As a positive example, I like how many YouTube channels take 30 seconds out of a 10 minute video to talk about some company's product that might be of interest to their audience.
Not an annoying video ad that has nothing to do with what I'm reading / watching, just the host talking about a product or service that their audience has a chance to be interested in.
No tracking of my habits was required, and all parties involved probably got detailed-enough feedback as to their campaign's effectiveness.
When I see someone I trust on YouTube recommend a product in a pre-roll ad or the like, then I don't really skip it and I do really keep those companies in mind in case I ever am looking for a product that they might offer.
> trust
This is the key feature. The advertiser risks trusting that the promoter will show their product in a positive light. The promoter risks their own reputation on the quality of their endorsements.
This element of risk hopefully serves as an incentive to maintain a high quality of both the product and the endorsement. Bad manufacturers and poor quality (or scandalous) services should lose access to trustworthy endorsers with (hopefully) large audiences. Good manufacturers and high quality services won't want to associate with cheap shills.
The only problem I see is that places like YouTube are not optimizing for quality endorsers. Their focus on engagement is driving away people that try to create quality; may have already left. Those that remain are becoming increasingly formulaic, uninspired, risk-adverse, and "advertiser safe".
[1] Which is required by law![2] This probably also means disclosing if you got a free copy for review purposes[3] of a product that would otherwise cost money or have a non-monetary value. The disclosure must also be "clear and conspicuous"[4].
[2] https://www.ftc.gov/news-events/press-releases/2017/04/ftc-s...
[3] https://www.ftc.gov/tips-advice/business-center/guidance/ftc...
[4] https://www.ftc.gov/tips-advice/business-center/guidance/ftc...
I like how many podcasts have ads with predictable lengths, so I can skip them with the "fast forward" button without hearing them. NPR is 4 skips; Slate is 8-10. It seems like "deep learning" should have figured this out by now.
> make it indistinguishable from content
I kind of hate the idea of having marketing-planted material appearing indistinguishable from real content. Most of that is the general distrust I have towards large organizations in general.
That said, one could argue that all content is "marketing-planted" even when it's not an ad. There's always an agenda or at least bias. That's a fair point. Carry on good sir.
They're all creepy and they're all unwelcome.
I work with a publisher on these sorts of things (I am not an ad network, though). Some of those networks have strict prohibitions against showing ads on pages featuring certain types of content, and so we've had to build our own system to display ads on those pages which work with networks which are fine with that content. The publisher does sell its own ad spaces to interested local parties, but these are usually accompanied by periodical ad sales because the company ultimately makes more money selling ads through networks.
This sounds like an argument for sponsored content. At least with the current ad strategy, it can be relatively easy for a user distinguish ads from content. How do we distinguish a YouTube video that ends with the statement "this video was brought to you by [sponsor not relevant to the content]" and a YouTube version of an infomercial?
Tie the advertisement to the type of content that the user is already there to read about. Then you can serve ads from the content web server and make it indistinguishable from content, completely unobtrusive, completely noninvasive, and maybe even welcome.