If you have used the Google Search app on Android or iOS devices since the start of 2017, it’s likely you’ve encountered Google Discover in one of its several different forms.

The curated content feed uses machine learning (also known as artificial intelligence or AI) to recommend web pages it thinks you might be interested in.

It does this in several ways:

  • Based on your past behaviour.
  • Based on your current location.
  • Based on local and global trending topics.

The aim is to give you a screen filled with pages you’ll want to read. These can be serious, such as headline news stories, or simply for entertainment.

Over time, Discover learns from you, to further improve the relevance of the content it suggests. In this article, we’ll look in more detail at Google Discover, how it works, and its implications for webmasters and SEO.

What is Google Discover?

The history of Google Discover can be traced back at least as far as December 2016. That’s when a post on the official Google blog announced an update to the search engine’s mobile app.

Two new sections were added: one, an area for upcoming personal events like appointments and flights, and the other a feed with curated updates on news, sports, entertainment and so on.

By July 2017, Google had improved the artificial intelligence behind this feed and introduced some new capabilities, such as the ability to recommend items not based on your personal interactions with the app, but on more contextual data like your current location.

In September 2018, the feed was given a name: Google Discover. The general design was updated, with new ways to deep-dive into a topic of interest, and ways to explore topics directly from the app’s search results rather than going to the Discover tab first.

Discover has always aimed to provide a vibrant and engaging, personalised alternative to the traditional process of searching, and it continues to put more value on visually impactful pages including the relatively recent introduction of Web Stories.

But at its core, Discover is still what it has always been: an automatically curated feed of web pages designed to match closely with the user’s personal preferences and interests.

How does Google Discover traffic vary from ‘traditional’ traffic?

Discover is intended to give users results tailored to their interests, without even needing to run a search for them.

For content publishers, the flipside of this is that traffic from Google Discover is likely to consist of people who are already highly engaged with your content on arrival.

This is reinforced by the more graphics-based and colourful nature of tiles on the Discover tab, as well as the ability to like and share results the user finds particularly interesting.

In this way, Discover is similar to a social network, but instead of following other human users, the home feed is curated by AI.

Segregating Google Discover traffic in your website analytics can be difficult; however, Google Search Console provides some data that can be helpful when optimising campaigns for Discover and planning new content that targets this channel.

Understanding Google Discover traffic to your site

Patterns of traffic from Google Discover can appear quite erratic, but there are a few common characteristics that can be helpful if you are trying to optimise your website for Discover:

  • Not all of your pages will be included in Discover results (no matter what you do).
  • Traffic will typically spike just a few days after a page is published (3-5 days).
  • Residual traffic may continue long-term if the content is evergreen.

Unlike organic search traffic, you usually won’t see a gentle curve or ‘ski jump’ shape if you chart your number of hits over time.

Instead, a typical page will have a few days of zero hits from Discover, followed by a high peak lasting 2-3 days, which then drops off completely.

A further 2-3 weeks later, the residual traffic starts to arrive, and if your content is evergreen or remains timely, you should achieve a low but steady long-term average of arrivals via Discover.

Keep publishing content that does not lose its relevance over time and you can accumulate a much larger amount of residual long-term traffic across your entire website.

How to generate Discover traffic to your site

Google Discover is a relatively new platform and as you can see from the timeline at the top of this article, it’s evolved quite a lot in just a few years.

Producing content for Discover is not an exact science, but by committing to high standards on every page you publish, you can maximise your presence in Discover feeds and generate the most traffic as a result.

All you can EAT

There is no such thing as a single page optimised for Google Discover. That’s because to get your content included, your whole website should score well on EAT:

  • Expertise
  • Authoritativeness
  • Trustworthiness

These characteristics all help you to rank well in the organic Google Search results, so they’re worth your attention on other channels besides Discover.

Generally speaking across Google platforms, other search engines and social networks, the first and best step towards optimising your content is to make sure it is engaging, informative and well written.

What Google says

Unusually, Google provides a fairly detailed list of how to improve your chances of getting included in Discover results, on Google Search Central.

These methods include:

Accurate titles

Pages should have titles that accurately describe the content, without using typical clickbait title structures.

Google say your page title should “capture the essence of the content”. Think about how a reputable newspaper might get the core meaning of an article into a headline of just a few words – but avoid using tabloid techniques to titillate and tantalise users into clicking.

Ethical content

Keep your morals in mind when producing content – you should not attempt to increase clicks by focusing on morbid or outrageous subject matter.

Again, it’s about publishing content that is genuinely engaging on an intellectual level, rather than encouraging people to click through by saying something egregious and socially unacceptable in the page preview.

Added value

Make sure your content has added value in one form or another, for example:

  • Unique insight or professional expertise.
  • Timely coverage of a trending topic.
  • Exceptional story-telling and high-quality content.

Evergreen content can also appear in Discover results, but the biggest short-term gains come from timely content in the first few days after it is published.

Image quality

Discover favours large images and “compelling” content, so try to use high-resolution illustrations or professional photography with vibrant colours.

Images used on your pages should be:

  • At least 1,200 pixels wide.
  • Use the max-image-preview:large setting (in a page-level ‘robots’ meta tag).
  • Alternatively, enable large image previews using AMP.
  • Do not use your website logo as the preview image.

Each page you publish should have at least one large, high-quality image to be used as the page preview, which should not be duplicated significantly across other pages on your site – use a unique image if you can.

Non-clickbait previews

The preview tile for your page should avoid clickbait techniques in the title, snippet text and images, for example:

  • Exaggerating key features in preview images to make them more appealing.
  • Withholding core information so the user has to click through to find it.
  • Providing misleading details that are not actually in the content.

Your preview should be a fair and accurate representation of your content – and if your content is good, this should not be problematic.

Complete metadata

The more information you provide about your page, the better. If possible, try to include as many of the following as you can:

  • Author name
  • Author biography
  • Date of publication
  • Information about parent publication/website
  • Information about the publisher/company
  • Contact information

All of these help to show that your content can be trusted, which will encourage Google to include your pages in Discover feeds.

The future of Google Discover

In October 2020, Google announced that it was integrating Web Stories into Discover in the USA, India and Brazil, with a plan to roll out the immersive, visuals-orientated Stories to more Google products and other countries around the world.

Web Stories now have their own carousel in supported countries in the Google Discover tab, and tapping a tile launches a full-screen, immersive experience driven by multimedia content.

They can be created within mainstream content management systems (CMS) with drag-and-drop editors available for WordPress, NewsroomAI and MakeStories, or can be manually coded by a proficient web programmer.

At the time of the announcement, Google had already integrated over 2,000 websites into the Stories carousel, and had worked directly with major online publishers like Forbes, The Dodo and Now This to build up the amount of content available.

Final thoughts

The trend is one towards multimedia-rich, immersive online experiences, especially on mobile devices that offer unique ways to view content and interact with pages.

As Google Discover continues to evolve – and paralleling trends on Google Search – publishers should focus on high-quality, engaging content, with good media assets and support for mobile devices and screen resolutions.

By putting mobile first and producing the best content you can, you stand the best chance of ranking well in search results, getting featured on Discover, and building your audience over the long term.