Facebook Ads Don't Rock

(If you like this post, come check out my new blog, Marketing With Meaning, where I am trying to literally "write the book" on a new way to do business by creating marketing that people choose to engage with, marketing that itself improves people's lives)

Elvis Presley once sang, “Before you abuse, criticize and accuse…walk a mile in my shoes.” Last week I walked a mile in the shoes of a Facebook advertiser and can now officially criticize Facebook as an option for marketers who seek the Holy Grail of digital marketing. The purpose of this blog is to drive improvement through challenge; and I hope that by challenging the hype I can help improve the future of digital marketing.

Facebook is definitely on a roll. The company was recently valued at $15 billion in an investment by Microsoft, it is gaining traffic, stealing employees from Google, and just took up its revenue forecast and staffing levels.  Forrester claimed the sheer mass of the company's page view inventory would make it a winner, especially with small businesses.

All of this money and attention is a gamble that the millions of daily page views will be populated by paying advertisers. The hope is that, like Google Adwords, Facebook members will pay more attention to ads that are closely linked to their personal interests.  For example, the manager of a punk rock band who is getting ready for a show in Detroit can create an ad targeted to the people near Detroit who say they like punk.  These 40 people (true number) should look at this ad as added value.

In my day job as Chief Marketing Strategist at Bridge Worldwide, my role is to help clients navigate from traditional to digital marketing. That means I have some personal incentive to jump on the bandwagon of anything digital. I am also a frequent Facebook user, so I have some personal interest in seeing the service succeed. But my experience in running a Facebook campaign leaves me unable to recommend the service—and makes me fear that a new bubble might be forming. Let me share my experience.

My Experiment:

Two weeks ago, my company ran a 24-Hour Rock-a-Thon event to benefit the Cincinnati Fine Arts Fund. It was basically an excuse to play Rock Band and Guitar Hero all night long (you can see the entire story here). A few days before the event I decided to invite my network of Facebook friends to either join in person or watch us on a live feed online. After setting up the invitation, Facebook asked if I would like to advertise the event. I clicked “yes” and decided to run a little experiment.

I really enjoyed the Ad Manager interface. I was able to quickly create an ad, play with different targeting audiences, set a budget, and monitor progress. My main goal was to drive traffic to our event via an ad that targeted people over 18 who say they are fans of Guitar Hero. There are a little over 42,000 of these people within the entire Facebook population of 24 million. You can see my ad at the top of this page.

My hypothesis was that this highly-targeted ad would drive a good deal of traffic (clicks). After all, Guitar Hero fans are really into their craft (yes, I am one) and it seemed to me that getting a targeted ad would drive much higher clicks than the industry average of around .1%. I ran three test legs:

• Targeted CPM (to the 42,000 Guitar Hero fans)
• Non-Targeted CPM (to all 24 million-something U.S. Facebook users)
• Non-Targeted CPC (same as above, but pay-per-click)

Results:
You can see for yourself what I encountered after a ten day campaign...

So, with a highly-relevant ad served to a highly-targeted audience, only 12 of the 71,023 ad impressions led to an action. That’s a .02% click-rate.  As an advertiser I was completely under-whelmed. Sure, I only paid $22, but I paid almost $2 per visit. I achieved no scale, no efficiency, and no real results.

But it gets more interesting, my completely non-targeted ad earned 23 visits among 149,081 impressions. That’s the exact same .02%.  Extreme targeting made zero difference. 

What does this mean?

I believe this means Facebook advertising is a failure; and others are beginning to agree.  A few other advertisers are sharing their poor experiences. Take Venture Capitalist, Fred Wilson, and Gawker’s Nick Denton, who also find Facebook ads to be worthless. Both of these guys invest in ad-supported net companies. Throw in Google’s announcement that they have not met expectations in MySpace or YouTube ads. Another buddy of mine in an unnamed Fortune 100 company reported similar click results for a huge, recognized brand.  MySpace is even running a half-off sale to try and move unused ad inventory.

The big moose we need to put on the table is that people have learned to ignore information on the web that is irrelevant to their task at hand. Google Adwords work because people are actively searching, and the ad units sometimes add value. When updating our profiles on Facebook or watching videos on YouTube, we simply have no reason to give ads any attention.  That's why at least one fellow blogger called Facebook "the worst traffic on the net."

Some will say that “clicks are the wrong way to measure internet media” and that studies show people exposed to banner ads buy more product, just like they buy more product when they are exposed to print or television ads. Maybe. But digital marketing and extreme targeting is supposed to offer more than a different way to hit an eyeball.  That's not worthy of a revolution.  Rather, we should expect digital marketing to build deeper engagement and drive action.

I wish more advertisers would come out and share their (poor) results, and more investors/analysts would challenge these valuations. Perhaps user experience legend Jacob Nielsen nailed the issue in Advertising Age in March:

"There’s a huge financial incentive to say advertising works. To say that it doesn’t work—I don’t get anything out of that."

I believe we digital believers do have a huge incentive to not waste dollars either as advertisers or investors. If we fail to closely analyze digital alternatives we might blow ourselves another bubble and hype ourselves back to the unemployment line.

Instead, we need to challenge the impression-based advertising model and create marketing that consumers actually want to engage with. Now that’s a marketing model worth the hype, and certainly an improvement for both marketers and their consumers.

(UPDATE: George Colony at Forrester has some great counterpoints to the belief that advertising will continue to drive business models.  Jon Fine at BusinessWeek reports more on how making money on YouTube will take longer than expected.  And Chris Anderson at The Long Tail reports that broad interest targeting is working better on focused networks through Ning.)

(MORE UPDATES: Paul Soldera takes my challenge to do better and while he improves his click rate, he agrees that Facebook is in trouble.  AND in this week's Advertising Age my new company-related blog, Marketing with Meaning was featured and included a piece on this Facebook experiment.  Neat stuff.)

(If you like this post, come check out my new blog, Marketing With Meaning, where I am trying to literally "write the book" on a new way to do business by creating marketing that people choose to engage with, marketing that itself improves people's lives)

Comments

[Andy Zilch]

Bob,

I agree w/ you that Facebook targeting is probably over-hyped as the solution to world hunger. Couple of challenges whether the experiment supports the conclusion that Facebook is a total failure.

Was the targeting relevant? Does Guitar Hero (the targeting criteria) = Rock Band (the CTA)? The CTA also calls out Cincinnati, could it be that our hometown is less relevant to 18+, Guitar Hero fans? Seems like a Cincinnati Rock Band arrow aimed at a Global Guitar Hero target.

Was the idea compelling to the target? Do Guitar Hero/Rock Band fans watch online Rock Band sessions?

Look forward to your thoughts!!!

Andy

[Andy Zilch]

[edit of above] ...supports the conclusion that Facebook *advertising* is a failure.

[raman]

am really glad you wrote this article bob. agreed w/ a lot of what andy wrote though. was either the creative or the targeting (global VS local) compelling enough for your audience to engage?

regardless, you're dead on RE. the context of facebook ads. when consumers are on facebook, they are more actively engaged w/ their environment, and an ad that does not ad value to the social networking experience is more likely to be ignored (VS the search ad during the search experience).

there's a lot of myth-busting that needs to happen, but also a lot of work that needs to be done around quantifying what engagement ultimately does for moving cases out of the door, one bottle of detergent at a time.

[Bob G]

Thanks for weighing in, guys!

I agree that there could be some reasons why this ad might have performed a little better. But we're really talking about the margins here. Check out some of the other blogs I've linked to and you can others are not faring much better.

And I do not believe "marketing on Facebook" is a failure, but that pure advertising like I did IS a failure. Bridge has used Facebook very well in recruiting, for example, as it allows us to quickly and cheaply provide an inside-look at our agency for prospective hires. And I think you can look at the case study of Colbert Nation for more evidence that there is some marketing potential.

The irony is that some of the best things marketers can to on Facebook are FREE. For example, creating a brand page, getting fans, or creating an event. And those who have praised Facebook or valued it sky-high keep saying that "it's all about the eyeballs."

My meta-point is that "eyeballs" is meaningless as a measure of value in online properties. And even "really, really targeted eyeballs" is not proven to do much better in this case.

What do you think?

[Andy Zilch]

I do agree w/ the main point that pure advertising on Facebook is mostly wasted because the ad is not baked into the experience.

Curious how you think about Yahoo's Consumer Direct service then to place contextually relevant advertisements. Seems like the same process: target people based on their demo's (FB's run a little deeper) with an online display ad. I've always seen very strong ROI's behind a Consumer Direct campaign.

What's different? (besides the fact that it costs several $100,000 to get into Yahoo and an army of people -- read: unaffordable to small companies/brands)

[Alberto Cottica]

Great post, very useful. Thanks for sharing.

[Raquel Hirsch]

Excellent article, Bob. Thank you for calling for the sharing of data and the acknowledgment that “exposure” does not automatically lead to business success!

However, here is a thought for you:

Clearly, as you demonstrate, marketers do not always have power over to generate *visits* to the site (no matter how much they “offer” to pay), but they *can* control what they present to web visitors.

If that is the case, why not run online experiments to optimize the conversion rate?

In other words, it is entirely possible that the ad you created for Guitar Hero was not ideal to the target market. In a conversion optimization scenario, you would create multiple versions of that ad, serve them randomly to web visitors who qualify and determine – based exclusively on statistically valid experiment results—which version has the highest conversion rate.

Anyway, this is precisely what we at www.WiderFunnel.com do and *all* our clients have seen conversion rate lift ranging from 10% to 270% running these sort of experiments… and we are so very often surprised by experiment results!

[Bob G]

I completely agree, Raquel, that there are things I could do to test and improve my results on this ad. However, let's look at the numbers.

If I was to get a 270% increase in my conversion rate (the high end of what you quoted here), I would have received 44 clicks instead of 12 clicks. While better, such a number is still meaningless in terms of business impact.

We can try to be better and better at interrupting people, but we're still just playing at the margins of something that is ineffective overall.

[Paul Soldera]

Hey Bob, nice post and interesting argument. It caught my attention so I wrote about it on my blog. My point was that your results are more easily explained by the fact that Facebook just isn't good at identifying GH fans, not that everyone (GH fans or not) are uninterested in ads. Same ultimate conclusion, but gives some hope that FB could improve its service.

[Bob G]

Thanks for the comment and link, Paul - is there something you can share with real results? What do you think would be a "successful" click rate if Facebook better targeted?

[Paul Soldera]

Personally, given the nature of what FB is, I would expect the CTR's on ads to a properly profiled and targeted audience to be slightly better than non-targeted banner CTRs - maybe .1-3%? Basically because you encounter them in a similar way - as an interruption to what you are doing, but they are about something you at least MIGHT have an interest in. On a badly profiled audience (which is what I expect you have here) obviously less (.02% seems low, but maybe it just is really badly profiled?).

I don't have any real-world numbers to share (just a few , probably bad, guesses). But if I come across any, I will post back here.

[Tim Maly]

Great article. As you mention in the comments, the funny thing here is that Facebook is actually a fantastic marketing platform.

Having a group of people who are interested in what you have to say and having an efficient system to invite them to events and for them to share those events to others is pretty much perfect.

I know that I end up going to a lot more shows and other events that I wouldn't otherwise go to if it wasn't for Facebook invites.

Too bad for the people at Facebook that the best marketing vehicle the provide is one they can't earn income from.

[Paul Elliott]

As other comments have more tactfully put it, your click-through-rate sucks because your ad and event suck.

You need to mention your target keywords (Guitar Hero, Rock Band, Cincinnati) in the title and graphic of the ad.

[Bob G]

Paul, so I'd improve results by 400%? Those are STILL have meaningless numbers.

Post some real results please. You can create your own perfect ad and test it for about $10 and 20 minutes. Prove me wrong.

[Paul Soldera]

Is it hard to believe that only 1 out of 4 people on FB is properly profiled? Seems reasonable to me. And yes, they are meaningless, but you did asked what I thought, so I gave you a guess :)!

But I love a challenge, so I will do some investigations to see if I can't actually get some real numbers.

[ak]

Bob,

My background is with Adwords and I've built nice businesses around it. I dabbled with Facebook advertising with the platform they had before they officially launched advertising. We were targeting pet owners and there were thousands of them too. I wrote good ads but got incredibly poor CTRs and conversions. I have banner ads that performed better than Facebook ads.

I can't help wondering though if part of the problem is just that Facebook ads are predominantly located in parts of the screen that aren't typically looked at?

[Bob G]

Thanks, Paul - I'll be anxiously awaiting :)

AK - I think you're hitting on part of the issue, the ads are not "in the consumer's face". However, by putting ads more upfront, Facebook risks worsening users' experience. Catch-22...

[Frox]

> If I was to get a 270% increase in my conversion rate
> (the high end of what you quoted here), I would have
> received 44 clicks instead of 12 clicks.

also, while those 12 clicks cost you a few dollars, thow much would that optimization cost you?

[Raquel Hirsch]

"also, while those 12 clicks cost you a few dollars, how much would that optimization cost you?"

Could cost you nothing: Google Website Optimizer, for example, is free. But you need to know what to test.

[Javier Broch]

Bob - Agree. What if the %.02 are clicks by mistake? Any statistics on those? What did the users do after they visited your site? Overall attention could be worse. I'll bet there is a large % of Facebook traffic caused by social pressure, unwanted time wasted, just to belong. Banner blindness stats are worsening as more social media is available. Huge bubble.

[Paul Soldera]

Got some numbers.

http://insightbydesign.blogspot.com/2008/05/do-facebook-ads-rock.html

[Peer Lawther]

Interesting article, thanks Bob. I managed to get 0.15% CTR on adverts based on interest, but still, that's not good enough. What I did find was that the best audience to target wasn't those who you'd associate with Facebook, but older demographics - more here:
http://www.rubbergenius.co.uk/2008/05/facebook-adverts-dont-target-young.html

[One User a Day]

I don't think it's fair to quickly jump to a conclusion that facebook ad is a failure. Facebook ads and beacon are still young and still need a lot of work.

I've also ran an ad on facebook to see how it works and compare it with google adwords. I wrote about the results and conversion rate here on my blog.

http://www.oneuseraday.com/2008/05/facebook-social-ads-vs-google-adwords/

Google adwords wasn't an overnight success. Anything disruptive like this will take a while to take effect and when it does, it will be good for everyone. In the mean time we just have to continue to give facebook our feedbacks like this and hope that they listen.

[Bob G]

Thanks, Tri, I think you actually proved my point again with the same .02% CTR on your experiment.

Why isn't it "fair" to jump to a conclusion that the ads fail? Isn't it "unfair" for people to expect that just attracting eyeballs will lead to results for marketers? Hype is unfair to our industry and our advertising clients.

I'm not anti-Facebook, I'm anti-interruption and anti-dot-bomb-hype.

[Jason]

I disagree with this ad, I believe it depends on the audience you are trying to reach. For example, I use facebook ads to target college students 18-24 for my website http://www.mybooksforcheap.com

By far facebook ads have been the best type of advertising thus far. Also you can set the settings to only charge you per click.