Simon’s Blog

Facebook has just revealed the details of its new social advertising platform. At first glance, the ideas seems pretty smart, making excellent use of the key strengths of Facebook. However, I think there’s actually a massive potential problem with the new offering: a problem that might mean Facebook has no chance of reaching the much talked about future $100B valuation. This is the offering that Facebook has bet the company’s future on - it’s a huge gamble.

Before looking at what that potential key problem with the new ads is, how does the new system actually work? Here’s the short version (as I currently understand it). If you’re an advertiser you can do a couple of things:

1. You, as an advertiser, can tap into Facebook user profiles to target ads to particular types of audiences. For example, if you are selling fizzy sugar-water (e.g. Coke), you could choose to show ads to any Facebook user that is a member of the Facebook group - Fizzy Sugar Water Drinkers. So, those users would then see ads such as - “Love Fizzy Sugar Water? How about an ice-cold Coke?” In principle, then, these ads can be highly targetted.

2. You, as an advertiser, can buy the right to have ads embedded in Facebook user’s mini-feeds. So, if a Facebook user puts something like, “I’m really loving Coke with ice at the moment”, then Coca Cola could choose to embed an ad for Coke there. So, all that user’s friends would then see their friend saying, “I’m really loving Coke…”, and also see a big ad saying, “Enjoy Coca Cola! It’s not too late to go out an grab a bottle right now!”

Now, it’s early days, and I might have misunderstsood… but I think that’s how it’s going to work. So, at first sight, it seems quite smart. The ads aren’t random - rather, they’re highly targeted: either in terms of an individual user’s stated interests and who they are, where they live etc.; and/or in terms of being relevant to what a user’s friends are doing (which should be of interest to the user). If you own a small vegetarian restaurant in Paris, you could imagine targeting vegetarian people living in Paris. Or if you sell widgets, and someone who bought your widget says, “I love these widgets - they ROCK!”, then you could put up an ad for said widgets for all their friends to see. Such a fine level of granularity in targeting hasn’t really been achievable before.

However, if you think about this for a second, it’s actually far from obvious that these kind of Facebook ads will be effective compared to Google-style ads, which are targeted via keywords. The point is this: if you are a Facebook user, nothing in the way these ads are displayed relates to whether you’re interested in buying a product or not. Compare that to Google. If you’re searching Google for “pizza”, there’s a good chance you’re looking to buy some pizza… which is why you’re searching… which is what makes the advertising so effective. It’s because Google ads are so genuinely effective, that so many businesses spend so many billions advertising with Google.

Facebook ads are different. They actually seem rather similar to the kind of ads you probably already receive by e-mail and SMS. You know the kind of thing. Imagine you’ve bought something on-line or by using your mobile; so, the company now has your contact details, and they know you bought what they have to offer… Armed with this information, then then send you related information about their products/services e.g. - “Hi! Don’t forget to order flowers for that special occasion - we’ve got a two-for-one offer on at the moment!” These ads are highly targeted, as are Facebooks. Actually, you could argue they are even more hightly targeted than are Facebook’s because these ads are actually targeting previous customers. The question is - do they work? On balance, I’m not sure they do. Speaking personally, these kind of ads often work as a net negative on me; on the other hand, they do help to keep the brand in my mind.

So, it looks like the new Facebook ads might be the kind that will drive sales on the occasions they happen to catch the customers at the right time. The big risk is that they end up annoying potential customers by being presented at inappropriate times. Advertisers will have to be careful about how they handle this. The bottom line is this: people want to see ads when they’re considering/open to making purchases; and not so much at other times.

It will be pretty interesting to see how many, and what type of, companies end up buying this kind of advertising. Right now, it’s hard to see it making a lot of sense for many orgnaisations other than a handful of big brands… and a handful of big brands isn’t going to spend ten billion a year on Facebook ads. Having said all that - this type of super-fine targeting of ads hasn’t really been possible before. There’s a chance that the sweet-spot for this technology is a lot bigger than I think; and that it doesn’t overlap with Google’s keyword-based search ads.

One thing is for sure - it shouldn’t take too long for advertisers to figure out whether their money is better spent with Google or Facebook. It’s really easy for companies to experiment with these kind of technologies; and given that it’s so easy, it would be pretty dumb not to do the experiment to find out what works in individual cases. So, whether it works well or not, kudos to Facebook for trying something new.