Why does Google Discover traffic seem to vanish overnight?

What exactly is Discover?

Google Discover is this personalized content feed that appears on the mobile home screen of the Google app and on google.com in the mobile version. Unlike traditional search, no query is entered by the user — it's Google pushing content based on history, interests, and geolocation.

The algorithm alone decides which content appears, and this decision rests on shifting and opaque criteria. No keywords to target, no SERP to analyze. An environment where traditional SEO methods only work on the fringes.

What explains this extreme volatility?

The volatility stems from the fact that Discover does not rely on a stable search intent. You do not position your page for a specific query — you are in a flow that changes based on dozens of signals: news, content freshness, recent engagement, user behavior.

An article can soar one day because an event brings focus to the subject, then completely disappear the next day without anything changing on your site. This is not a penalty signal, it's the native functionality of the system.

Mueller emphasizes: there is no guarantee of presence in Discover, even if you follow all the official recommendations. The algorithm may decide that your content is no longer relevant to its users, and you will have no lever to reverse the trend.

What types of content have a chance to appear in Discover?

Google has published minimum criteria in its Help Center: content that complies with Google News policies, high-quality large images, no clickbait. But meeting these criteria doesn’t guarantee anything — it’s just the entry fee.

Content that performs well in Discover is typically fresh news, impactful visual reports, and analyses of recent events. Classic evergreen content has much less chance of success, unless it resonates with a sudden spike in interest.

• Discover is not a reliable traffic channel — any editorial strategy depending on it is structurally at risk.
• Volatility is a characteristic of the system, not a malfunction — don’t try to “correct” a Discover traffic drop as you would for the traditional SERP.
• Eligibility criteria are necessary but not sufficient — you may tick all the boxes and never appear.
• Discover traffic can represent significant volumes, but it remains unpredictable and uncontrollable — never build a business model around it.
• No third-party tool can predict or measure your exposure in Discover outside of the Search Console.

Is this statement consistent with field observations?

Absolutely. For years, we have observed sites losing 80 to 95% of their Discover traffic in a matter of days, without any technical or editorial modifications being made. This is not an isolated incident, it is systemic.

Media outlets that have heavily invested in optimization for Discover — image format, publication speed, AMP structure — have experienced phases of euphoria followed by brutal downturns. The pattern repeats: a site may dominate Discover for 2-3 months, then become invisible. No recovery is guaranteed.

What nuances should be added to this position?

Mueller remains aligned with Google’s official stance: caution, no guarantee, normal volatility. However, there is an important unspoken element: certain types of sites seem to benefit from relative stability in Discover — notably established large media with strong authority.

[To be verified]: Google does not provide any data on Discover traffic distribution by type of site, nor on stability criteria. There is a lack of transparency on the signals that influence exposure in the medium term. The official recommendation to “check the Help Center” is insufficient — this document lists exclusion criteria, not optimization levers.

The reality is that Discover operates like an opaque recommendation system where user engagement likely takes precedence over everything else. If your contents generate few clicks or interactions in the flow, the algorithm sidelines you — and you have no direct means to correct this.

In what cases does this recommendation not apply?

If you are a pure news media outlet with real-time publishing capability, Discover may represent a significant complementary traffic source — but never an exclusive one. Traffic spikes related to news can be massive, but they remain sporadic.

For e-commerce sites, themed blogs, and SaaS sites: forget Discover as a strategic lever. You may appear there sporadically if a piece of content hits just at the right moment, but it's never systematically reproducible.

What should you do if you receive Discover traffic?

First, monitor without relying. Activate the Discover report in the Search Console to track volumes, but never build business forecasts on it. Treat this traffic as an unpredictable bonus, not as a stable channel.

If you find that certain contents perform well in Discover, analyze their characteristics: image format, editorial angle, timing of publication. But don’t over-interpret — what works once may never happen again. Reproducibility is virtually non-existent.

What mistakes should you absolutely avoid?

Never revamp your editorial strategy to “optimize for Discover.” Sites that have shifted to short viral content in hopes of capturing this traffic have often destroyed their classic SERP positioning without gaining stability in Discover.

Another classic mistake: panicking during a drop in Discover traffic and urgently modifying technical elements (images, Core Web Vitals, structure). In 99% of cases, it changes absolutely nothing — the drop is algorithmic, not technical.

Finally, do not compare yourself to competitors who are excelling in Discover. You do not know how long it will last for them, and trying to replicate their approach is likely to lead you into a dead end. Each site has its own trajectory in this flow, and it is largely uncorrelated with classic SEO quality.

How can you build a resilient SEO strategy in the face of this volatility?

Massively focus on stable transactional and informational queries — those that generate predictable traffic year-round. Discover can be a temporary amplifier, but your traffic base must come from traditional search.

Diversify your sources: newsletters, organic social networks, editorial partnerships. A site that depends 70% on traditional Google SEO is already fragile — if it also depends on Discover, it becomes structurally precarious.

If you operate in a highly competitive sector or your business stakes rely on reliable traffic forecasts, these technical and editorial decisions can become complex to manage alone. Hiring a specialized SEO agency can help you build a balanced acquisition strategy and avoid dangerous dependencies on unpredictable channels like Discover.

• Activate the Discover report in the Search Console to track volumes — but never draw long-term strategic conclusions from it.
• Do not alter your editorial line to “please Discover” — focus on the classic SERP and stable search intents.
• If your Discover traffic drops sharply, change nothing — wait 2-3 weeks to see if it’s a temporary fluctuation.
• Build traffic forecasts completely excluding Discover — consider it a random bonus.
• Diversify your acquisition channels to reduce overall dependence on Google SEO, whether traditional or Discover.
• Regularly analyze the share of Discover traffic in your overall mix — if it exceeds 15-20%, you are in a risk zone.