Remove OpenGraph metadata tags from "Coming Soon" sites

[gamebits]

What

If a wpcom site is marked as "Coming Soon", it should output no OpenGraph tags that social media sites can use to generate thumbnails, descriptions, and other previews of the site.

Judging by #56779 (comment), I presume the current behavior is expected, and that a site should instead be set to "Private" to remove this metadata. Hence I'm submitting this as an enhancement and not a bug.

Why

The front page of a site that is marked as "Coming Soon" still has all the metadata loaded necessary for Slack and Twitter to generate previews of the full site. Here are the OG tags of one such site marked "Coming Soon" with Yoast (wordpress-seo) installed and active:

    <meta property="og:locale" content="en_US" class="yoast-seo-meta-tag"/>
    <meta property="og:type" content="website" class="yoast-seo-meta-tag"/>
    <meta property="og:title" content="Home - by Nijaeri" class="yoast-seo-meta-tag"/>
    <meta property="og:description" content="Who am I? “Zim, Zima, who got de keys to my bimma?” Well, lol, I’m not Beenie Man. My name is Nijaeri (Nigh-jair-ree). It’s an African American variation of the African name Njeri, and so far, it is 100% unique to me!" class="yoast-seo-meta-tag"/>
    <meta property="og:url" content="https://bynijaeri.com/" class="yoast-seo-meta-tag"/>
    <meta property="og:site_name" content="by Nijaeri" class="yoast-seo-meta-tag"/>
    <meta property="article:modified_time" content="2024-06-04T00:16:07+00:00" class="yoast-seo-meta-tag"/>
    <meta property="og:image" content="https://bynijaeri.com/wp-content/uploads/2024/05/about-nj-1.png" class="yoast-seo-meta-tag"/>
    <meta name="twitter:card" content="summary_large_image" class="yoast-seo-meta-tag"/>

Without Yoast:

    		<meta name="description" content="Who am I? “Zim, Zima, who got de keys to my bimma?” Well, lol, I’m not Beenie Man. My name is Nijaeri (Nigh-jair-ree). It’s an African American variation of the African name Njeri, and so far, it is 100% unique to me! Read About https://open.spotify.com/artist/4iHNK0tOyZPYnBU7nGAgpQ?si=A3hldOKJQVGe65vMWJFfsA https://open.spotify.com/track/4EI8VuxUuIHKfafU72emqz?si=71b0e3ae964b4268 R&amp;B Memoirs R&amp;B Memoirs View all Posts Be the&hellip;" />

<!-- Jetpack Open Graph Tags -->
<meta property="og:type" content="website" />
<meta property="og:title" content="by Nijaeri" />
<meta property="og:description" content="Music and life those are my genres" />
<meta property="og:url" content="https://bynijaeri.com/" />
<meta property="og:site_name" content="by Nijaeri" />
<meta property="og:image" content="https://bynijaeri.com/wp-content/uploads/2024/05/about-nj-1.png" />
<meta property="og:image:width" content="378" />
<meta property="og:image:height" content="567" />
<meta property="og:image:alt" content="" />
<meta property="og:locale" content="en_US" />
<meta name="twitter:text:title" content="Home" />
<meta name="twitter:image" content="https://bynijaeri.com/wp-content/uploads/2024/05/about-nj-1.png?w=640" />
<meta name="twitter:card" content="summary_large_image" />

How

The site's settings:

Excerpts of how it appears on Slack and Twitter:

[mrfoxtalbot]

Great catch @gamebits!

I can reproduce it on simple and on AT. Could someone in @Automattic/t-rex or @Automattic/lego take a look at this? Thank you!

[arthur791004]

I'm proposing Automattic/jetpack#39012 to remove the og tags when the privacy of the site is configured to one of below:

• Coming Soon
• Public & Prevent 3rd party sharing

[arthur791004]

@gamebits @mrfoxtalbot Do you think it would be better to remove OG tags if the site is Public but disabled Prevent 3rd party sharing? Do our users expect this behavior when they want to prevent 3rd party sharing?

[gamebits]

@arthur791004 The "Prevent third-party sharing" setting has this definition:

This option will prevent the site's content from being shared with our licensed network of content and research partners, including those that train AI models.

I think it would be poorly received by our customers if we told them: "For social media optimization and to ensure your content looks correct when shared with your followers, you need to opt into AI models using your content."

Let's keep OG tags enabled or disabled independent of that checkbox.

[mrfoxtalbot]

+1 to what @gamebits said. We should only remove opengraph tags based on the site's visibility. Private and Coming soon sites should not "leak" any information about the site.

Thank you @arthur791004!