Keywords will die, welcome your machine overlords and have a happy 2018

Jim Kirby

February 5, 2018

Anyone who says that 2017 wasn’t a roller-coaster is lying. What we can be most certain of in 2018 is change itself. It’s going to be a wild ride. Working in the strategy and insight team, we think that we’ve got some chops in this area between our time in trendspotting, forecasting and opportunity analysis…


…that and our chronic penny-arcade addiction…


So here it is, what you can look forward to (or be part of building) in 2018.


Audience targeting


I’m going to go big or go home: we’re going to see keyword targeting in search marketing die this coming year.

Why? Google realised a long time ago that a search for “best mouse” means different things for rodent enthusiasts and for PC gaming enthusiasts. The solution to this is personalisation. Google knows a bunch about you and whether you like this or not, it means you’re going to be able to select more and more abstract demographics rather than approximating these with keyword targeting. Facebook is ahead of the game on this by advertising against their massive demographic database, Google is going to catch up this year, probably through rolling out more demographic filters in AdWords and if it’s being ambitious then it will offer a pure demographic targeting option probably for less CPC to encourage uptake.


Google can even tell I’m a productivity freak rather than a gamer


The question is :

Why bother with keyword targeting when google and other services aren’t using keywords to serve them data anyway?


Voice search SEO



Everyone is telling you it’s going to be big. We’re telling you how to make the most of it.

Get in answer boxes. Write your content to answer questions. Make your site have crawlable schema and try to keep an eye on the search engine splits in search console because voice search is definitely going to turn up there soon. If you have the flexibility then run tests and learn how to target the answer boxes and voice answers in your niche, if not then learn from how your competitors are doing it. Hopefully there’s going to be an option to open answers on phones or laptops to enable user retention but that’s up to the voice search developers.

Search analytics chart

coming soon to a search console near you


Data products


2017 was the year of data science. It was the sexiest job and it’s the reason for the rise of R and Python (the key languages used in data science) in coding language popularities this year.

Google Trends chart showing Python VS R

Google Trends with polynomial forecast


However, not every business can or should have a data scientist in house. Luckily software as a service is going to solve all this. In 2018 we can look forward to having user-friendly SaaS products or plugins to Excel or Google Sheets that will do simple machine learning and statistical analysis based on a question-based model.

You’ll input your keywords and the product/plugin will group your results and find the most cost-efficient topic for investing budget into. You’ll input your list of keywords and SEMRush or a competitor will group them into hub topics. The methodology is already out there in researchers, capitalism is going to productise, distribute this knowledge and make bucks out of it.


Machine learning is going to make your life better; trust me.


It’s going to be built into your data processing tool of choice (Excel is getting it in Q1  and Sheets shouldn’t be too far behind it – and for good reason… The good news is that you’re going to be able to get better insights faster from your data by augmenting your field knowledge with a machine that’s willing to trawl through thousands of rows looking for a quick win. The bad news is that your client is going to spot your problems quicker so you’re going to need to get on this bandwagon or face some awkward calls about proactivity…


Get on board the train of machine learning, data mining and predictive analytics or miss it.


Talking of data mining, if you’ve not heard about GDPR yet then where have you been all year? New rules on data use will affect every business that captures PII (personally identifiable information). We’ve not got space to cover it all here, but there are some useful blogs we’ve linked to below worth reading. Compliance isn’t optional and poor planning could cost you a lot of money if you haven’t got your house in order.


Read more here:


GDPR and digital marketing:

GDPR and web analytics:

GDPR official guide from ICO:


On the other, more positive, hand from a data modelling perspective, there are rumours that tracking the impact of channels and campaigns is getting just that little bit easier;


Attribution 360 from Google is on the way!


We don’t know the feature set but our wish list is that it will have an API backend to pull data into Data-Studio and Google Sheets. This would mean a Supermetrics plugin and potentially an actual reason to recommend a client to buy 360 in the first place! We’re unlikely to get back to the good old days of linking keywords to conversions (something that AudienceView from Connexity can get close to) but we’re likely to see un-sampled data on first click attribution for your monthly reports


So, what are your new year’s resolutions?


  1. Look into optimising for voice search
  2. Look into automating repetitive data analysis
  3. Give up learning to code in R or Python, excel is probably going to do it for you.

Joking aside, it’s time to look into drawing out a data pipeline, think about the regular questions you’re going to be asking and making the process as smooth as possible; strategy is for life not just for Christmas.

Related Articles