The lean startup methodology (popularized by Eric Ries in his book The Lean Startup) calls for startups to develop products iteratively—that is, to build a "minimum viable product" that will gauge demand and maximize the ways a business can learn from its customers while minimizing investment on the product itself. Customer feedback is then applied to quickly develop and launch the next version.

The thinking is that by focusing on discovering customer needs rather than predicting them, companies won't waste significant resources building products that nobody wants.

The concept of "lean," however, doesn't end there. The rise of real-time bidding (RTB) for online display advertising now allows marketing departments and agencies to build lean display ad campaigns.

In the past, creating an online display ad campaign meant first conducting extensive market research both to define a target segment and to identify the aspects of a campaign that resonate with that audience. Then, creative was developed, ad space was purchased, and the campaign was run.

The problem with that approach? Advertisers have no clear idea of how successful a campaign will be until it's over or until a sizable amount of data has been obtained.

Though it's possible to optimize a campaign based on the past performance of similar campaigns, the catch is that no two campaigns are alike; moreover, making accurate predictions this way can be very difficult and tedious. Whatever happens, in the traditional method, by the end of a campaign, significant resources will have been spent on the project with little control over its operation and optimization.

Real-time bidding, on the other hand, allows marketers new ways to optimize and adjust display ad campaigns on the fly. It's an approach much like the one Ries spelled out in The Lean Startup. With RTB, marketers can quickly and cost-effectively test each aspect of a campaign individually, and only scale all aspects and bring the campaign up to cruising speed once it's clear they have a winning formula.

Here's how it works.

1. Test

The lean marketer's first step is to quickly develop creative content and begin running it across a wide range of sites. The key here is not scale, but variety: You are looking for a representative sample.

Create multiple campaigns with various targeting parameters. Diversify the placement of your content across major media, long tail sites, blogs, and social networks. Experiment with ad position and unique page placement. Reach publications that target different demographic profiles.

The goal is to capture baseline metrics that can be used to test against.

2. Optimize

Next, the lean marketer optimizes each component of the creative itself via split-testing.

Split testing, or A/B testing, is essential to improving click-through and conversion rates. It involves running parallel campaigns in which the only difference is a single variable: for example, background color, wording, or a central image. The campaign that performs best becomes the benchmark.

You then go on to sequentially test other variables, until you have a creative optimized for peak performance.

You'll identify elements of your campaign that can cause friction and so slow down conversion; for example, the number of form fields, copy length, and less obvious hidden frictions such as load time should all be taken into consideration.

3. Scale

Once the creative itself is optimized, the marketer needs to refine the scale of the campaign. By monitoring how the campaign performs on all of the sites and with various audiences, you can divert spending away from audiences and sites that do not perform to satisfaction, and invest more in sites and audiences that perform particularly well.

Identify the elements that can be monitored, as well as the campaign goals, so you don't become overwhelmed by incoming data. Most real-time ad platforms and analytics packages give digital marketers the ability to monitor how many impressions an ad has received and how many clicks the ad has earned, as well as site engagement and demographic information. Internal data like online purchases or actions should also be monitored.

This is Ries's validated learning process for launching a business, where feedback from the market informs each successive stage of a business strategy. Today, marketers can apply the same principles to their display ad campaign strategies.

With programmatic buying and selling, and the ability to optimize on the fly, an advertiser can confidently scale a campaign, knowing that every component has been tested and tweaked to maximize the chances of success.

* * *

Real-time bidding is empowering a whole new generation of "lean marketers," who are fundamentally changing the way display ad campaigns are run. If the online display ad industry provides convenient and cost-effective opportunities for all types of marketers to adopt these methods, display advertising may undergo the same accelerated growth that paid search underwent a decade ago.

It can happen... if marketers are given the freedom to roll out iterative ad campaigns and the opportunity to move nimbly from one step in the process to the next.