Post of the Week
Why Your SEO Process Needs to be Agile - SEO in Focus
This interview was spurred from a great twitter thread here. Kevin's approach is quite contentious, as auditing processes, and the way we consult on them, have largely stayed the same for years. Check out the interview, and see what you think!
Content
Natural Questions: a New Corpus and Challenge for Question Answering Research
A lot of interesting stuff coming from the Google AI blog recently. Google has introduced Natural Questions (NQ), a corpus for training and evaluating open-domain question answering systems. This is so interesting as it give us some insight into how Google is using machine learning to answer search queries, and provide featured snippets.
Tech
Inspect URLs for Search Console Accounts You Don’t Have Access To
Now this some out the box thinking! This posts describes a method for inspecting URLs for search console accounts you don't have access too - check it out!
35 Quick Tips about Web Performance
There will undoubtedly be tools you don't use and processes you don't follow in this brilliant post by João Cunha. For example, the tool in the first tip looks insaaaaaane: https://speedcurve.com/ - can't wait to trial it out.
Tools
How to rank track Google's hotel search
Good one this. Google is forever adding new 'meta search engines' such as hotels, jobs and flights, but how do we track rankings on them? In this post, Dominic Woodman shows you how you can with a custom bookmarklet, before describing how to scale it.
Not a load of links this week, but quality over quantity eh!
Our post of the week is actually an interview with Kevin Indig from the guys at ContentKing. Much like this issue it's quite short, but the message is pretty powerful nonetheless.
Kevin talks about an agile approach to SEO:
"The full agency SEO audits have to stop. The time in which you wait 1-2 months for an audit with 30-40 recommendations is over."
I'd say the majority of you reading this issue still carry out an all singing all dancing technical audit covering every issue you can think of. That's cool, and it's still got value, but do we really need to do it? Is it overwhelming for the client, and is it an inefficient use of client budget?
What if we only gave the client 3-4 recommendations, but really went to town on them. Surely, that's a better approach right? I also think by minimising the volume of technical recommendations, it allows you to better understand and report on progress. Measuring and reporting on 3 issues, is a hell of a lot easier than 100.
Let me know your thoughts on Twitter
Andrew Charlton