In This Issue:

🔍 Google loophole exploited by drug dealers to hijack websites

💻 Microsoft seems to be testing hiding ad unit labels outright

🛑 Reddit restricts users from opting out of targeted ads

📱 Artifact introduces user posts, competing with X and Threads

🚫 X considers ad-free premium subscription option

🗳️ Elon Musk and X CEO clash over election integrity team's status

🌍 OpenAI brings live web browsing back

 📅 Google Ads introduces user accounts with expiration dates

How Drug Dealers Use YOUR Brand's Web Site to Move Product

A recent change to Google’s algorithm is what some people believe is behind a new wave of web pages advertising the sale of crystal meth and other illegal substances.

To back up a little: When you do a search on a regular web site, many sites will return that result as its own web page — dynamically generated based on your search. It’ll have your search terms in the URL and it’ll also put that search term as the page header. Sometimes with the phrase “Results for” ahead of it.

This means you can make any web site that works like this say pretty much anything you want. Just type in a search, copy the resulting URL, and presto.

And when I say anything you want, I mean it. I tested it on our site, searching for the phrase: “High Quality Weed Available for Sale — contact Tod directly!! (but not actually)” and it generated a web page with that as the title and in the URL:

The search itself came up empty — and said so, with a “No Results Found” note. But the headline, advertising drugs, was at the top.

You can make many web sites do this — almost all WordPress sites generate search results pages like this, and that’s 40% of the web right there.

Among those sites where this trick works: The American food and drug regulator, the United Nations, Interpol, the government of Ontario, and more.

How are these getting into Google?

Turns out, a recent change to the search algorithm now lets these auto-generated pages get indexed.

Insider.com reported on this today:

And that’s why hundreds of web pages advertising illegal drugs, along with contact information for the seller, is coming up in Google searches.

Not just any web sites, but web sites that Google trusts and ranks near the top of search results pages: Government sites, educational, and media organizations.

Mushrooms for sale, on the CDC web site

Cocaine and fentanyl, on the National Institutes of Health site

Crack, on a Cleveland clinic’s site

Again, to be clear, these web pages are not static pages hosted by these web sites. They’re generated on the fly, based on the words the user searched for and embedded the URL.

Indeed when we searched for “buy heroin”, the third result down was from the official web site of the Canadian province of Ontario.

It provided a web site, email address, and Telegram account of where, presumably, you could find someone to sell you it.

Google shifts the blame to web administrators

For its part, Google told Insider.com that it was up to the web site administrators to keep those pages from appearing in Google search — using the usual methods: noindex, robots.txt, and the likes.

It also said its search index is 99% spam-free.

That may be true, but 1% of Google’s index is still a lot of pages. Insider did a Google search for the Telegram handle of one illicit drug advertiser in Google and found more than 7,000 results across 24 web sites — some sites being hit thousands of times.

So what can you do about it?

First, check to see if Google has indexed this kind of dynamic content on your brand’s web site. Type words like heroin and meth, followed by site:yourdomain.com

Second, whether you find some or not, it’s probably best to try to send Google signals that you don’t want those dynamically generated search results pages from showing up in Google’s index.

Unfortunately, most CMS systems out of the box, don’t provide that level of granularity, so you might need to install a plugin like Yoast to control the indexing on specific types of dynamically generated web pages.

But if Insider’s report is correct and Google’s crawler is just ignoring those signals anyway, it may be harder to keep those off your site.

Microsoft Keeps Playing “Spot the Ad Disclosure”

Last week, we reported on Microsoft’s newest ad format — an AI-generated product comparison table — and how the ad disclosure was nearly impossible to see at first glance: buried in the top-right, in light grey text, the word “Ad.”

This tiny text likely isn’t something trade regulators around the world will like, most of which say ad disclosures must be prominent.

But hey, at least it was a disclosure.

Now, Microsoft is being accused of running ads with no disclosure at all on the ad units in search results that what they’re showing has been paid for.

SEO consultant Frank Sandtmann found some examples of this in the wild on Microsoft’s Bing search engine and posted a screenshot on Mastodon:

There are three ads shown at the top, and none of them had any indicator within the ad unit that they were ads. The only hint was small grey text at the top of the whole page saying “ads related to” and the search term.

But the individual results were completely indistinguishable from organic results.

Is this the future of online advertising?

For better or worse, shrinking disclosures do reflect an industry trend.

Bing itself, months ago, was found testing ad units where the word “Ad” appeared only in tiny print at the end of the ad unit description. If you didn’t read that long description, you wouldn’t see the disclosure. I had to look at it for 30 seconds before I could find it:

It wasn’t even three weeks ago when we reported on Twitter’s shrinking disclosures — first, changing it from the word Sponsored under a link to a tiny “Ad” snuck into the top-right corner where nobody’s eye goes:

Then, some ads ran on X with literally no disclosure at all. It wasn’t clear if it was a glitch, a test, or a new policy, and X did not respond to media questions about it.

The double-edged sword

From a marketing perspective, is non-disclosure actually good for brands?

We’ve known for decades that the less an ad looks like an ad, the more people are willing to trust it.

Back in the 60s, David Oglivy broke the marketing mold when he started running full page magazine and newspaper ads, loaded with text — the exact opposite of best practices at the time — and it worked.

These days, the platforms may be different, but the advice is identical. TikTok’s tagline for its advertising department: “Don’t make an ad. Make a TikTok.” Indeed, the best performing TikTok ads look like organic posts.

So then why should we worry about a search engine or social platform doing the same? Because on TikTok today, and in those magazine articles years ago, there was still disclosure. Sure, they look organic, but they aren’t, and they say so.

Straight-up removing ad disclosures entirely — besides picking a fight with every national trade regulator — may work in the short-term for sales, but you’re trading quick wins for long-term brand damage.

All we have as marketers is trust.

If that’s a currency we’re willing to spend, then we get what we buy.

Reddit Bolsters Its Ad Targeting by Removing Users’ Ability to Opt Out

Speaking of spending the currency of trust, Reddit has decided to cash some brand goodwill in for the sake of better ad targeting — the company this week telling its users it was “simplifying” some of their ad privacy options.

And by “simplify,” Reddit means “remove.”

Previously, users have been able to tell the site’s advertising platform to not consider what communities they join, which posts they engage with, and how they spend their time on the site when creating a user profile for marketing.

In other words, spend all your time talking about gardening, and you could tell Reddit to not use that data for ad targeting.

All snark aside, we have to acknowledge that as marketers, having the ability to target by interest is valuable. As the company continues its unpopular but inevitable move toward more business value, this move was an easy one to see coming.

Improved targeting won’t include third-party transfer

So advertisers will, in theory, get better performing campaigns. What they won’t get is data — Reddit said it wouldn’t send that activity information directly to advertisers.

Then again, Twitter once said it would never do a whole lot of things. Then, some guy bought it outright, and started doing a whole lot of those things.

The move, as you can expect, isn’t popular with Reddit users who have had a number of things to get mad at site management about this year.

Also, Reddit is not immune to legislation, and did acknowledge that they would still let residents of some countries keep that ability to prevent their activity from being used for ad targeting. They wouldn’t say which countries, specifically, though these will almost certainly be European nations.

A strange way to handle restricted categories

They also say they’re adding the ability to opt out of specific ad categories, like alcohol, gambling, or weight loss…

…though peculiarly, Reddit says this ability won’t be absolute — some ads from restricted categories might seep in as they work on their classification system.

This is odd because ad categories are pretty binary. You’re either in them or you’re not.

Every ad platform has you indicate what business category you’re in. You’re a news organization. Or you sell children’s clothing. Or you sell guns. Some platforms require you to additionally specify if your campaign is about protected topics like housing and employment.

Reddit could simply exclude businesses in the gambling category, or the alcohol category. For them to ignore that simple and clear classification, and instead try to build a machine learning algorithm to detect topics, seems like a peculiar first step to take.

Still, the writing really had been on the wall for these kinds of changes to their ads offering ever since they signalled their plans to go public soon.

Artifact: The News App Goes Social

A relatively new app built by Instagram’s original co-founders is starting to look a lot like a social platform.

The app is called Artifact and it started as a news app, similar to the Apple News or Google News apps. I’ve used Artifact, and it’s pretty solid. Like the others, it tries to show you stories it think you’d be interested in, based on your past engagement of other stories.

But its algorithm, in my experience, is much better and stronger than the others. Sort of like how TikTok really nailed its discovery algorithm, this app is doing the same for news articles.

It has been slowly building itself into a social platform. You can comment on stories, and a few weeks ago, added the ability to submit a link — like to a blog post you wrote.

Now, it’s letting users publish the full text of a blog post or company announcement or how-to guide directly to the feed.

If you have the app and want to try it out, go to the Links tab, tap the + icon, and from there, you can add your own images, title and text.

This is another contender’s shot at filling the void created by the drop in Twitter usage. Substack added a similar social feature some months ago.

So should marketers try it out?

There’s no reason not to, if you have the time, and in fact it may prove valuable. Often when these types of new social features and apps are released, there are early adopters there to see your content, and few other marketing teams there posting.

That said, there don’t seem to be any brand safety controls which would prevent your blog post or company message from being positioned beside a news story about a tragedy or something similar.

Also, this is mobile only, there is no way to post your content via desktop, and no API to use any third-party tools.

X’s “No Ads” Plan

X, formerly Twitter, may be considering a paid tier which would eliminate ads from subscribers feeds.

It’s not a crazy idea. Meta is said to be considering the same thing for users in Europe.

This is something X’s majority owner, Elon Musk, has mused about in the past. But now, some software engineers are seeing text in the codebase indicating that this is actively being worked on.

Currently, there is only one tier of membership, and it promises fewer ads in a couple of places in the app, but not most of it.

X’s ad revenues amount to about $US12 per user, per month — so any ad-free version would need to charge at least that to break even.

It would also make the platform, I would think, less attractive to advertisers, given that their campaigns would be guaranteed to reach fewer eyeballs than the current setup.

Currently, about half of 1% of X’s user base subscribe to its Premium plan.

Since the start of his tenure at the company, Musk had tended to make sudden announcements, only to ignore or backpeddle later. Last week, he said they might soon charge “​​a small monthly payment for use of the X system.”

He also said a cheaper version of the current Premium plan is coming soon, though it’s not clear what would be stripped out.

Did X kill its integrity team? That depends on who you ask.

Meanwhile, yesterday, X’s CEO Linda Yaccarino disputed media reports that their election integrity team had been slashed from 25 people to just four people. She told a conference that the group is a “robust and growing team.”

Except her boss, Elon Musk, earlier that same day not only confirmed that the team had been cut, he said the whole unit got disbanded.

Replying to a post about the cuts, Musk posted:

Oh you mean the “Election Integrity” Team that was undermining election integrity? Yeah, they’re gone.

— Elon Musk (@elonmusk)
Sep 27, 2023

More than 50 countries have national elections next year.

On Anti-Semetism: “Everybody deserves to speak their opinion”: X’s CEO

At the same conference, Yaccarino was asked about concerns that the amount of hate speech, particularly anti-Semetic speech, had been increasing on the platform.

Her response probably wasn’t what brand safety managers and advertisers wanted to hear. You know, something like “We denounce hate speech.”

Instead, Yaccarino told the somewhat shocked audience “Everybody deserves to speak their opinion.”

Then she looked at her watch, and said she had to catch a flight.

Two Small Items

ChatGPT can see the web again

First, a welcome change to ChatGPT — actually, bringing back a change that they had put on hold for a bit… you can now once again use results from the live web, so no more of that “My knowledge stopped in September 2021” message.

OpenAI paused that ability when they discovered people were using the feature to bypass news paywalls. Since then, they’ve taught the engine to respect robots.txt exclusions, and given enough time for publishers and web sites to put that in place, if they so choose.

Second, Google Ads has added a nice feature: You can now set up temporary users to be able to see your account — and you can automatically expire their access after a set period of time.

Some links in this newsletter may provide affiliate revenue to us.

This newsletter is powered by Beehiiv.