Ads API Profile: Efficient Frontier’s Portfolio Management Across Channels

Facebook’s performance advertising system is quickly becoming an essential part of any marketing mix. To allow companies to build more powerful tools for designing, targeting, and managing Facebook ads, the company gave a limited number of developers access to an ads API last fall. Advertisers seeking to optimize ad bids across channels using portfolio theory can do so using Efficient Frontier.

A mature marketing management platform, Efficient Frontier developed its ads API tool to use the same applied portfolio theory it had be using in its system for paid search and display ads clients. It tested the beta of its Facebook tool for a few months with several existing clients interested in getting more conversions from Facebook on the same budget, or the same volume of conversions for less.

The core of Efficient Frontier is a system which models the potential return for any ad, automatically bids in response to the potential, and then optimizes that bid against other channels where that money could be spent. Unlike other ads API tools which are rule based, the portfolio-centered system “models expected performance for different bid levels and places the optimal bid”, allowing advertisers “to bid down when it’s less important, and bid up above the threshold of [the] max bid when it gets [them] the best position, so it could give [them] more conversions for the same cost” says Justin Merickel, Efficient Frontier’s VP of Marketing and New Product Development.

Company Profile

Efficient Frontier was founded by Anil Kamath in 2002. Before starting the company, Kamath worked as a vice president of a hedge fund, where he honed the skills and understanding used to develop Efficient Frontier’s risk-return portfolio model. Headquartered in Sunnyvale, CA, and with offices in New York City; London, Paris, Hamburg. and Teynampet, India, Efficient Frontier now has 50 full-time engineers. The company’s latest funding round was a $6 million Series C led by Mitsui & Co., and joined by Redpoint Ventures and Cambrian Ventures.

Efficient Frontier  handles $1.2 billion a year and sets 5 million bids a day across paid search, display, and Facebook ads. Its 250 clients include some agencies, but mostly consist of direct advertisers representing industries including performance marketing, financial services, travel, retail, and automotive. Salesforce.com, Discover Financial Services, Capital One, Crate and Barrel, Match.com, and Travelodge are some of its clients.

There is no minimum spend necessary to use Efficient Frontier, though the company prefers enterprise-class marketers. Most of its clients spend $100,000 a month across channels. Efficient Frontier doesn’t publish a rate card, but the percentage of total spend they charge is tiered by total spend by the client.

Efficient Frontier allows advertisers to integrate their search engine account data, ad exchange campaigns, web analytics, paid search, display ads, and now social media spend. By managing all marketing channels on the same platform using a unified portfolio system, the platform can fluidly allocate budget to where cost per acquisition is the lowest.

Efficient Frontier for Facebook Ads

The new “Social” tab of Efficient Frontier currently allows advertisers to run Facebook ads, but the company says it is also looking to provide LinkedIn and Twitter ad management in the future.

Portfolios

Clients can create a different portfolio for each performance metric for which they want to optimize. Within a portfolio, campaign budgets are dynamically shifted to provide the highest return for the selected metric. For instance, clients would set up one portfolio for attaining exposure, measured by impressions or clicks, and another for conversions or acquisitions, measured by sales or CPA.

Once a day, Efficient Frontier calculates what would happen if each ad’s bid was increased or decreased, and alters spend to optimize for the desired metric. On any given day, Efficient Frontier may spend more or less than what that day is paced for in the total budget. Over time, though, the difference between the predictions and the actual traffic won’t exceed 10%.