Facebook Ads: How to Get 1 Cent CPC (and why you might not want to) | Hacker News Reader

Facebook Ads: How to Get 1 Cent CPC (and why you might not want to)

tech.momentgarden.com

82 pointsgarbowza25 comments

this whole article is practically useless without data on how the incoming traffic ultimately converted on momentgarden.com. (how many sales did you generate from picture B over picture D?) just like eve online's a/b testing that led to using pictures of attractive, loosely clothed women indicates you can always increase click through rates by getting closer to porn. what good are CTRs when your traffic is more interested in the picture they clicked than the product you're selling?

It's true that ultimately for our product, conversion into customers is what matters in the end. But I don't think that makes this article useless.

Almost all businesses want to get people "through the door" to your website or product, and our experiment shows how you can do that for a relatively low price.

A logical follow on would focus on the landing page and conversion, but that is beyond the scope of this original experiment, which was to show you the mechanics of lowering your CPC.

I think the parent is right: Businesses don't just want to get people through the door, why attract people who will never buy anything at all, ever, and moreover will drop their food wrappers, mess up the produce abuse your staff, etc.. I think there are similarly toxic customers for online businesses (though I'd guess they tend to waste less resources) - e.g. those who will use things and return them for a refund even when they're not faulty.

Slightly aside, aren't the ads targeted to Pakistan cheaper because the ad costs as a proportion of revenue per item are likely to be far higher if the ads are simply charged as if for a company operating in a "richer" nation? Isn't ad competition higher in richer nations?

To make it clearer - using the Big Mac index - Denmark comes in at $5 and Malaysia at $1.50. I'd expect advertising costs to people in Malaysia to be at least a third of those to people in Denmark (barring other over-bearing factors).

Definitely that's the reason the ads are cheaper in less developed and less wealthy countries. It's simple supply & demand: obviously there are more people targeting ads towards users in the U.S. and other wealthy countries.

We touched on a lot of this in the article, and in fact the title clause "(and why you might not want to)" was referring specifically to that point!

That being said, there is a tremendous amount of useful techniques you can learn from this approach, and you can choose apply those towards whatever demographics you are marketing to.

It's up to each person to decide the specific goals of their ad campaigns, but I'm hoping this article outlines some techniques that will help you achieve those goals.

EVONY, not Eve.

You're looking at this the wrong way, let's say you have some rotting hops, malt and barely and need to sell this fetid stuff. The key is to get your product associated with attractive women. if you do it long enough there will be an association. I bet the wine industry could get men drinking wine if they used the same approach.

I run Facebook ads all day at my job and you're definitely going to have to pay much more for ads in U.S. (over Canada or anywhere else). In addition, starting your ads at a lower CPC will get you a lower quality score for your ad -- so even if you get served initially, the life-span of your ad getting served on Facebook will be much less. You will then have to reupload and start spending at a higher CPC much quicker. Of course if you're a small business, starting out at a much higher CPC (>$1) and running a lot of initial volume might not be feasible for most people. You need to start out with greater than .100% CTR if you want to have the ad sustain for a long period of time. Running celebrity ads for a short period of time will work, but you could be looking at C&D at some point if you're running a huge amount of volume.

Edit: At the end of the day CPC and CTR are useless if you can't drive enough volume from the ads based on your targeting.

Edit 2: Targeting 30+ women will get you a higher landing page conversion and CTR (obviously depends on the vertical, but in general).

"And by cowbell, I mean clickable-ness. Examining the results of our first round, and our research of what makes a photo clickable, we experimented with a new version. This one took our best performing ad (Ad B) from round 1, and zoomed in further to put more focus on the baby and the, ahem, clickable-ness."

Best part. Great article, much appreciated. Reminded me of the infamous "gray shirted girl": http://www.whattofix.com/blog/archives/2011/04/why-does-face...

That's exactly what I was thinking. I was hoping they would compare it to more subtly sexy mother pictures, just to see if the celebrity effect actually makes a bigger difference.

The interesting thing is that the least recognizable celebrity (Isla Fisher) ad performed the best. But examining the photos themselves, when combined with what we learned about the things people click on, it became more obvious why that was the case.

With more time and a bigger budget, we'd like to do further experiments to determine what other aspects mother (or even father) photos draw clicks.

High clicks alone, even if they don't turn into customers, are worth something in that you can bid less for advertising. Depending on the numbers, it may be worth styling the ad and targeting the ad partly to include those high clicks even if they aren't turning into customers. You'd still get to bid less than if you just targeted for whoever is most likely to become customers. You'd have to crunch the numbers to see if it is worth it, however.

I definitely agree. It all comes down to user intent: Either you can try to sell Facebook users on your product in the small ad space that they probably end up ignoring, or you can get them to your site first and convert them into customers once they are on your website.

Personally, I think the latter approach works better. Get lots of cheap clicks to your site, and work on conversion when they are there. It's too hard to create a convincing pitch for your product in the tiny real estate provided by Facebook's ads, when the users' attention is already distracted by a lot of other stuff going on.

Really good post. Keyword/interest targeting on Facebook is a gold mine that is completely ignored by 99% of advertisers.

Facebook Ads reminds me a lot of the early days of AdWords. Whoever was first to bid on long tail keywords must have made a killing. You can be that person in 2011 bidding on long tail keywords on social ads.

Results seem to vary considerably depending on the market and keywords, although it does seem that with some patience and experimentation you can find that high value/low cost sweet spot.

They key is patience. I had tried Facebook Ads several times to no avail, but after reading one of your blog posts as well as several others, I went back to the drawing board until I started finding the variables that made it work. I'm hoping this post expresses some of the lessons learned.

Keywords are used by a majority of advertisers in my large sample.

I had no idea that you could view demographic info of people that clicked your ads. In fact, I was about to request it as a feature. Even after I read this, it took me 5 minutes to figure out how to get there. Makes me wonder if they buried it on purpose, or if it's another example of very poor FB UI.

Great article, but I think your marketing message needs to be adjusted for conversion a little, which will backward-waterfall into CTR and CPC. Your ad is missing your value prop, it does not communicate what it will do that FB does not do already. Secondly, since you're already talking to moms, so why show moms more moms? If your message is privacy or continuity or whatever, why not show an ad with someone that fits your physical bill having the problem that you're trying to solve so you attract customers (who are having the problem you're trying to solve).

BTW, the photos with high CTRs you describe will alienate family oriented people - that is why very few businesses go there.

> That way it can milk higher CTR from ads with CPC bids by putting them at the top of the same page, and still get just as much money from your CPM bid at the same time. That’s my theory, at least. Sneaky.

It's just the result of their auction, not intentionally sneaky. Ideally when you choose CPM, you'd be allowed to specify different bids for different positions, or at least be randomly inserted into the list since there's no way for CPM bidders to specify their value for each position.

While most of the points made are valid, the main reason they likely got cheap CPC was targeting countries where there's little competition.

Yep, that's essentially the moral of the post. While cheap clicks do exist, and you can find them, you need to ask yourself whether it's really worth it for your business.

Before you decide to copy these tactics, you might want to read up on the law regarding using the name or likeness of another without their permission (i.e., grabbing celebrity photos off the internet for your ad).

http://www.citmedialaw.org/legal-guide/using-name-or-likenes...

I was wondering the same thing. How are they getting around using a celebrity's image and likeness without permission?

Are they trying to make the "thumbnail" argument?

I'd really be interested to see how they got around this without paying for a license....

so for 1 cent you can get few clicks from Egypt and Pakistan.

nice.

you are just 2 steps away from becoming next internet {b/m}illioner. if only you can:

1) attract millions them @ 1 cent price. screenshots shows just few clicks at that price

2.1) find advertisers willing to pay >$10 cpm for these visitors

2.2) make them spend >1 cent on average.

Reply on news.ycombinator.com