
38. Will we see an AMP e-commerce carousel in the near future?
Illyes: We are looking into e-commerce features, and we’re very aware that this functionality is needed. If you want to see what’s planned, check out the AMP roadmap.
39: Is authorship really gone as a ranking factor?
Sullivan: Dead, dead, dead, deader than Google+.
via Search Engine Land http://ift.tt/2ohQQmM