- Google Discover is not classic SEO; it rewards scroll-stopping topics, titles, and images, not keywords and backlinks.
- Your headline and lead image decide 80% of your Discover traffic, so you must design them for curiosity and emotion, not just relevance.
- Audience signals power Discover; you win by building real engagement and testing, not by flooding the web with more articles.
- Discover is volatile but still one of the biggest free traffic levers for publishers that learn, adapt, and treat it like its own channel.
Google Discover can send more traffic from one article than a year of ranking for mid-tier keywords, but it plays by its own rules, so you need to think less like a traditional SEO and more like a YouTube creator who understands feeds, hooks, and retention.
What makes Google Discover different from Google Search
Think of Google Discover as a personalized newsfeed where your article has to win a tiny auction for attention in a fast scroll, instead of a search results page where users already typed in a question and want a specific answer.
Search vs Discover at a glance
| Aspect | Google Search | Google Discover |
|---|---|---|
| User behavior | User types a query and compares results | User passively scrolls a feed with no query |
| Core question | “Does this answer my question?” | “Do I care enough to stop and tap?” |
| Main hook | Keyword relevance and SERP intent | Emotion, curiosity, timing, image |
| Trust signal | Backlinks, on-page signals, history | Engagement, repeat visitors, click behavior |
| Content style | Structured, thorough, intent-focused | Angle-driven, sharp hook, strong narrative |
On Android, Discover sits in front of people all day, and that constant visibility makes it powerful, but it also means mediocrity disappears fast, so if your content feels like generic SEO copy, the feed will quietly stop testing it.
Discover is closer to a TikTok or YouTube home feed than to a classic search results page, so stop expecting keyword tricks to work there.
I know the SEO instinct is to think in terms of queries and clusters, but with Discover, it helps more to think in terms of hooks, angles, and emotional triggers than in terms of exact match phrases.
Stop treating Discover like SEO and start treating it like a feed
If you plan Discover content with the same mindset as your evergreen SEO pages, you will likely underperform badly, because the feed is trying to guess what people might like, not answer what they just asked.
Discover serves people, not keywords
Every card on Discover is based on a pattern of behavior around topics, devices, locations, Chrome history, YouTube history, and a messy mix of signals we will never fully see, which is uncomfortable but also a big chance if you like testing.
Instead of thinking, “What keyword does this target?” ask, “What type of person is this for, and what are they already obsessing over this week?”
Think like a social content creator
Social creators do not start with “What keyword should I rank for?” but with “What headline and thumbnail will get my audience to stop?” and that mindset is much closer to what Discover rewards.
When I look at publishers who crush Discover, they act more like small YouTube studios than old newsrooms: they test hooks, recycle strong angles, and are not scared to kill ideas that do not move numbers.
Discover rewards content that is both relevant and emotionally charged, not dry recaps that feel written for a textbook.
Where SEOs overcomplicate Discover
Many teams try to reverse engineer every little bump or drop with complex theories, but in practice, the patterns that matter are usually simple: people either care and click, or they scroll past without a second thought.
If you track your best Discover hits for a few months, you will likely spot recurring themes like repeating actors, recurring storylines, specific content formats, and times of day that your regular readers respond to first.
A simple mental model
Use this quick model when planning a piece that you hope will land in Discover.
- One clear audience segment: know exactly who this is for.
- One sharp emotion: curiosity, fear of missing out, surprise, or relief.
- One strong promise in the title: what they get if they tap.
- One image that reinforces that same emotion.
If any of these are fuzzy, your chances drop, because Discover does not give you the luxury of a second impression; your card either works in that half-second or disappears into the feed.
How to craft Discover titles that actually earn clicks
Your Discover title is not a label for a search result, it is a tiny pitch to stop a distracted thumb on a crowded feed, so you need to build it around emotion, not only information.
Why classic SEO titles fall flat
A typical SEO headline like “Best standing desks for home offices” works fine when someone just searched “best standing desk” and is ready to compare products, but in Discover, that same line blends into the background because there is no tension or story.
The person scrolling Discover might be half-interested in productivity, but they are also looking at sports, local news, and recipe ideas, so you have to give them a reason to care right now.
The one emotion, one promise rule
For Discover, I like to use a simple rule: your headline should focus on one main emotion and one clear promise, not three weak ideas packed together.
Here are some patterns that tap into that.
| Intent | Dry SEO-style title | Discover-friendly title | Emotion |
|---|---|---|---|
| Productivity | “Best morning routines for productivity” | “I tested 7 morning routines for 30 days, and only 1 actually helped me work less” | Curiosity + relief |
| Personal finance | “How to save money on groceries” | “The grocery rule that quietly saved my family $3,200 this year” | Surprise + gain |
| Health | “Benefits of strength training” | “My doctor begged me to lift weights for 6 months, here is what changed” | Personal story + intrigue |
| Tech news | “New AI tool launched for writers” | “Writers are quietly using this AI shortcut, and clients have not noticed” | FOMO + mild controversy |
You might notice that these titles are usually between 70 and 95 characters, long enough to carry a real hook but short enough to avoid getting cut off in the feed.
A strong Discover headline is not about tricks; it is about stating a clear, emotional promise and then fully keeping it in the article.
Headline ingredients that tend to work
There is no magic formula, but some ingredients keep showing up in high click-through rates when we test across different niches.
- First person angles: “I tried,” “We tested,” “My coach made me”
- Clear numbers with context: not just “7 tips,” but “7 changes that cut my rent by 18%”
- Hidden or unusual angles: “what nobody tells you about,” “the rule my accountant hates”
- Contrast: “I spent 3 years doing X, then one small change fixed it”
- Direct reader benefit: “here is the script you can copy,” “the checklist I send clients”
I sometimes see people swing too far and create titles that feel clever but shallow, and those may spike clicks for a day, then collapse once users realize there is no real value inside.
A simple test for your next Discover title
Before you ship an article that you hope can hit Discover, ask two blunt questions.
- If I read only this title, without context, would I save it to read later?
- Does the first screen of the article clearly deliver on the promise of the title?
If the answer to the second question is no, fix that before doing anything else, because clicks without satisfaction are a quick way to send bad engagement signals back into Discover.
Why your images can rescue or ruin Discover performance
Your image in Discover acts like a YouTube thumbnail that anchors emotion and context, and if you treat it as a last-minute decoration, you are probably leaving a lot of traffic on the table.
What a Discover image must do
A strong Discover image earns its place by making someone pause and feel something that matches the headline, without confusing or misleading them.
That means the bar is a bit higher than just “not ugly” or “on-brand”; it has to pull weight in the tiny real estate of a mobile card.
Stock vs custom: why it matters
Generic stock photos tend to feel invisible because people have been trained to ignore them, especially the ones that look like stiff business meetings or staged pointing at screens.
Publishing teams that switch from bland stock to tailored visuals, even simple ones like close-up hands, real workspaces, or annotated screenshots, often see a nice lift in Discover click-through, sometimes faster than any change to the copy.
| Type of visual | Example idea | Why it helps |
|---|---|---|
| Human faces | A real person reacting to a bank statement or a cluttered room | Faces catch attention and convey emotion instantly |
| Before / after | Side-by-side of a messy desk vs a focused workspace | Visual story that reinforces the promise in the title |
| Annotated screenshots | A tool interface with a big circle around the crucial button | Signals practicality and concrete help |
| Simple charts | Clean chart showing “time spent” dropping over months | Backs up results claimed in the headline |
| Objects with context | Shopping basket with highlighted items from the story | Connects the abstract topic to real life quickly |
Size and clarity
Google recommends large, high quality images like 1200 x 800 pixels so they can show properly on bigger screens, which is easy to hit with modern devices and design tools.
What matters more in practice is clarity: the subject should still be recognizable on a small phone screen, with no messy clutter that looks like noise in a tiny thumbnail.
Ask yourself a harsh question: would you stop scrolling for this image if it did not belong to your brand?
Faces, feelings, and context
Faces almost always pull extra attention, but they should match the mood of the story, instead of just being a smiling person next to a laptop for no reason.
If your article is about stress from hidden subscription fees, then a worried or skeptical expression works better than a generic happy pose that tells no story at all.
Testing images in a lightweight way
You do not need a big design team to test images; a small set of templates and a simple internal rule can be enough for steady improvements.
- Create 5 to 10 visual layouts that fit your brand: face close-up, object close-up, chart, screenshot, before/after, etc.
- Rotate them across different topics and track which ones correlate with higher click-through in Discover.
- Document what seems to lift performance, even if you are not 100% sure why yet.
Over a few weeks, patterns usually show up, like certain colors popping more, or particular compositions that your audience seems to notice faster.
Audience signals: the real fuel behind Discover traffic
Discover is not a charity traffic source; it needs proof that your content resonates somewhere first, often with your own audience, before it risks showing it to a larger pool of readers.
Why engagement beats volume
Publishing more content can look productive on a dashboard, but if most of those pieces get weak engagement, the overall signal that your site sends to Discover can get diluted fast.
I have seen bigger news sites that publish a hundred small stories a day lose to leaner teams that release several strong pieces with careful angles and consistent follow-up, and that pattern is not rare.
The sniper approach to topics
For Discover, broad is usually weak; tight focus on topics that your readers already care about gives the system a clearer picture of who to test your next card on.
You can think of each new article like a sniper shot at a known interest cluster: very specific subject, clear timing, and a message tuned to people who have engaged with similar stories before.
Publishing fewer, sharper articles that your readers finish and share often beats flooding Discover with forgettable stories.
Signals that help Discover trust your content
You cannot see the full ranking recipe, but some signals keep lining up with stable Discover traffic when we look across sites.
- Return visitors who spend time on the article, not just fly-by traffic.
- Internal recirculation: readers move from one article to another on your site.
- Direct traffic and email clicks: loyal audiences clicking from your own channels.
- Mentions and shares on other platforms that reflect real interest.
If you treat Discover as a bonus that lands out of nowhere, you might miss that your list, your app, or your social accounts can help prime that first wave of engaged readers.
Use your own audience as a Discover engine
One practical approach I like is to design each Discover candidate piece with a quick distribution checklist: newsletter feature, push notification, pinned social post, or a short YouTube teaser that leads people back.
That first 500 to 5,000 readers that care about you already can send a strong enough pattern that Discover starts testing the card to people who look similar, which is where the real numbers show up.
When Discover traffic drops hard
Feed products evolve fast, and Discover updates can be rough, especially by region or category, so traffic can drop 40 or 60 percent in what feels like a single night.
The worst move I see teams make is to panic and rewrite everything blindly or, oddly, to keep publishing the same way while hoping the old numbers come back on their own.
A calmer way to respond to volatility
When you notice a big Discover slump, try separating your content into three simple buckets and responding differently to each one.
- Evergreen hits that still have good engagement: keep them live, maybe refresh examples or images.
- Recent posts with weak metrics: either improve titles and images or accept that the angle missed and move on.
- Strong topics with weaker hooks: rebuild the hook with a new story format, like a test, a diary, or a “things I would do differently” piece.
You also want to look at where drops are strongest: category, country, device, or language, because sometimes the change is less about your quality and more about Google shifting traffic between regions or types of publishers.
Why selling something still matters
Relying only on ads around Discover traffic can feel fragile, because when the feed pulls back, your revenue takes the hit directly, without any buffer.
I am biased here, but I think publishers are safer when they connect their top-of-funnel Discover pieces to clear offers like courses, tools, events, or services, so the extra visits actually turn into stable customers over time.
Discover traffic feels large, but unless you connect it to owned channels and real offers, it can fade as quickly as it arrived.
Turning Discover readers into long-term assets
You do not control the feed, but you control what happens after the click, so part of your Discover strategy should be on-page captures that feel helpful, not intrusive.
- Content upgrades: a short checklist or template directly related to the article.
- Soft newsletter invites that promise ongoing coverage of that specific topic.
- Subtle product mentions where they actually solve the problem discussed, not bolted-on ads.
If your Discover article sends readers into a dead end with no way to stay connected, you are losing the real upside of that spike in interest.
Putting all of this together into a practical Discover game plan
Google Discover rewards publishers that respect the feed: strong hooks, focused topics, images that matter, and a calm response to volatility, instead of blind faith in keywords and volume.
A simple workflow you can start using this week
You do not need a huge team or complex tools to build a Discover-friendly process; you just need a routine you can repeat and refine.
- Step 1: Pick topics from real audience interest
Look at your top articles by time on page, email replies, and social comments, then plan new angles around those same subjects, not random trends you feel nothing about. - Step 2: Draft multiple titles before you write
Write at least 5 to 10 Discover-style headlines before drafting the article, then pick the one that feels most specific and emotional, not just descriptive. - Step 3: Design the image to match the emotion
Choose or create an image that clearly lines up with the headline promise, with a visible subject and clean composition that still looks clear on a phone screen. - Step 4: Front-load the value in the first screen
Use your opening lines and subhead to prove that the article keeps the promise made in the title, instead of warming up with generic context. - Step 5: Launch with your own audience first
Share new Discover candidates through email, social posts, and any apps or communities you run, so you generate real engagement signals fast. - Step 6: Track by topic, hook, and format
Label each article in your internal sheet by topic, emotion in the title, and content format, then review Discover stats each week to see which combinations lead to traffic and which die quietly. - Step 7: Keep your offers close to your Discover hits
Make sure every high performing Discover article points readers toward a related product, service, or resource that helps them act on what they just learned.
I know this might sound like more work than just writing another SEO article and hoping it ranks, but once you get into the rhythm of planning for Discover, your content tends to sharpen overall, which often helps organic search too.
You will probably have misses and random surges that make no sense at first; that is normal, and it is better to treat those as clues for your next round than as proof that the system is broken or that nothing you do matters.
If you see Discover as a channel you can learn and shape over time, not a mysterious lottery, you give yourself room to experiment without panicking at every swing.
Feed products reward persistence and curiosity; if you keep testing titles, refining images, and listening to what your audience actually responds to, you can build Discover into a steady part of your traffic mix instead of a one-time spike that you never manage to repeat.
The real edge comes from knowing your readers better than anyone else in your niche and having the patience to ship content that respects their attention, even when that means saying no to easy, forgettable stories that might get you a short burst of clicks today but teach the algorithm nothing about why your work deserves space in tomorrow’s feed.
Need a quick summary of this article? Choose your favorite AI tool below:
