Tweet
In the early days of Facebook, people wondered how we would make money. The obvious answer would have been the wrong answer. Here’s how we landed on the non-obvious right answer, and transformed the advertising industry in the process:
We ran banner ads in the right-hand column of Facebook.com to generate some revenue in the early days of the company. Banner ads were the standard way to monetize media properties at that time, and most people assumed that’s how FB would monetize long-term.
Advertisers loved banner ads because the format was standardized across the web. This is why Yahoo, AOL, MSN, MySpace and every other digital media property sold banners. It was easy for advertisers to create and syndicate their ads, and it scaled with publishers' page views.
But there was one big problem with banner ads - they sucked. Users learned to ignore them, so advertisers were motivated to make them ever more intrusive, which caused users to work harder to ignore them. "Punch the monkey," blinking lights & colors. It was a race to the bottom.
Shortly after I joined FB in ‘06, MySpace outsourced their banner ads to Google in a deal worth $1B. I thought we could do the same with our banner ads which were earning pennies through sketchy ad brokers. I knew we could get a lot more money and improve quality via partnership.
Zuck approved my plan, but he also had something else in mind. He wanted to invent an entirely new advertising product using a non-standard format, with non-standard pricing, designed to slot into News Feed. (Never mind that News Feed hadn’t even launched yet!)
So I negotiated a deal with Microsoft to pay us a minimum CPM guarantee for exclusive rights to sell our banner ad inventory. And I explicitly carved out all non-standard inventory which would be reserved for our own in-house efforts.
Microsoft wanted to repackage our ads with other properties, so they didn’t care about non-standard inventory. This is a classic lesson in negotiations: rather than slice an orange in half, ask whether one party wants the meat and the other wants the rind.

Within a year, Microsoft was profiting on our deal and very happy to have our scale to compete with Google, and we were earning hundreds of millions of dollars from the partnership. Meanwhile, News Feed had launched and we started working on a new native social ad product.
We designed ad units that looked like Feed stories, including social context from your friends. We knew Google had run tests where users preferred search results with Search Ads over those without, and that became our aspirational litmus test for Social Ads.
We also built an auction system for advertisers to buy Social Ads. Google had invented this pricing model for Search Ads, but it was anathema at the time for brand advertisers. They didn’t like our non-standard format or our pricing mechanism. It was a slog.
As VP of Monetization, I was optimistic we could make this new system work, but I was also pragmatic. In ‘09 when our revenue started to plateau after the recession, I pitched the idea of slapping banner ads at the top of News Feed to juice our business. Zuck’s response: “NFW.”
As FB grew, we attracted more advertisers who were willing to experiment. Some ads were great, eg “one of your favorite bands is playing a concert nearby and 3 of your friends are planning to attend.” But most were low quality as the best brands still resisted our bespoke system.
We knew if we could just attract great advertisers they would drive more engagement which would attract better advertisers who would produce even better ads. But until we could get that flywheel spinning, we remained dependent on banner ads for most of our revenue.
Then we woke up one day and realized mobile was the future of our business, and there was no space for banner ads on a small screen. One of our product leaders joked we would need to place ads on people’s thumbs to keep our revenue growing.
Fortunately News Feed turned out to be the ideal content format for mobile. But Social Ads quality was still too low for us to feel comfortable ramping ad load In News Feed. As mobile adoption accelerated, our revenue growth hit a wall, just as we were preparing for our IPO.
Zuck plucked one of our mid-level engineering managers to take over ads product leadership. Boz immediately grokked the importance of the ads flywheel, and he started searching for a crank. To his credit, he came up with an elegant solution: better targeting.
Better targeting would lead to more relevant ads which would drive more user engagement which would attract better advertisers. Rather than limit our targeting to interests expressed on FB, we allowed marketers to upload their own targeting criteria and/or find look-a-likes.
Suddenly users were seeing ads that felt relevant, timely and personalized to them, and they were clicking through at much higher rates. And with the shift to mobile incinerating banner ad inventory across the web, advertisers were more open to trying new formats.
We finally started to see users engage with ads at similar rates to organic content as small businesses and large brands competed in the auction. This gave us confidence to ramp ad load, which drove more clicks and attracted more advertisers, and the flywheel started spinning.
Sometimes innovative new business models fall from the sky and change the world overnight. More often they come from an early spark, followed by lots of iteration and stubborn persistence (and a bit of luck). It took us 7 years at FB to find our business model. Worth the wait!

Recommended by
Recommendations from around the web and our community.

Good thread on how FB went from generic banner ads to truly native ads and targeting