The State of SEO in Mid 2017

This is a review of the key factors for SEO, and organic search, as we reach the mid-point of 2017. It is a study of the meta factors driving change in the industry which will affect where Google and others push the market over the next 12-18 months. It is not a look at whether or not H1 heading tags, and the like, have increased / decreased in relative ranking power.

It covers the Mobile First Index; AI & Machine Learning; Content Quality; Technical SEO; and Local SEO.

Enjoy & if you’d like to discuss further, please get in touch.

#SEO #MobileSearch #LocalSEO #TechnicalSEO #ContentQuality #ArtificialIntelligence

Summary:

Google is deep into planning a large-scale revision of its core approach to indexing and ranking websites – namely a Mobile First approach. Note that this is going to be significantly different from a mobile-friendly website. Google is currently developing this new index and it appears to be taking a while longer than expected.

It had been expected mid-2017, but now appears to be due for release in late 2017 / early 2018. Google has stated it wants the new index, and resulting rankings to be quality neutral. Take care when interpreting “quality neutral”, it does not mean that rankings will not change, but rather that the quality of the pages ranked under this index will be broadly similar. New sites and pages could surface, and other drop. There will be a longer post on the Mobile First Indexed to be published shortly.

Key Factors to Consider:

• Content Visibility.
  • It will be critical that content is visible to the end user when they load a mobile page. This means avoid content hidden behind tabs, or read more. Making users click to get information they want is not considered good UX.
• Cascading Surfacing.
  • This means that Google has been pretty specific about its approach to surfacing answers to users’ queries. Namely that, if it can, it will display a direct answer in the search results, either by a direct answer (such as a simple calculation. The second class it wants to display is a Featured Snippet, pulled from a website. The third class of content it wants to display is Rich Snippets or Cards, such as for recipes, movies etc. Fourthly, it wants to direct users to the relevant app rather than website, before finally displaying one of the old-school 10 Blue Links, or normal organic links. Intermingled with these will be AMP results, which don’t currently get preference over standard web results, but it can’t be far away.
  • Make no mistake, Google does not want anyone to click away from its search results pages. It wants to provide the answers and not send traffic. This is bad for site owners.
• Ecosystem Battles.
  • A large part of this is driven by ecosystem battles. Google’s Android now runs the majority of smartphones. It really has no position in desktop. When you own the Operating System, it is natural to drive more people to use your OS and tie you into the ecosystem. That way your data is used for usage information and targeted advertising. Apple has a solid ecosystem, but is let down by its search functionality, which requires third party input. Microsoft appears to be abandoning non-business users as part of its ecosystem and has definitely abandoned mobile.
• But…overall visits are still much higher on desktop.
  • Although Google has been trumpeting the fact that more than 50% of its traffic comes from mobile, for most websites this is still not the case. There are huge swathes of the internet where desktop still out performs mobile by a 65-35 split. Of course, having a mobile-first index will change this, forcing people to adopt mobile as the default way to search, and driving people into the ecosystem.
  • It should be noted that most desktop vs mobile visit stats do not include App Traffic, and as discussed above, Google is surfacing App results before website results, so it is likely that some mobile traffic is disappearing down an App hole, skewing the split somewhat.
• But… overall conversions are still much higher on desktop
  • It really is unfathomable that 5 years into the mobile transformation, conversions are still so poor on mobile. The reasons for this are many fold, but ultimately it comes down to poor design, poor analysis and poor thinking on the part of website owners. They still do not know how / where to position a call-to-action on mobile, nor do they care that a user is going to be put off by a 10-field form to purchase. Both Apple and Google are making strides towards overcoming this with auto form-filling, including credit card details, but that is a band-aid solution which fails to deal with the actual issue.
  • When you take into account poor design and other factors, low mobile conversions become a QED proposition.
  • Again, be aware that mobile vs desktop conversion data may not take into account conversions which take place through apps.

Hoe to Build a Mobile-First Website:

1. If you have a mobile-friendly website, convert it to a Mobile First website. As this is likely to take some time for large-scale websites, you should have started this last year.
2. If you have a non-mobile website, go straight to a mobile first website, passing mobile friendly entirely. Again this is something that ideally you would have started a while ago, which means you need to start it now, if you haven’t already.
3. Work really hard on mobile conversions. Use A/B or multivariate testing to test design and response from mobile users, including form type and form-length – capture the right data at the right point in the process.
4. Design your forms so that the field inputs align to the name used by Apple and Google for their auto form-filling software. It’s not hard, but you’d be surprised how many big websites miss this simple trick.
5. Pay attention to Google’s priority in surfacing information. If you are going to rely on being one of the 10 Blue Links, be aware that you will be pushed further down the Organic SERPs, possibly to the point of traffic dropping substantially. Focus on being the authority for an answer, using structured data and rich snippets, creating an app (and getting it indexed), before AMPing your standard HTML pages.
6. Get in touch if you’d like to discuss further.

More info:

• Google Webmaster Blog – Mobile First Indexing
• Google Developers Mobile Friendly Guide

Return to Top

AI & Machine Learning Summary:

Everything for the last couple of years has been all about AI (artificial intelligence) and what is known as Machine Learning. In a number of instances AI should actually be Very Clever Algorithm, or even worse Just Lots of Data. There has been lots of chatter about Google’s algorithms, such as RankBrain using Machine Learning or AI to self-improve. This has been true to a certain extent, but it appears to be clear from Google that firstly “AI will never write the algorithm”, and that people at Google “understand the algorithm”. Indeed, if you look at Google’s services, AI & Machine Learning are used for very specific services, not the wider algorithm.

• AI & Machine Learning: Buzzword Bingo
  • AI and Machine Learning are big buzzwords currently, in the same way that Big Data was a few years ago, and they are equally misused. Essentially, a lot of organisations which use data to run their businesses, or to sell services, are saying they use “AI” to derive results. In fact, what they tend to be using are more sophisticated algorithms designed to compute more factors because of the scalable nature of cloud computing. Very rarely are they actually using AI, or even machine learning. Some are creating great things with Machine Learning, although I suspect some of it it really just more data to throw at middle managers.
• AI, Humans and Google
  • Google’s comments about AI and humans are interesting – this is a big company with lots of interest in AI. Indeed, at the last Google IO they said they would become an “AI-first” company. In the front-facing search engine we’ve seen big issues with fake news, and ads being placed next to offensive content which indicate their usage of AI is not as prevalent as imagined, or it’s not very good.
  • One issue I would note about Google understanding the algorithm is that this is true to a certain extent, but it is not wholly true.  There have been numerous examples of unintended consequences of algorithm updates, where big sites have been demoted, only to re-appear a few days later – this, and the flux seen in SERPs in the immediate aftermath of a Google update, are indicative of a large-scale algorithm not quite working the way it was thought it would.
  • Based on the evidence today, the major tech pkayers have a way to go before we actually have AI. This won’t stop a lot of vendors from claiming that everything is AI.

AI Actions to Take:

1. Monitor the situation regarding AI and Machine Learning. It is coming, but it is likely still a quantum leap away or two.
2. If vendors claim to use AI to produce data, query it – don’t believe the hype.
3. Google for the minute is still a human algorithm. This is unlikely to change in the near future.
4. Get in touch if you’d like to discuss further.

More info:

• Google Cloud Machine Learning
• Google’s AI Experiments

Return to Top

Summary:

Content is becoming ever more important to being discovered on the web. However, it is challenged by poor implementation and vast swathes of poor quality content.

• Content Floods
  • If you open almost any feed, or Twitter stream these days, you are almost certainly to be assaulted by an inundation of content. Some will be great (like mine), some will be mediocre, others will be poor. This is a race to the bottom as people think the quantity of content can outweigh the quality. This is wrong. Concentrate on producing good quality content at the right cadence and frequency.
• Content Length
  • The number of words per article seems to be drifting inexorably upwards. I have seen graphs claiming that all content should be 2000+ words, which is just silly, and more reasoned approaches, which indicate that content which ranks for significant numbers of keywords tends to have a sweet spot between 500 and 2000 words.
  • Remember that 2000 words of piffle is still 2000 words of piffle.
• Quality
  • With the rise in content length, we’ve also seen an inverse drop in content quality. Writers tend to wedge snippets of vaguely related information into their articles. This is poor.
  • To be seen as an expert, write deeply about a topic. Don’t cover the surface of 10 different things inadequately.
• User Journey
  • Since RankBrain, content which matches the user’s place in their journey has been promoted above more semantically correct articles. This is a good thing generally.
  • Always understand your prospective user journey and where they are on it when they query Google. Are they ready to buy, or just looking? Are they competitive, or do they want to read at length. Content should always match the user journey.
• User Experience
  • “Nice looking”, or swishly “designed” is often mistaken for good user experience. It’s not. Google knows that users want answers to their queries, and they want them fast – it’s their entire search ethos in recent years. If your content can give them that answer fast and in a clear, unambiguous manner, then you have given the user a great experience, increasing the likelihood of a bookmark, onward click, return visit, or conversion. To see this in action, you only have to look at Google’s implementation of Material Design and their Help Pages.
  • Focus user experience on answering the query as quickly as possible.

How to Create Quality Content:

1. Make sure you pace your content creation. Do enough to get noticed, but not so much that you sacrifice quality.
2. Keep content length reasonable for the subject. Detailed questions need detailed answers. Brief questions need brief answers.
3. Focus on the user journey. Understand the user, where they have been and where they want to go. Doing this will help users to find answers and move forward to the next phase.
4. Ignore designers who go on about user experience. It’s a myth that people can’t cope with reading content on mobile. Most mobile issues are IA and navigational related. Focus on given the user an answer as swiftly as possible.
5. Get in touch if you’d like to discuss further.

More info:

• Smart Insights Content Marketing Stats
• HubSpot Marketing Stats
• Basic SEO Content Planning
• Basic SEO Body Content

Return to Top

Technical SEO Summary:

Despite the content marketing shills believing that reams of content are the gateway to success, it’s becoming clearer that having a great Technical SEO strategy is regaining its previous importance. While Google makes efforts to overcome technical deficiencies on the part of webmasters, you need to be technically adept to deliver consistent crawling, indexing and ranking of your content.

• Google Search Console
  • This is becoming more important in the webmaster’s armoury of useful tools, especially if you want to test elements of technical implementation. Google has made great strides in making it a useful place to go, rather than just bare data.
• HTTPS and Security
  • It is time to implement HTTPS. It’s not difficult, well, it can be if your infrastructure is complex and third party vendors are struggling to implement HTTPS themselves.
• CMSs catching up, but still lacking
  • Many CMSs are catching up with the needs for Technical SEO, however they tend to be lacking in certain areas, namely:
    • Flexibility – it can be difficult to change default settings, or there are no settings to change.
    • Correct implementation – they are getting better at this, but they still seem to miss certain elements at times. This is frustrating.
• JavaScript Rendering
  • Ahhh, JavaScript, the saviour of many a server’s CPU. Google’s rendering of JS has improved enormously, however there are still concerns about crawling speed – allegedly a 10x efficiency impact, as well as whether or not Google can actually crawl and index content contained within some JavaScript frameworks.
  • Google is likely to catch up, but it would be wise to ensure that JavaScript rendering is only done when you absolutely have to. If there is no need for dynamic content, just use flat HTML, or use a CMS which delivers flat / cached HTML for non-dynamic content.
  • Even worse are the current fashons for SPAs (single-page-apps) and PWAs (progressive-web-apps). These are both in place because coders cannot think their way out of the corner they have put themselves in. Loading content behind the scenes to give an impression of speed is a broken workaround, as you rely on using too many user resources.
• Browser Wars
  • I had thought that the browser wars were done and dusted. But it seems that Apple’s Safari is no longer the golden child of web devs, whereas Chrome is, and Firefox is a fuddy-duddy old Grandpa sitting in the corner (luckily IE is dead. I’m not sure enough people use Edge to really form an opinion on it).
  • Clearly it makes sense in Google terms to develop for Chrome, but also remember that Apple has a large installed user base.
• Cascading SERP
  • As discussed above, Google has an order of preference of the content it displays in response to search queries. Technical SEO is particularly needed to deliver structured data, app indexing and AMP.
• Page Speed / Load Times
  • One of the big ticket items with the Mobile First Index, and using the web in general, is Page Speed and Loading Times. There are all sorts of stats on user abandonment if Load Times exceed x seconds. Load times are being bloated by overloaded web pages with too much JavaScript (or JavaScript which requires processing power to run), as well as static images not cached or CDN’d properly. This is why AMP with its stripped down code and JS is being promoted. AMP is not likely a long term solution, but it will remain until the web catches up with current users “need for speed”.
  • Where possible, heavy duty work should be undertaken to speed up the web experience – remember that AMP is a stopgap.
  • There are numerous studies showing the there is an exponential relationship between speed increases, visits and conversions.

Key Actions to Take:

1. Become intimate with all the information and tools that Google Search Console and related testing tools can provide. It will be worth it.
2. Push CMS vendors to provide better implementation of technical SEO requirements. If you don’t ask, you won’t get.
3. Stop using JavaScript when you don’t have to. Lightening a server CPU only places the load on the user’s CPU, causing delay for them.
4. Pay attention to browser issues. Make sure your content is usable and readable by the overwhelming majority of users.
5. Pay attention to the cascading order of SERPs and create technical solutions to appear higher up the chain.
6. Work to speed up your pages and and load times, from initial load to completed load in a  user’s browser. They won’t wait and it will increase visits and conversions.
7. Implement HTTPS.
8. Get in touch if you’d like to discuss further.

More info:

• Google Search Guides
• Google Developers – Web

Return to Top

Summary:

Local search is an interesting thing at the moment. It’s one of the last hold-outs of low-grade SEO, where keyword stuffing still can work, as can link manipulation, and often just outright spam. It’s like a directory strategy from the mid-2000s currently. Local accuracy and social are going to be very big for performance in the next 12-18 months.

• Proximity
  • Google is trying its very best to be right in your pocket when you pull out the mobile to search. Now if you ask for a bar, Google will use your mobile location to pull in the nearest results. This is a touch creepy.  One of the big problems is that Google does not necessarily present those close-by results on the basis of quality, an although Google occasionally filters by reviews, proximity still outweighs it.
  • There’s not a lot you can do to fix a proximity problem, beyond ensuring that when quality filter are applied, your quality is high.
• Reviews
  • Google’s big issues with reviews is that they are still eminently game-able, and it does not have the social graph to be able to rank based on reviews from your friends / connections, yet.
  • When a user thinks of a review site, they don’t really think of Google – and that in an inter-connected, personally driven world is a problem.
  • With the shift towards proximity, it is not likely to be long before Google uses information from your circle of friends to feed a quality filter for local results.
• Competing Ecosystems
  • One of Google’s biggest issues is that it has major competitors in the local search / reviews / information field, from Yelp to TripAdvisor, these apps provide very healthy competition.  What they lack in structured, or prosaic data about a business, they make up for in the depth of contextual quality from their reviews (even if there are instances of fake reviews / astroturfing on these sites).
  • Of course, there are also a number of social platforms making a half-hearted turn into local search as well. (Facebook, I’m looking at you).
• SoLoMo
  • I think I talked about SoLoMo first in 2010 / 2011. Everybody stared at me blankly then, but it is coming to pass, especially as we get more used to being tracked locationally by our mobile devices. I already have at least one app that pops up offers based on my location & SnapChat is currently offering promoted filters which are geo-fenced. Get used to this. SnapChat, and others which have a high degree of impulse and location combined will be prime candidates for local search and running local offers.

Key Actions to take:

1. The standard Local Search reccomendations still apply.
2. Make sure your NAP is correct and consistent.
3. Make sure you have citations, especially local ones.
4. Make sure you are engaging with all social networks and their local offerings.
5. Make sure you are engaging with all the review sites & apps, which would like to be social networks, but aren’t.
6. Don’t spam – even if Kevin The Home Renovations, Home Decorator, Home Builder does.
7. Work to get social reviews, not just drive-by reviews. This will help your local presence as the social gets applied to the geo-filters.
8. Get in touch if you’d like to discuss further.

Return to Top