Online Extra: Search Advertising by the Numbers

At online lender E-Loan, marketing is more about analysis than creativity, especially when it comes to buying keywords

Advertising. For decades, it has been an industry of expansive ideas, creative people, and chummy deals hatched in fine restaurants. But this world of relationships and high concepts has been under siege for three years by people up to their elbows in numbers.

Imran Khan is one of them. He heads search marketing at E-Loan (EELN ), the online lender. While the search advertising boom has focused attention on Google (GOOG ) and the Overture division of Yahoo! (YHOO ), it has also shaken up marketing departments in companies around the world (see BW Online, 10/21/05, "Google and Yahoo!: Rolling In It").

People like Khan are grabbing ever more of the budgets. Since the search ad boom began, Khan's share has risen from 20% to 50% of E-Loan's total marketing budget. He now spends $15 million a year. Not much of it buys fancy lunches. Instead Khan's department operates as a number-crunching lab.


  Search advertising's biggest advantage comes from the numbers it generates. It enables marketers to track customer behavior, and it replaces hunches with science. "The reason that we've been able to grow it, we've been able to show through scientific testing and measurement that we can optimize this channel, and we can scale it," Khan says.

The crucial calculation for a search marketer is to bid on keywords at the big search engines. The highest bidder for a word gets its ad placed next to the search results when a Web surfer types that word in a query. And then the marketer pays the amount bid when the ad is clicked (see BW Online, 3/7/05, "Making the Most of Web Ad Budgets").

Naturally, Khan bids for the keyword "E-Loan." He certainly doesn't want potential customers who start out with his company in mind to click on a competitor's ad. But how about "Eloan?" It's not spelled quite right, but he doesn't want to turn away business from bad spellers. He bids on that too. "Elone?" Ditto. In total he has bids on 3,000 variations of the brand name.

"We've spent a lot of money on building this brand," he says. "If they're looking for us, I want to make sure they find us."


  In total, Khan bids on about 250,000 key words and phrases at the major search engines. Working with Efficient Frontier, an analytics startup in Mountain View, Calif., he runs the numbers on them. Each one generates its own return on investment, and that number helps to determines Khan's bidding strategy.

But there are plenty of other numbers to throw into the mix. "I get gigabytes of click data," he says. "Cost, cost per click, and keyword rank." The ranking of each of his quarter-million keywords fluctuates during the day as bidding continues. Khan's team takes the flow of numbers coming in from the search engines and marries them with internal data. How many visitors from each keyword visit the Web site? How many of them spend money, and how much money do they spend?

With those numbers, Khan's automated system generates a return on investment for each keyword. Those help determine the bidding strategy. The process goes on and on, fluctuating much like a financial market. "It's the same math you use in managing a portfolio," says Ellen Siminoff, CEO of Efficient Frontier.


  Khan, an accountant by training, looks at the process like a linear programming problem. That may sound exotic to mathophobes, but the concept is simple. He sets a goal that he aims to maximize. In Khan's case, that's revenue. Then he factors in the constraints, such as cost per keyword.

This can get complicated. If he bids for the word "loan," for example, he competes with much of the lending industry. That drives up the price. What's more, it's a big, broad word, which means the results are much more complex than a far cheaper targeted keyword phrase such as "Tuscaloosa mortgage."

"If we can get one dollar out of an auto loan, we can get $20 if we do a mortgage loan," Khan says. "A keyword like 'loan' can deliver either of the two products. So should we buy the keyword based on potential of auto loan or mortgage loan?" He has his team do the math (see BW, 10/24/05, "Hard Questions From Google").


  Khan is continually testing and tweaking the system. He experiments with different types of ads and compares the results. He measures the online response to E-Loan's TV ads. He also tests the effectiveness of the Web site in turning visitors into customers. He can send them to various versions of the site, compare the results, and make adjustments.

It's all based on numbers. And that's what Khan looks for when he recruits. "Marketing is still considered a soft skill," he says. "But there's a shift in focus from creative to analysis in marketing. When I go out and look for search marketers, I look for computer science and economics backgrounds. Investment bankers are good.

"A new kind of marketer has emerged," he says. "I call them analyst marketers." In Imran Khan's fast-growing realm of marketing, math is the crucial skill.

By Stephen Baker in New York

Edited by Phil Mintz

    Before it's here, it's on the Bloomberg Terminal.