Tom Anthony - Dotting the (A)I frenzy

by Gianluca Fiorelli

Tom Anthony evangelize about the new shapes Search is assuming since few years. At The Inbounder London he will tell us what SEOs should do in order to follow up with the evolution of Search.

Many times SEOs are considered as some sort of alchemists, who are trying to discover the secret formulas behind the Google search ranking factors.


Many times that perception is correct, and we all know the myths created by SEO (Keyword Density or LSI anyone?).


However, there are also SEOs, who can legitimately can have a trustable voice about how search engine works and how the search industry is evolving.



One of these persons surely is Tom Anthony, Head of R&D at Distilled, and not only because he is member of the Adaptive Systems Research Group at the University of Hertfordshire and researching on Perception-Action Loop of Evolved Agents using an Information-Theoretic approach (I've headache only reading it), but also because he actually experiments with Machine and Deep Learning algorithms and SEO A/B Testing.


If you are imagining Tom Anthony somehow as the classic nerd-looking guy, well... somehow you are right. 


However, do not let your first impressions blind you: Tom is one of the smartest mind in SEO and one of most giving and generous person in the industry. This is why I am particularly proud of having him speaking at The Inbounder London on May 2nd.



Since few years now, in your talks and posts you are underlining how much the search landscape is dramatically changing. 

From the description of your The Inbounder speech, this change is going to be even more evident especially due to the use of Machine Learning and the rise of Personal Assistants.

On the other hand, and despite of the general enthusiasm SEO are showing for ML, I see still some sort of resistance by those same SEO in transforming this generic frenzy into something that really influence their daily job.

What do they risk from this sort of passive attitude toward the change Google and others search engines are experiencing?



On one hand I think people are bored of hearing about Machine Learning; we've been getting excited about it for a 2-3 years now and people are wondering when is this dramatic impact going to actually happen!?



However, I think that is is very similar in that way to 'mobile' - we talked and talked about it and waited for the 'year of the mobile' but what we actually got was a transition into the era of the mobile. 


With mobile was no watershed moment where everything dramatically changed, but now we are in a mobile first industry. Machine Learning isn't going to have a 'moment' but 2-3 years from now everything we do will have been shaped by Machine Learning.


There are two sides of it that will effect us. Firstly, its increasing impact on search engines (both the traditional web search engines, but also emerging paradigms of chat bots, Intelligent Personal Assistants, and others). Secondly, will be the increasing opportunities for us as marketers to deploy Machine Learning to help us.


Keyword research, personas, analytics, and other core parts of SEO will be able to leverage it, and in ways that are entirely accessible where you don't need to learn about neural networks, Tensorflow, or anything else like that.



I usually say that SEO won’t ever die, because it is strictly related to a basic human need: searching for answers.

What will change is the medium through which people will ask and try to find those answers.

As these media seem multiplying themselves into different forms, you aim people to move to a hub and spoke-API approach. Could you give us an example of this approach?



Sure! Imagine that you are a health and fitness website, and you are concerned with capturing new users.


You've traditionally had this website that hosts your service for paying users, but also has dietary, nutritional and exercising information to try to capture searches by people who may be interested in your service.


Well now, these people are maybe going to ask Amazon Echo about nutrition whilst cooking, or have wonder if they burn more calories when it is sunny whilst running and speak that query into a smart watch. Maybe they want to see a video of the correct form for an exercise on their tablet whilst at home, or perhaps (gasp!) they are on a laptop computer and searching for some information.

Each of these searches has a different set of input to output. Echo is voice in, voice out, whilst the smart watch is voice in, display out. The tablet and laptop are both different again.


There are now two distinct considerations - our raw data and the interface to that data. Traditionally with website focused SEO we have dealt with them as one entity, but we need to start thinking about changing that. With a hub and spoke model you can start setting yourself up for that even when we don't know exactly what is coming next.




Gary Illyes recently said that Google won’t probably entirely base its algorithm on Machine Learning, because the biggest problem of Machine Learning is debugging.


I am sure – despite of the probable scepticism you may have about this declaration – that you may agree with the problem of feeding the Machine Learning algorithms with biased data, so my question is: how can SEO learn to debug what presented by machine learning based tools?



I don't think they will need too. Gary is absolutely right - it is a real pain to debug ML at times.


At The Inbounder I will share an example of how challenging it can be based on my own days in academia when I faced challenges doing it.


For Google they have traditionally liked to understand the algorithm, and that was especially driven by Amit Singhal who was previous Head of Search until last February.


John Giannandrea, who replaces Amit, was previously Head of Machine Learning at Google - so it is hard not to believe that ML is going to have an increasingly massive influence on the algorithm.


So it is going to be a challenge for them, for sure, and it will be very interesting because I think we'll see a period where there is some regression in terms of the ability for black-hat leaning marketers to manipulate the algorithm more. 



Finally, tell me in 140 characters why people should come to The Inbounder London on May 2nd

.@theinbounder in Valencia was amazing. Now the same quality and talent is coming to London for a crazily hard to beat accessible price. 

The Proust Questionnaire


What is your favorite word?

Schnürsenkel! It means shoe laces in German but it sounds like a big furry creature!


What is your least favorite word?

I'm not sure I have one - most words have a time and a place.


What turns you on?



What turns you off?

Compliance paperwork 


What sound do you love?

The children giggling. 


What sound do you hate?

Polystyrene being rubbed together! 


What is your favorite curse word?



What profession other than yours would you like to attempt?

I can't do it because it is too late, but being a medical doctor would be cool.


What profession would you not like to do?

Patent lawyer 


If heaven exists, what would you like to hear God say when you arrive at the pearly gates?

"Yes, of course we have the second season of Firefly!"



Gianluca Fiorelli
SEO Consultant
Keep me updated
Subscribe to our newsletter to receive all the news for the next The Inbounder event.