We put Team Koozai on the ground live reporting from Search Elite 2018. If you couldn’t make it to this year’s conference on SEO and CRO, we have gathered everything you need to know below:
Did you know that 72% of people who own voice-activate speakers use their device daily including booking appointments with Google Assistant?
Try “Alexa let’s chat” and see what happens. This gives consumers an insight to what we’ll be exposed to in the next few years and how machine learning works. But don’t worry, Alexa isn’t always listening. Devices tend to have set commands that they listen out for in order to activate.
How to prepare content for voice search:
– Prepare long-tail keywords when producing content
– An approach targeting the Featured Snippets (position 0) works here
– Content should read like a conversation and be thoughtful and useful. Think like a customer and help them problem solve.
– Be trustworthy and authoritative. Use reviews and engage with people in a meaningful way on social media
Are we going to get to the stage where people will want Alexa embedded into their new home when searching for a house? Watch this space…
Tom took us through the update to HTTP/2 and explained how latency motivated Google to make the first major update to the HTTP protocol in 20 years.
– HTTP/2 can positively impact site speed and it doesn’t require web development resource as you can roll out a CDN like Cloudflare for easy deployment.
– HTTP2 requires HTTPS – they are ways to do it but all major browsers require HTTPS to use HTTP2
– Moving to HTTP/2 is not a migration. Links don’t need to change as they’ll be HTTPS already
– Googlebot won’t benefit from this, but it will notice.
Google is making a play to turn image search into product discovery
1 in every 5 searches has an image block, training consumers to see that image results are there and important.
Intercept the consumer at the point of desire, before users try and search for something with keywords. Some examples of this are:
– Search by camera on Pinterest. For example, if you take a pic of red shoes, it will show you similar shoes and tell you how to style them
– Samsung understands this and has created Bixby, powered by Pinterest Lens
– Google Assistance also has Google Lens. Their AI are about to launch functionality similar to Bixby and this will challenge text-based search
Google needs to have high quality images and if we use product structured data, this will improve the likelihood of these showing.
In short, think in visuals, not in keywords.
Voice devices are getting adopted faster than mobile. Wow, that’s quite a statistic. So what makes a good aligned Alexa strategy?
– Get voice device yourself to test for yourself or at least use https://echosim.io
– Focus on featured snippets (position 0) and schema markup for voice search and align with your featured snippets strategy
– Register your brand terms across voice platforms now before they’re taken by someone else (as these can’t be duplicated!)
Voice search shouldn’t be used as a stand-alone strategy but more to support the other digital marketing activities you’re already carrying out.
Gianna takes us through issues with Google Translate and why artificial intelligence can’t always get it right.
– Number formats (decimal points, for example)
– Document structure (salutations in letters)
– Different legal requirements for contracts
– Brand names
– Synonyms and connotations
– User intent
The main things to consider that AI can’t pick up on:
A key point from this talk was: Don’t invest in technical if you aren’t willing to invest in content that can be read internationally.
– Computing power means a lot of servers
– Prone to issues
– Downtime = ranking loss
– More complex and difficult from an SEO perspective
– Requires a lot of SEO knowledge
Why do log files make a massive difference if your domain is bigger than 50k URLs and larger? They help to:
– Understand crawl priorities
– Prevent reduced crawling – behaviour. frequency – if something’s wrong
– Understand global issues – can find patterns and crawl shortcomings
– Ensure proper crawling – you want to make sure Google is crawling everything important
– Ensure proper linking – you want to pass link equity one from URL to the other – and no redirect chains, etc
In summary, log files help to:
– Give more insight than a crawl
– Spot anomalies in specific timeframes to allow for further investigation
– Find what the top crawled pages by Googlebot are
Fili takes us through the do’s and don’ts of Schema Markup.
– Google Tag Manager is a good way to inject Schema, however it isn’t there for every crawl so it won’t appear each time
– Data highlighter is dangerous as we could change content and forget to update it and no one but Google sees this
Both are handy to experiment with but it’s better to implement it within the code.
Don’t mark the same information up with both in-line and JSON-LD as this will create conflicting signals. For example, don’t point to two different URLs within Schema. The structured data testing tool will tell you that it’s correctly implemented, even if it isn’t actually correct in terms of context.
Schema is always growing. pending.schema.org will show you the Schema that’s about to be released.
– Don’t exclude recommended fields as they could be a required field in the future
– Avoid marking up hidden content
– Use deep crawl and screaming frog to extract the way schema has been implemented
– Can also use structured data outside of a webpage. Can be incorporated within an email.
Why do it? Putting it simply, it will help Google to understand your content.