46 comments

[ 3.1 ms ] story [ 68.1 ms ] thread
We are building the new social ad unit at Referly - and every word is written and targeted by humans, not demographic algorithms
The trouble with that approach is that while algoritms aren't great in this area (at the moment), they are at least trustworthy (in that they'll do what you tell them). People on the otherhand are often untrustworthy, and their definition of targeting is often likely to be quite different to yours and mine. It will be interesting to see how you solve the problem of untrusted users, my feeling is the ideal solution will end up being a combination of humans (users and staff) and algorithms. How to make that scale of Twitter/Facebook will be challenging.
I love hot dogs.
Read this before lunch and hoping that the author was opening an actual restaurant in his postscript.
MySpace's problems are due to their user base defecting en masse, I don't think their ad strategy was really a factor. Could be wrong, though.
This is properly the best example of why hn needs tags. You think you are going to read an interesting article on food and culture and maybe learn a valuable lesson, when in reality you get a piece of whyning crap delivered by a marketeer.
I agree HN needs tags but still think this was an interesting post. While the analogy could have been more thoughtfully constructed, I think the topic is very relevant right now for startup business models (ad-supported versus paid).
Another stupid post about fb and their business without mentioning the fb ad exchange? I'm having trouble understanding how Dalton isn't misinforming his audience.

Read this for why I think fb has a great future in ads.

http://news.ycombinator.com/item?id=4279900

From the Authors website linked to by his P.S.

"App.net is building a better way to share your status."

What an immediate turn off! The first thing I think when I read this 'ugh not another status sharing service'. I think the author needs to rethink how he pitches his idea...

Thanks for your feedback. We have had an internal conversation about copy in the header and are currently using Optimizely to test a few variations. We are definitely open to suggestions :)
What I want out of app.net would be a community building machine -- like most web forums used to be. In 2012, I don't know anyone personally that follows a classic forum anymore; I want that type of connection without the archaic style. Building community on my own website, Facebook, YouTube, Reddit, and/or Twitter can be exhausting, tedious, and time consuming. If app.net let me create a community and interact easily so that I can focus on producing content, I'd love it.

Just my two cents.

Before reading the postscript I thought it was an interesting, if not revolutionary, critique of ad services. Then I read the last line and realized the whole "article" was yet another ad for app.net. I felt like I had been tricked into reading a sales pitch, which irritates me and makes me less likely to join your project.
(comment deleted)
This blogpost was not intended to be another ad for App.net. Dalton's recent posts are a byproduct of his firm belief that the ad-supported business model is unsustainable. In fact, the footer was not originally included in the article. He decided to include a short sentence connecting his critique of ad services to our project after receiving feedback from the community. (http://twitter.com/johnsheehan/status/227816577607823360)
We get it... the ad model is broken, a subscription model is, in theory, always better. It's simpler. Generally, simpler is better. However, I don't think the app.net experiment will succeed.

I'll be brief:

- Outside of phones (& WhatsApp), I can't think of a single paid non-business communication platform.

- App.net will not have a good one-player or even two-player mode (good 1-player: DropBox, Github, Evernote).

- All paid services have at least 10x less users than Twitter.

- If it does succeed, twitter could just copy you.

The biggest problem with Facebook web ads is twofold:

1) They display the same ads on every page (regardless of context)

2) They rely on the advertiser to know their target market. (Most often, the advertiser doesn't even know)

I'll try to write up something on this later.

[edited for formatting]

If we agree that the ad-supported model is broken, why not try to create an alternative?
(comment deleted)
You absolutely should. But it doesn't look like this IS the alternative (at least the viable one).

That said, no one appears to have any other firm ideas about what the viable alternative should look like (only what it shouldn't). And as it's your money and not mine, I say go for it, good luck!

> We get it... the ad model is broken, a subscription model is, in theory, always better... However, I don't think the app.net experiment will succeed.

I agree that if App.net succeeds it will be construed as a data point for this line of reasoning, but I think Dalton's business experiment should be regarded as independent of his technical experiment. Things fail for lots of reasons, not just because free is better than paid.

> - If it does succeed, twitter could just copy you.

By that logic, MySpace could have just copied Facebook.

> - App.net will not have a good one-player or even two-player mode (good 1-player: DropBox, Github, Evernote).

That seems like a given, but I wonder if it's a foregone conclusion that App.net will turn into the same kind of thing as Twitter. After all, technically there isn't a huge difference between HN and Reddit, but the way they're used is very different. Even cultural differences can be dramatic. I haven't invested in App.net, but I do find myself toying with the idea of having something I could use like Twitter, but without being Twitter, and I admit I find the idea attractive. If I could control whether the messages are public or private, I could see using it as a pub/sub framework if the payloads were essentially encoded API messages. That's not something I'd be inclined to try out with Twitter, even though it's free.

I guess I'm saying subtle differences can be meaningful; maybe they can be the difference between success and failure.

1. In the past even five years, we've seen huge changes and experiments in business models (just saying that these have been tried at scale in the past 5 years). Some have been successful, others have not: In-app purchases (Zynga), SaaS (37 Signals), Freemium (DropBox), Daily Deals (Groupon), Early Advantage Buying (Pinboard), User's Choose (HumbleBundle), Razor/Blades (Nexus 7, PSP), $.99 (iTunes, Google Play) etc. I'm not arguing that free is always better than paid, just that in app.net's case, it probably is.

2. The core difference between "MySpace + Facebook" vs. "Twitter + App.Net" is that the reason (at least published) why App.Net is better is because it isn't supported by ads (& that the Twitter team is focused on how to sell ads diminishing the importance of the user). I don't believe that this is a strong enough reason to change my user behavior.

3. Subtle differences can be meaningful. Agreed.

I guess who knows what will happen with app.net but I think the great biz model experiment will only accelerate in general.

I, as a company owner, would like to pay Twitter, as a news distribution service, to make my tweets more visible to my followers. Stick 'em, emphasize 'em, repost 'em - I don't care, but do something that would help me reach these people. They are already targeted. In fact, they have targeted themselves, voluntarily.

So here's the money, Twitter, take it. Help me help my subscribers get the information that they have subscribed for. Thank you.

Yes, exactly that.
I think they'll get around to that. A number of their UI nudges discourage people from taking a 'completist' view of reading their stream -- no mark-as-read, no persistent latest-read-indicator, endless encouragements to follow more than could be read. That plus control of the client would let them offer some kind of paid pinning as a service.
Hi there. Could you drop me a note at brad@nowspots.com? We're building stuff like this.
I think Facebook and Twitter ads are being unreasonably benchmarked against Google adwords ads. Facebook in particular.

Google ads typically convert to a total handoff of the user from Google to the ad buyer. That's the only dynamic Google really admits to. But that's OK, because people routinely engage with Google in order to find products and services. Very low conversion rates are repaid by purchase intent.

People don't engage with Facebook to find products and services. So the ad dynamic Google has doesn't work very well on Facebook. And so, story after story about how nonperformant Facebook ads are compared to Adwords.

But Facebook has a compensating strength: it admits to other dynamics.

So for instance, after "Liking" almost everything (I actually liked) from the "Recommended" block on Facebook, my feed is awash in announcements from brands. Not an entirely welcome development, but not a horrifying "hope you like hot dogs" experience either, since those announcements tend to actually be announcements. So far as I know, every one of the brands that got something on my page that way paid nothing to Facebook to do that.

People do engage with Facebook in order to "Like" things, for the same reason they fill out personality quizzes: benign narcissism.

So Facebook ads can have the objective of generating "Likes", which give advertisers many possible bites at the apple for each user down the road. I know there are brands that have exactly the same process on Twitter. The "like" count isn't all that interesting, but any measured conversion from announcements to "likers" establishes an ROI for Facebook ads.

It's not at all surprising that lots of big companies haven't figured this out; they're still throwing Google ads at Facebook to see if they'll stick; the ads don't stick; they drop the ads; the only ads left are from a few "savvy" brands trying to generate "likes", and from skeezy "hail mary" inventory that generates so much return from a single scammed conversion that it's worth posting ads anywhere.

To make matters worse, social media ad inventory is rapidly moving to phones, where advertising is even more challenging.
Mobile advertising is only in its infancy. If you think ads are invasive now, just imagine what life will be like when they move to a screen 1/10th of the size!
I can accept bad metaphors and analogies in HN comments;  this a technically focused community and the comments are so far above above most online forums that nitpicking about literary devices would be ridiculous. But I can't forgive the hotdog/hamburger/caviar analogy in this full length blog post. It does not convey his message well and really does the article a disservice. This writing should have been more thought out considering it is a follow up to his previous posts on this subject that got quite a bit of attention.

I realize this is somewhat off topic. But as an entrepreneur I feel this is worth noting because I can't overstate how affective a good metaphor (or good, well revised writing in general) can be in persuading others. And as a reader how easily a bad one can turn me off especially when woven throughout your entire piece.

"I can't overstate how affective a good metaphor (or good, well revised writing in general) can be in persuading others."

Muphry's law[1] strikes again.

1. https://en.wikipedia.org/wiki/Muphry%27s_law

Affective - causing emotion or feeling.

I was conscious of my usage although ill be the first to admit i make typos all the time. To elaborate, writing can evoke feelings good or bad, feelings of familiarity or confusion, and these feelings can lend to a more persuasive writing. This was my point in the second paragraph.

What's wrong with the metaphor? How would you have described the problem differently?

I thought it was a really good description of the situation Facebook is in. People have speculated for years about all of the different ways they could produce insane revenue/profits (the hot dog -> caviar machine), but their current revenue (the hot dogs) is very lackluster.

I think the issue is that metaphors work best when they are about real situations. The idea of a metaphor (usually) is to help a person understand one situation by mapping their knowledge of another, more familiar, situation to the first. No one is familiar with the problem of turning hot dogs into caviar, because no one is trying to do it. If anything, the actual problem is more familiar to those on HN than the metaphor.
You were not persuaded. Fair point. Blaming the metaphor is probably going too far. Perhaps you weren't the intended audience?
It's a pretty good analogy. FB can't increase the volume of its users all that much, so it has to drastically increase its ARPU to compensate.

But I (and many others) don't think it can.

A new “social ad unit” will be created at Facebook or Twitter in the next 12 months that will manage to be far more profitable than current ad units, not piss off users, and immediately be embraced by advertisers.

I call them "tweets". Seriously, most tweets are ads, so just charge people to send them. If you have over ~1,000 followers then you are a de facto marketer and you could pay per tweet per active follower.

  > If you have over ~1,000 followers then you are
  > a de facto marketer and you could pay per tweet
  > per active follower.
Then we start seeing "follower bombs" where the goal is to increase someone's number of followers until they are forced to pay for Twitter.
If you have over ~1,000 followers then you are a de facto marketer and you could pay per tweet per active follower.

So, so wrong. I'm not far off that, and while I tweet now and then about work stuff, it's all from a point of view of "this is what I'm doing", not "this is what I want people to see". If I changed job to a completely different industry tomorrow, I'd tweet the same stuff about the new work, regardless of whether it fits my followers.

I like twitter enough that I'd pay for a premium service if there was one and its benefits were slightly interesting, but I wouldn't pay just because I breached some arbitrary line in my account.

Won't that just incentive quantity of tweets rather than quality of tweets? Where is the value in that service?
If you're paying to send tweets, I'd think you'd want to maximize their quality.
It's not just the prospect that FB or Twitter figure out a new wildly-profitable promotional model themselves, but the prospect that someone else figures it out, and FB or Twitter's large audience/social-net/technology defensible asset is still the best way to deploy it.

Compare: Goto.com (Overture) pioneered pay-per-click auctions, and were already a profitable public company even before Google adopted a similar model. But only combined with Google's natural-search advantages did that model become a profit supervolcano.

"The advertising units Twitter and Facebook are selling today are “hot dogs”: poorly targeted, poorly performing."

Maybe that's true for FB, but apparently Twitter's mobile ads are doing just fine: http://on.wsj.com/LxMLwc

Something related on a literal level to the article's metaphor- there's a bar in Midtown Manhattan called Vanderbar that sells a $50 Kobe beef hot dog topped with caviar.
That has to be the best hotdog in all of Manhattan.
If I could use an analogy like Dalton does that doesn't make much sense;

Dalton is trying to sell us a better mousetrap, BUT--atleast in the beginning--he was advertising it as a totally different way to catch the mouse.

The problem is that most people either:

a) aren't bothered by the mouse, or

b) didn't know they should be trying to catch the mouse in the first place.

Substitute 'mouse' and 'mousetrap' with 'Facebook & Twitter's advertising model.'