Marketing The Internet

Twitter Interests and the scam of interest-based targeting

a screenshot of the Twitter Interests page with some interests deselected
I don’t have any special interest in space telescopes, but I think astrophotography is pretty and like science generally. I like Star Trek TNG, though I’m not a Trekkie. But tanning? I am anti-interested in tanning. And sports — absolutely zero interest. I would prefer to avoid all sports news except to hear if it’s going to make traffic in Seattle a nightmare.

Social media silos are harvesting our personal data, selling it and using it to target us with ads and filter the information we see through their algorithm — and they’re crap at it. We sacrificed our privacy for this? 🤔

Targeting problem: a significant portion of supposed interests — which I presume are used to support algorithmic decisions and targeted advertising — are incorrect.

I recently looked through my Twitter Interests and removed everything that I was actively not interested in. About half of my interests were totally wrong. It appears it hasn’t improved much since 2019.

As the old saying goes, half of what you spend your money on in marketing is a waste, but you don’t know which half. Portraying targeting based on personal data extrapolated from users’ online actions as inherently more accurate or higher quality than other forms of advertising feels like a scam.

I can extrapolate where some of these mysterious false interests came from because they’re recent enough to remember:

Data problem: they cannot distinguish an interest in current events from actual interest in a topic.

The Try Guys, who I’d never heard of until one of them apparently cheated on his wife and got fired, were a trending topic recently. With enough people posting subtweets I was curious enough to open one of the threads. But I’m not ever going to watch a Try Guys video.

Likewise, it thinks I’m interested in UK politics and current events. I’m interested to the extent that they affect world politics and I can see comparative trends with the US. (For example, right now I’m interested in how English people feel about the monarchy, and seeing everyone shit on the colonialist institution 😎 I’ll take anything that makes me feel good about America these days; not having a monarchy and (theoretically) having freedom of speech are two things we got going for us.) I also tend to read news articles on The Guardian rather than paywalled American news sites. Outside of that lens, I don’t have any special interest in UK political events 🤷‍♀️

Data problem: their software can discern the nominal subject of a post, but not the actual subject when something is being used as an example or case study.

Twitter thinks I’m interested in Waffle House. I live on the West Coast, where we don’t have Waffle House. I am highly unlikely to ever patronize a Waffle House. This is not a useful interest to evaluate in an algorithm, or for advertisers to target me with content.

I suspect they believe I’m interested in the brand because an author I follow recently posted about Waffle House as an example of using retail / capitalism as a lens of emergency severity, and as an institution you can rely on to get hot food when you need it. A human would understand that I’m interested in commentary about resilience and capitalism, not Waffle House.

This is just as bad at the old Amazon classic of showing you vacuum recommendations after buying a vacuum. When we’re given the option to target a level of granularity that *feels* like we’re selecting exactly the people who will be interested in our message, interest-based targeted marketing feels like it’s a better deal for marketers. But is it really?

Frankly, I think that at its current level of quality, interest-based* data targeting is a scam on shareholders, marketers, and small business owners. By giving marketers some number (as opposed to the total black hole of advertising in physical media), we feel like we are getting a return on our investment — but the number of people who saw your post, or even who clicked on it, is not very meaningful. Results in the area you are targeting are what is truly meaningful, whether that is sales, prospects, calls, or some intangible result like “brand awareness.”

* I would guess that demographic targeting, or data based on actual purchases, is higher quality.

Sure, some targeted advertising “works” — for a certain type of product. Instagram’s targeted ads seem to be decent at direct sales — for made-for-the-internet cheapo brands. But there could be a better way to earn sustainable profit from online advertising, if they prioritized the long-term success that would come from building relationships with small businesses and organizations over the short-term returns that extract money from the local level with their pay-to-play system:

Social media is a fully exploitative system. The silos steal data and privacy and attention from their users. They then extort small brands and creators to spend their limited time creating the content that keeps viewers returning to the site, and then charging those same brands to show their content to the people who asked to follow them, or to users who supposedly match their targeted interest areas. Culturally, they flatten artisans, artists, writers, and thinkers to monetizable units as producers of content.

(And we users buy into the ad-based system because we’ve fallen for the trap of free.)

If I was spending ad money online today, I’d probably put it into niche publications with a specific audience, like email newsletters and podcasts. Of course, finding those is an investment of time — but it seems a more trustworthy way to reach your audience than the questionable quality, unethically-harvested extrapolated data through silos.

So much of the internet’s supposed market value is betting on this personal data; if social media companies can convince small businesses that targeted ads are the only way to reach their audiences, and large brands that “having a presence” (augmented by sponsored posts) online is essential to their success, social media companies capitalize on that data, regardless of its real quality.

By Tracy Durnell

Writer and designer in the Seattle area. Freelance sustainability consultant. Reach me at She/her.

2 replies on “Twitter Interests and the scam of interest-based targeting”

Leave a Reply

Your email address will not be published. Required fields are marked *