Google Discover keeps showing up in traffic reports as a mystery channel. It sends huge spikes of visits, then goes quiet for weeks. Most publishers treat it like luck.
It is not luck. Google Discover optimization is a real practice with its own rules, and those rules are different from anything you know about ranked search.
This guide breaks down what Discover actually is, what kind of content it favors, and how to build a Google Discover optimization approach that survives the feed’s mood swings instead of getting wiped out by them.
What Google Discover Is and How It Differs From Search
Google Discover is a personalized content feed that lives inside the Google app and the mobile homepage. There is no search query attached to it.
Google decides what to show a user based on their search history, app activity, and interests, not based on something they typed.
This is the single biggest thing to understand before attempting any Google Discover optimization work. Traditional SEO ranks pages against a query. Discover ranks pages against a person.
Two users can see completely different stories on the exact same day, from the exact same site. A few structural differences matter here, and each one shapes what a working Google Discover optimization plan actually needs to include.
- There is no keyword to target. Discover does not match content to a search term, so classic on page keyword mapping barely applies.
- Mobile only. Discover is a mobile surface, so any Google Discover optimization work has to assume a mobile screen and a thumb scrolling fast. Desktop traffic from this channel does not exist.
- No click through from a results page. Users tap a card inside a feed they were already scrolling, which changes what makes them stop.
- Topic authority matters more than page authority. Google leans on your site’s track record covering a subject.
Because of these differences, a Google Discover optimization plan cannot just be an extension of your normal content calendar. It needs its own decisions about topic choice, timing, and visual presentation.
Teams that skip this step tend to apply search logic to a feed that does not work like search, and the results usually show it. A page can rank well and still never appear in a single user’s Discover feed, simply because it never matched an interest signal Google was tracking for that person.
Content Types That Perform Well in Discover
Discover rewards two very different kinds of content, and confusing them is where most sites waste effort.
- Evergreen explainer content. Broad, curiosity driven pieces that answer a question a large audience already has. These can resurface in the feed for months.
- Timely news and trend pieces. Content published close to a moment of public interest, where speed and freshness carry more weight than depth.
What almost never works, and what undermines most Google Discover optimization attempts from the start, is thin, generic content written only to hit a topic.
Google’s own guidance for this feed leans heavily on E-E-A-T signals, meaning the content needs a visible point of view, real expertise, and a byline that reads like a person wrote it.
A practical Google Discover optimization checklist for content selection looks like this, and it applies whether you run a news site or a niche publication working through Google Discover optimization for the first time.
- Pick topics with proven search or social interest, not guesses
- Cover the topic with a distinct angle, not a rehash of what already ranks
- Publish consistently in a subject area instead of one off pieces
- Attach clear authorship and expertise signals to the piece
Sites that publish across wildly unrelated topics tend to see erratic Discover performance. Sites that build a recognizable lane, even a broad one, tend to get pulled back into the feed repeatedly.
This consistency is often the difference between a one time spike and a Google Discover optimization program that keeps paying off month after month.
Image and Headline Requirements for Discover
Discover is a visual feed first. A card with a weak image gets scrolled past in under a second, regardless of how good the article is.
This is the part of Google Discover optimization that most teams underinvest in. Google has specific technical requirements for images to even be eligible for large, high visibility cards, and skipping these is one of the most common Google Discover optimization mistakes teams make.
- Images must be at least 1200 pixels wide
- The max-image-preview meta tag needs to be set to large
- Images should be high quality and relevant to the actual content, not stock filler
- A featured image tag should be present in the page’s structured data or standard HTML image markup
Headlines matter just as much as the image in Google Discover optimization, but for a different reason. A headline written for Discover has to work with zero context.
There is no surrounding search results page giving the user hints about intent. The headline needs to state the value clearly without resorting to bait.
This is one of the clearest rules in Google Discover optimization. Clickbait style headlines tend to hurt Discover performance over time.
Google’s systems appear to weigh post click engagement, so a headline that overpromises and an article that underdelivers can suppress future visibility. Any serious Google Discover optimization effort treats the headline as a promise that has to be kept, not just a hook.
Why Discover Traffic Spikes and Disappears
This is the part that confuses almost every content team. A page can sit quietly for weeks, then suddenly deliver tens of thousands of visits in a single day, then drop back to almost nothing.
This happens because Discover distribution is driven by real time interest signals across a huge number of users, not by a fixed ranking position, which is exactly why Google Discover optimization has to be judged over months rather than by any single day.
When a topic starts trending among people with related interests, Google can surface a relevant page to a much wider slice of its user base than that page would ever reach through search alone.
The flip side is that this interest window closes. Once the trending moment passes, or once a newer, fresher piece captures the same interest signal, the older page drops out of the feed almost as fast as it entered.
A few patterns show up consistently, and understanding them is a core part of any serious Google Discover optimization work.
- Fresh publish dates get preference during a trending window
- Evergreen pieces can get repeat waves of traffic months apart, tied to renewed interest
- A single spike rarely repeats on the same page unless the topic resurfaces
- Sites with strong topical history in an area seem to get pulled into more of these windows
Because of this volatility, treating Discover traffic as a core forecasted channel is risky, and this is a point worth repeating in any Google Discover optimization plan presented to leadership.
The more realistic way to think about Google Discover optimization is as a multiplier on top of a content strategy that already works for search, not a replacement channel you can plan a business around.
Measuring Discover in Google Search Console
Google Search Console has a dedicated Discover report, but it only appears once a site has received a meaningful amount of Discover traffic, so newer or smaller sites may not see it at all.
The report shows impressions, clicks, and CTR specifically for Discover, separate from the regular Search performance report. This separation matters because Discover CTR behaves nothing like search CTR.
A page can have a low CTR by search standards and still be considered a strong performer in Discover, since the baseline expectations are different.
A few things are worth tracking closely inside this report, since this is where most of the real Google Discover optimization decisions get made.
- Which pages are pulling repeat Discover impressions over time
- Whether spikes correlate with a specific image or headline change
- How CTR shifts after a headline or featured image update
- Which topics keep resurfacing across multiple articles
Because there is no keyword data tied to Discover clicks, teams have to rely on page level patterns instead of query level ones.
This is a meaningful shift for anyone used to working from search query reports, and it is a core part of learning Google Discover optimization from real data instead of guesses.
It also helps to track this data over a longer window than you might for search. A single week of Discover numbers rarely tells you much on its own.
Looking at a rolling ninety day view of the report tends to reveal which topics and formats are actually earning repeat visibility, which is the kind of signal a real Google Discover optimization program should be built around.
