Digital Uncovered #4: Facebook Ad auctions
Facebook Ads is one of the world’s biggest online ad platforms, not only owing to Facebook’s 1.6 billion users but also to its ever-extending portfolio of social networks such as Instagram and Snapchat, as well as apps that have made their ad space available to the internet behemoth.
If you’ve somehow found yourself working in an advertising or marketing role, you own a business or simply want to know more about paid online advertising, understanding the core of how Facebook Ads work may just be the thing that tips the scales in your favour.
We get our participants creating their own live Facebook ads in our digital skills workshops, so to build on that and provide a useful resource for understanding what actually happens, in this Digital Uncovered article we are going to attempt to explain the single trickiest part of how the Facebook Ad platform works — the ad auction!
Creating ad relevance
How can you make sure ads aren’t just tolerated — but even valued? This is a key goal of the Facebook Ads platform. Targeted advertising means running shoes may show up in your newsfeed after you’ve been talking about getting back to the exercise habit.
But when it comes to ensuring the ads are actually something you want to see, the process gets a bit more complicated!
Traditionally, advertising has taken a spray-and-pray approach: whack them in between TV programmes or newspaper articles, and hope they reach the right people. Technology has changed this, meaning we’re now able to see ads loosely based on our interests, as judged by our online activities. But just because you bought new bedding or flights to Thailand, it doesn’t necessarily mean you’re going to value seeing ads for those items following you around for the next week.
When it comes to value-adding ads, the Facebook Ads platform has broken the previously set mould with its experience-first approach.
The Facebook Ad auction system
Facebook uses the Vickrey-Clarke-Groves (VCG) auction system to sell and select ads for the network. With VCG, Facebook is able to place ads not just based on certain keywords, but also on how they fit in with everything else on the site.
VCG operates as an auction, similarly to other online advertising platforms, but unlike some it is a closed auction.
A closed auction means that you are submitting a sealed bid, in other words the highest price you are willing to pay for 1 person to click (or 1000 people to view) your ad is hidden from other bidders. Likewise, you cannot see the exact amount other advertisers are bidding. And intriguingly, VCG isn’t necessarily looking for the highest bidder.
Showing the right ad at the right time is the goal of any advertiser, however the VCG system throws a new factor into the auction mix: how much value is lost from the rest of the system by having that particular ad on that page? In Facebook’s case, it calculates this by looking at how relevant the ad is, as well as how much other advertisers are willing to pay for the placement.
The complexities of VCG
If this sounds complicated, rest assured that it is. In fact, The VCG system is known as a black box that’s near-impossible for advertisers to “game”. In other words: no one on the outside truly understands the specifics of how VCG works. Even the Wikipedia page for VCG isn’t much help, except maybe to provide a peek into the complexity of the algorithms for economists and mathematicians. A more useful explanation what we, as users need to know about this side of VCG came from Facebook chief economist John Hegeman, in a 2015 interview with Wired:
“If you’re an advertiser and you’re getting a chance to show your ad, you’re going to take away the opportunity from someone else. The price [of the ad] can be determined based on how much value is being displaced from those other people. An advertiser will only win this placement if their ad really is the most relevant, if it really is the best ad to show to this person at this point in time.”
What this does is create a system where advertisers are best served by launching bids that reflect the true value of the slot they’re aiming for.
This is something that takes some time and testing of your ads and campaigns to master, but it also adds something of an honour system to the process of online ad auctions.
In other words: offering to pay lots of money won’t necessarily get you anywhere with VCG if your bid is far higher than what others are willing to pay. This works best if advertisers know this is a VCG auction, and trust that the auctioneer is setting fair prices with the motivation of getting good ads.
When this works as intended, advertisers are rewarded for placing ads people actually get value from seeing into the ad auction — which is the only way to get ahead on Facebook: the social network simply won’t let you spam its users by placing ill-fitting ads just because you have cash to spend.
Think of how different your Facebook news feed would be if only the companies willing to throw the most money at their campaigns would win the ad auctions!
This may cost Facebook in the short term, but the idea is that building trust will be worth more over time. Hegeman concluded:
“For anyone who is trying to build a platform with a long-term focus, where the goal is creating value for the participants rather than short-term revenue, [VCG] is going to be a pretty good option.”
Who invented the Vickrey-Clarke-Groves algorithm?
The roots of this very modern system actually stretch back to a pre-internet age, to the 1960s. Nobel Laureate William Vickrey, alongside Edward Clarke and Theodore Groves, created the algorithm that carries their names, but it was mostly an academic exercise. It didn’t really have a practical application until Facebook came along, and decided to use this system for its online ad auctions.
The ideal outcome of this kind of weighting is that ads will slot so seamlessly into the newsfeed that Facebook users don’t mind seeing them in between posts of friends’ holidays and children. They may even interact with them! The fact that the system can’t theoretically be gamed also means advertisers are encouraged to make sure their ads are a great fit for their audience, instead of spending time and resources trying to trick the system to accepting less-optimal ads.
Of course, not every ad you’ll see is perfectly relevant — anyone who uses Facebook will be very aware of this fact. But the system is constantly being tweaked, not only by experts, but also by the site’s users. You’ve probably seen this: you can click on ads you don’t like to hide them, and the reason you give for doing this will feed back into the system to give you better-fitting ads next time.
Digital ads: what does the future look like?
VCG continues to evolve, and it’s also gaining in popularity: Google has started using it to sell ads in places beyond its search engine. The ideal version of VCG is a system that benefits everyone: the public gets better ads, and advertisers get better return on their investments — the result that any advertising platform is striving to achieve.
But while there have been many improvements, many issues remain unresolved. As advertisers chip away at the problems around ad relevance, the next piece of the puzzle will be dealing with the sheer amount of ads we are shown in a day. Could there be an improvement to the system that could also account for “audience fatigue” ie. people being bombarded with so many ads they begin rejecting them? It will be very interesting to see what the future holds!
And we can’t talk about the future without talking about the upcoming advances in tech. From VR and AR to AI, there are very exciting frontiers and possibilities emerging for brands and customers to create a better experience.
Freeformers Founder Gi Fernando, who also founded Techlightenment — one of the first ever technical Facebook advertising solutions — says the following:
“The potential of AI in the future of advertising is very exciting.
With the continuing development of AI, it could enable ads to become near-seamless in-story or in-experience content adaption rather than an imposition into what a person was actually trying to do or enjoy.
The AI would stitch together different brand-related content and make it flow with the experience that you are currently having.
This means a combination of different types of digital content could become a seamless part of your experience on an Oculus Rift, a social media news feed, or even a WhatsApp conversation.”
Found this installment of Digital Uncovered useful? Share it and tweet us your comments!
Got something that you’d like us to explain in the series? Send us your ideas to firstname.lastname@example.org or contact us on Twitter, LinkedIn and Facebook!