A few days ago I came across a talk by Maciej Ceglowski titled Notes from an Emergency. If you have not watched the talk, stop reading this and go watch it. In the talk, he makes a suggestion that the Government should regulate ads to target content, not users. There was a discussion about the talk on Hacker News where Maciej, in response to a comment that implied ads that target content will be worse than ads that target users, said that’s why he feels it’s a regulatory argument and not a business argument. That got me thinking, are ads that target content worse than ads that target users, looking at it purely from a business perspective? I don’t think so, and in fact feel that ads targetting content will be better from a business perspective. In this post, I propose a model for a privacy oriented ad network that targets content, not users. This is not a novel proposal and it’s likely that people have thought about this over the years.
I will briefly explain why I feel the current model does not have the best interests of the users and maybe even businesses in certain cases. Ad networks thrive by profiling users and letting businesses target users based on age, location, interests, browsing history, shopping history and so on. Looking at how much money the Google and Facebook ad networks bring in, it seems to work very well. What are the potential advantages of targeting content instead of users?
Relevant ads : Copying my comment from HN : At it’s core this is essentially like sponsorship. Anecdote : I’ve always found sponsored content more useful than the ads that ad networks show me and they are also not creepy like traditional ads. Eg- I search for strawberries on google, click through a few links. Separately I visit a tech blog to learn about how to monitor a VPS, guess what shows up on a site about servers and software? Strawberries! It’s very creepy and completely out of place.
Improved CTR and engagement : Without enough data points to backup these claims, it’s somewhat anecdotal, but more relevant ads increase CTR and engagement. People who have or had sponsorships on their websites report that sponsored content is more relevant to their audience and the advertisers report CTRs/signups from sponsorship to be much bigher than traditional ads. Eg- Troy Hunt’s blog, IndieHackers.com. Obviously, running ads that target content is not exactly sponsorship, but it’s closer to sponsorship than traditional ads.
Privacy : Since the proposed ad network focuses on displaying ads based on the content of the page, there is no need to profile users. On a related note, there exist large companies whose only job is to profile users and sell this information to advertisers or ad networks. And these companies use downright illegal means to profile users. The need for such cancerous companies will not exist anymore if ads instead target content. Then the focus will shift to matching ads with content, so companies will need to focus on the ML/DL of “profiling content”.
Inevitable future : With time, people’s tolerance for creepy ads that profile users continues to go down even among people who are not tech savvy. At the same time, companies still want to advertise and get their product in front of as many eyeballs as possible (and they are happy to spend obscene amounts of money to do so). Is there a middle ground somewhere? Yes. Stop profiling users and still display relevant ads by targeting content. Allow (savvy) users to completely disable ads for a monthly subscription. The reality is that the internet literally runs on ad money allowing the average internet user to use it without ever paying for any content or services on the internet. Yes, that’s how the average user uses the internet, more so in Asia, Africa and South America. People who use the internet for the first just can’t spend $1 on an app or a service.
Proposed ad network : what are some of the properties or features of the new ad network?
Profile content, not users.
Allow people to opt out of ads for a monthly fee.
Companies that want to advertise pick a budget, choose the websites that are part of the ad network and that’s it. The websites that have the adsense like code will display the ads. Eg- A hypothetical email marketing company will choose to advertise on blogs about email marketing. The ad is relevant to the users and there is no user tracking involved.
Companies can also advertise on individual pages and if there is interest from multiple advertisers, a bidding process will decide who gets the spot.
Websites that want to display ads will just embed a snippet of code to display ads, just like adsense.
The only tracking will be CPM and CTR to begin with. Users will not be profiled.
Since it’s a marketplace like site, for payments I will use Paypal for charging advertisers and also paying out people who make money displaying ads. Minimum ad budget will be $100 and the minimum amount for a payout will also be $100. Why Paypal over Stripe Connect? I will need to register using Stripe Atlas which costs $500 + $300 post incorporation setup charges. I will definitely move to Stripe if this takes off though.
Code will be open source and licensed under the GPL. All development related activities will be discussed transparently. The site will be built using Django.
The commission rates are not finalised, but 30% seems reasonable. For reference, depending on the type of content, Google takes between 32% and 50%.
People mentioned that ads that target content instead of users won’t work as well as the status quo. My theory is that it will work way better, even if you consider it purely from a business PoV. Of course, the only way to find out is to collect data about ads that target content and compare it to the status quo.
If you’re interested in seeing such an ad network being built, fill out your email in this Google form and add any comments about what you want to see in the ad network. I also embedded the form below if you prefer filling out the form without leaving this page. If there is sufficient interest in seeing such an ad network come to life, I will move ahead with the project.
Let there be no confusion about this project : if I decide to go ahead with this project, it will be a for profit venture. This post is a way to guage interest in such an ad network.
Any feedback is welcome, so please fill out the above Google form and/or reach out to me on twitter or email.