Who 'likes' my Virtual Bagels?
bbc.com
bbc.com
1) The author made a Facebook page for a fictional product that was rather obviously silly - downloading bagels - and talks about it as if it were a real product.
2) The FB ads campaign ran with an absolutely horrid targeting set:
> I chose the United States, the UK, Russia, India, Egypt, Indonesia, Malaysia and the Philippines. I narrowed it down slightly by targeting under 45-year-olds interested in cookery and consumer electronics, but was told that would still give me a potential audience of 112 million customers.
This is completely silly. Trying to target 112 million people means you won't be able to draw any sort of conclusion or get valid data - nor expect a meaningful response - without spending tens of thousands. If he wanted a response in the US or UK, he should have started there and focused on it to the exclusion of all else.
3) His statement "it seems that Facebook adverts can be very effective in generating interest in your business from certain countries but not in the US or the UK" is borne through no real testing or time. He spent less than $100 - you cannot draw any sort of conclusion from that information.
If his sample set, consisting solely of Ahmed Ronaldo, is representative of the rest of the people 'liking' his page, I'd say he has a point - and there's something fishy going on that warrants further investigation.
OTOH, if Ahmed is one of Facebook's 1.5% 'undesirable' users, then I'd say I agree with you, and that India and SA is just full of people who love virtual bagels.
Or, users in the U.S. might just be so jaded from having seen too many joke products advertised on Facebook that they don't bother liking them anymore, while they may still be a novelty in Egypt.
At one point it was talking about the Reticulum (Internet), botnet ecologies and how some of these would sublty modify information such that you couldn't really trust anything you read. I find this an intriguing and somewhat alarming idea.
Facebook was good early on because bots weren't sophisticated enough to create real-looking Facebook profiles so if you had a Facebook profile you had a person. Now they're better but still a human can pick a bot pretty easily.
But bots will only get better.
One idea that's been raised is the notion of trust. Usage patterns are analyzed to identify bot or bot-like behaviour. You see this in identifying sock or meat puppets on reddit, HN and other social news sites.
People have talked about the hordes of Twitter accounts used to make content appear more relevant than it is. This sort of thing is already happening.
But how do you pick out the bots when they're the majority and thus there is little to no statistical norm to compare them to real people?
At what point will bots start modifying Wikipedia articles and have other bots or, more likely, some stooges, approve those changes? What about establishing fake sites with wrong information and having them appear valid to search engines?
This will I believe be a big problem.
One particularly problematic aspect to this is that the site owners themselves, within reason, have little incentive to expose bots (or inactive accounts for that matter) because so much importance is placed on metrics like "# of active accounts".
This only becomes problematic if the user experience suffers beyond a certain threshold.
It's going to be interesting to see just how much of a problem this becomes and what we do to solve or at least mitigate it.
[1]: http://www.amazon.com/Anathem-Neal-Stephenson/dp/0061474096
Also known as asshole metrics (which are, are at most, 40% true). They only exist to manipulate you into thinking the company is more important than reality. Most companies will tell you "total registered users" but few will give "active users." If they do give you active users, then they'll count "active" as something silly like "we saw their cookie during the month" meaning they were just delivered some external asset passively.
Stop lying and make something valuable (except lies can keep companies in business much longer than most people realize).
After targeting US and Canada only, CPAs skyrocketed but each signup actually became worth something.
I'm confused. Who benefits from these click / like bots that seem to plague every advertising platform, besides the platform itself?
Another thought: An advertiser could use a bot to click on ads for keywords they are targeting. If they could exhaust their competitors daily budget they could potentially get higher placement for less CPC [later in the day]. I'm thinking of Adwords here.
The guys you pay, most obviously.
The worst part is I submitted some of this data to Google and they basically said, don't worry we've already detected all the click fraud and filtered it out so you never were charged for it. Uh, ok. No way to prove or disprove that. But if you identified these bozos as fraudsters, why do they still have AdWords accounts weeks later?
Seriously? Seriously?
I mean... no, never mind. "Confused" doesn't even begin to describe it. Given that the US doesn't have an official language, India is more of an English speaking country than the US is.
Relax, no need for that. Obviously India is an English-speaking country, but when we targeted "English" speakers we expected a higher percentage of our traffic to come from Western sources like the US and Canada, instead of 95%+ from click farms in India. We found that wasn't the case, so we switched to country-specific targeting. Thanks for educating me, though.
(This is especially insulting if the site in question represents a company/brand that actually operates internationally.)
Also, English speakers are not limited to countries with English as an (un)official language. Just an example, LinkedIn had 3 million Dutch users before they introduced a Dutch language interface. I don't know what business you're in, but that's a lot of creditcards...
It would look a bit weird if thousands of users always liked exactly the same thing and nothing else.
... but then you'd assume that'd be something Facebook ran automated checks for.
1. Facebook - obviously they benefit from higher CPMs in any country. I seriously doubt they would create fake profiles on their own system though (that would be fraud).
2. Like resellers/advertisers - A client comes to you and says they want tons of likes. While there are a lot of ways to work around this system, one obvious way is to generate fake profiles in countries where the CPMs are low, then target ads at your own bot army. You can sell the likes, generate the likes yourself, and profit. Otherwise you're waiting for quality people to like your page which as this article illustrates, is extremely costly. The bottom line: pay for cheap likes, receive cheap likes.
I should add that these malevolent actors in the system are beneficial in the short-term for Facebook (CPMs rise) but bad in the long-term. As genuine advertisers see their effective CPMs rise, the ROI decreases. This forces benevolent actors out...
Bottom line: We were happy to generate 5000+ likes with little money, and that helped us look like more legit and popular business. However, real value of generating business from these ads was minimal.
Many people believe social signals help increase their search engine rankings and/or conversion rates. Some of those people buy likes, shares, etc. for that reason.
> "What is the incentive to fake-like advertisers on facebook?"
Perhaps people who offer those social-seo services control thousands of robot-facebook accounts and they want their robots to seem like real, active users? It seems possible that they would automate some random activity (in this case liking virtual bagel's ad) while they're waiting around for orders.
Just guesses.
Same thing happens here on HN - there's lots of young accounts that dutifully push a couple generic links from the usual tech sites every day.
There's nothing to disagree about. The guy just pointed out what happened.
Besides, "I completely disagree" is just code for "You're wrong".
I completely disagree with that statement :-D
"I disagree" is a statement of opinion -- I believe something different than you do. "You're wrong" is an assertion of fact -- your belief is incorrect. The former is much less of a direct challenge to the other person than the latter is.
Perhaps they have bot farms that try to drive up the likes of ANY ad targeted to those countries? I'm not sure, but the summary of the posts over the last few days seems to be that ads perform legitimately in the US and EU, and with some questions elsewhere.
Then he acts surprised!? What exactly did he expect?
As others also point: his page is a joke, somebody will always "like" the joke, maybe even recommend it to other friends!
From his target population, one of 70000 potential "eyes" "liked" the joke in 24 hours. So what?
I'll try to explain the author's point. He was confused by and suspicious of the vastly higher liking rates in certain countries than in US/UK.
So: 1600 likes from India and Egypt in one day: $10 == 0.6 cent per like. 350 likes without that area: $16 per day == 4.6 cent per like.
Where's the surprise? If he pays for cheaper likes, he gets the cheaper likes.
Did he expect to pay for cheap likes and get the expensive ones? Does such an expectation have any economic sense?
But his complaint isn't about the price per like. It's about the fact that likes per impression are ten times higher in certain countries. Can you explain that disparity?
We can theorize all day "why" but we don't have any information about that. What he presents is only how much he paid and how much he received, and there everything seems consistent.
People are looking at the market doing something strange and your contribution is "the market functions, who cares".
Sure, I'm not very surprised that if the likes seem highly fraudulent the price per like is lower, but that's not the issue. The issue is these fraud-seeming levels in the first place!
1. likes vary in price per area
2. click rate varies per area
Of course he's not claiming anything doesn't 'fit' in his data. That would require screwing up basic arithmetic about clicks. The author didn't claim to be tricked by facebook. He was surprised by the order of magnitude differences. Why don't you see that surprise as valid? And of course the article doesn't give the data to explain why, because the author doesn't know.
I was thinking the same thing. Maybe it's not so much that people in India click a like button ten times as much, but that more people in India get to see the ads. I assume Facebook somehow reserves ad space for you, but since many more people advertise in the UK, they can only show the ad to the target audience every 1 out of 100 page views, while it might be 10 out of 100 in India. If you advertise in India and the UK at the same time, Facebook will probably reserve a lot more slots in India, because that's cheaper for them. If they want to maximize their revenue, it would be smart to sell the sought-after UK slots to companies that can pay a lot. Indian likes are cheaper because more people want to advertise in the UK but there are fewer slots, so UK likes become more expensive.
.55 is an Adops executive daydream.
(The .059% from UK visitors is about what you'd expect)
I wonder: why Cairo?
I am sure the "pay for a less spammy ad riddled FB experience! Only $1.99 per month!" is in the near future. And I'd bet a fair number of users would pay for that.
Is it really just the fb ads or the entire system of how ads work in 2012? Is it time that someone changes how ads work as per pg's list of ideas ?
Sponsorship has seen a resurgence in recent years. I know that P&G really likes the sponsorship model, but there are only so many things out there to sponsor.
Why do you, and "any marketer," think we're headed away from conversion-oriented, highly quantitative models of advertising?
Now, some people think they can compensate for that. But, for me? I don't understand how you could tell that a user learned about me from word of mouth (before they clicked on an adword that got them to my site and a signup.) - To me, figuring that out is so hard that I might as well just focus on the brand-building kind of advertising.
Now, a lot of this is because most people are willing to pay a premium for the 'quantitative' advertising. they are willing to pay extra for things they think they can track. This means that if I'm right (and who knows) then the brand-building advertising, that most people consider largely untrackable, is underpriced.
I mean, I'm not saying I'm right and you are wrong... just that there are a number of businesses that feel this way.
Now this gets heady: "People are willing to pay a premium for things they can track." Perhaps advertising vectors which are easily tracked, thus optimized, yield more valuable per dollar spent. If this is the case, demand for this inventory (Adwords, etc) increases. You would expect it to become more expensive by forces of the market. This is why you don't see people lined up to fill billboard inventory, especially in the last 10 years.
It is a chicken-egg problem in some ways, but I think people are willing to pay more because their advertising/conversion math enables it, and that math is easier to back up with facts.
Yeah. This makes /absolute/ sense if you are actually tracking what you want to track. (e.g. I think that when I buy adwords, most of the conversions are from people that had a good chance of buying from me from the organic results; I believe that people only buy from me if they've heard of me before; if my brand reputation, in their mind, is good enough.)
My theory is that most trackable advertising that exists now does not really track as well as you think it does.
If I'm right, then trackable advertising is overvalued (because the actions that are tracked are only peripherally related to the buying decision.) - of course, right now most people disagree with me.
I disagree. Online display advertising, which can be tracked effectively has had incredible downward pressure on prices as tracking has gotten better.
I think most advertising is highly overpriced (brand advertising especially) and as tracking gets better prices will drop.
Ad supported business models get harder and harder in that world.
that's essentially what I was trying to say. The part where we disagree is that I think tracking is not as good as you (and most people) think it is.
My point was once it can be tracked well it will come to light that it's not very valuable. Because that's the pattern we've seen with online display.
But, as I reread your comment I realize that I'm thinking TV style mass broadcast advertising here, and if you mean branding advertising on online platforms then I agree with you that in that context it's under priced.
I'm mostly talking about offline brand building advertising. "Nobody is lining up to fill billboards" - I just bought some advertising on the checkout lane for the Safeway nearest google. (If that works, I think it will mostly work because it's so unusual. That, and the density of potential customers in the area.)
Really, I think sponsorship is an extremely effective way to build your brand; way more effective than billboards. I mean, if you can sponsor something that the people you sell to care deeply about. But, that's nearly impossible to track. My theory is that a customer might click on an adword (or organic search result) and sign up because they saw me at SCALE or something... but there is really no way for me to prove that.
I guess you didn't read this: http://irvinebroque.tumblr.com/post/28415393877/how-i-made-1...
Shouldn't he have said useLESSness not usefulness? It's pitifully useless, let's not be politically correct and say it's not that useful.
In the context of that sentence, they are interchangeable.
aren't US & UK markets more lucrative too to bots makers?
I suspect someone somewhere has CTR statistics for certain keywords on certain regions and kind of map it out. Bot makers may have ILLEGAL access and probably do it to avoid suspicion.