Marketers all want the same thing—to build relationships with their potential customers and to drive long-term value and sales.

To distribute content successfully in digital, it is important to be able to work with platforms that have two key things: scale and user data.

Targeting audiences in the search and social arenas is relatively straightforward, thanks to the vast scale and rich data of closed channels like Google and Facebook.On the open web, it’s not so simple.

Outside of the walled gardens, you might be able to target by context, vertical, device or location, but it’s often not enough to move the needle. To reach the right audience, you’re working with your own data or through a direct relationship with another data company, both of which can yield complicated and unclear results.

It gets even more complicated if you aren’t 100 percent sure who your target audience should be—and you’re probably not. If you’re going to accurately segment, you need to test your own data or data from third parties in a scalable way.

Marketers, this is how you target an audience that performs.

Sometimes, you don’t know your best target audience at the start of a campaign. Chances are that without a hefty amount of data, you’re not going to know who will respond best to a specific campaign. The first step in targeting an audience that performs is collecting that data through your own campaigns.

Once you’ve collected that data, the next steps are to find audiences that have interesting reach and scale, and find a reliable way to segment that audience.

Step 1: collect data on consumer behavior; lots of it.

To really understand consumer behavior like interests, purchase intent, and demographics in an effective way, you’ll need access to data from top data providers—this is the only way to scale.

You’ve probably already heard of some of the best data providers out there. Oracle Data Cloud provides demographic and interest data, Neustar can connect online behavior with offline behavior, Bombora can target users by intent, and Acxiom presents real-time data trends and there are plenty more out there.

Step 2: Identify Target Audiences

When you start your first campaign on the open web, you’re going to be tempted to be as segmented and conservative as possible with your budgets and targeting, since that’s generally a strong approach for social media and search marketing.

It’s important to take a different approach for content discovery on the open web. There, users are primed to discover something new—you are likely to be surprised at the audience segments that perform for your campaign.

Start with as wide of an audience net as you can, with more liberal goals. As consumers start to engage with your campaign, you can try narrowing your campaigns down to audiences you think will perform best for your goals.

Ideally, you would be able to match campaign data with audiences from data providers that have corresponding CPAs, CPCs or other performance metrics available. This may require a lot of A/B testing between audiences that are available to you and to wider-targeted campaigns.

This is how you know which data is delivering the best ROI in real-time.

Taking the measured approach to segmenting your audience can have interesting or unexpected results. For example, a very male-oriented CPG client discovered, based on real-time audience performance reports, that women were actually buying its products in droves. Based on that revelation, it began buying female audiences and targeting them as well with tailored creatives and landing pages.

The ability to see how your campaigns are performing against various audiences is a powerful thing. Advertisers need to be able to make informed decisions instead of assumptions about their target audiences.

We know that the open web lacks an efficient way to reliably segment audiences. Earlier this year, Taboola launched the Taboola Data Marketplace. The goal of the Marketplace is to bring transparency to this process. The new Marketplace allows 3rd party data providers a chance to offer their data services to advertisers in an integrated platform. It improves the decision-making tools available to the advertiser but lets the performance of the provider and their segments speak for themselves.

So the open web—if approached in a data-driven way—can provide marketers exactly what they are looking for—and consumers as well.