Earlier this year, Winterberry Group released the results of its programmatic survey, which revealed that data-driven marketing will increase 17.1% this year, reaching an astounding $20.6 billion in revenue. Moreover, 91% of marketers expect to deploy audience segmentation over the next two years. But to me this is all contingent on using quality data, a potential weak spot in some parts of the data ecosystem.

There's no doubt that audience data drives campaign performance. But is all data created equally? Apparently not. When we tested an otherwise reputable third-party data provider's gender segments against Shopzilla's fashion audience, the provider said only half of the visitors to our sites -- featuring women's skirts, dresses and handbags -- were in fact women. Our internal data showed otherwise. No wonder marketers are skeptical.

How can marketers ensure that the audiences they buy through programmatic channels really represent their target segments? Here are six questions every marketer should ask their data providers:

1. Do you collect data yourself or aggregate it from other sources?

Companies that collect data directly from consumers have complete control over how their data is classified and scored, and can target by individuals when activating that data in ad campaigns.

When third parties aggregate data from multiple sources, you won't have insight into quality or accuracy. Ultimately, the closer the provider is to its data, the more accurate it will be.

2. How recent is the data?

Recency is a key issue for marketers, especially for performance campaigns, though it is also relevant for branding initiatives. Ask your data provider for the look-back window (length of time that has passed since the consumer took an action that indicates intent) of the audience segment you're considering so you can determine if it's appropriate for your buying cycle.

For an automaker, a look-back window of three months might make sense, but for a home appliance retailer, not so much. You don't want to spend limited marketing dollars messaging consumers who are no longer in-market.

If your data provider can't provide the scale you need within a specific look-back window, you'll need to expand it. Obviously this will affect campaign strategy, creative and performance, but don't you want to know that up front so you can set expectations accordingly within your organization?

3. How is your data sold -- via public data exchange or privately?

Data that's packaged and sold on public data exchanges may be oversold. What does that mean? Messaging a group of consumers over and over for the same product category can lead to banner blindness and affect campaign performance.

4. How do you classify consumers as in-market?

We all look at products and services that are out of our reach; we're curious by nature. That means lots of consumers visit sites recreationally, not necessarily because they're in the market for those products or services. Ask your data provider how they distinguish between casual reading and actual shopping behavior so that you don't spending your budget on consumers who will never convert.

5. Will my ads pass through multiple tech platforms before they are actually seen by consumers?

This question addresses efficiency, which is always diminished when you hop platforms (i.e. buy data from one provider but execute it on another's demand side platform). Inefficiency will also affect your costs, since you have to pay both the company that provides the data and the DSP that buys the inventory. The more hands that touch the campaign, the higher the cost.

On the flip side, when data is collected, classified, segmented and activated on the same platform, not only do you gain efficiency, you also get the ability to "listen in" and optimize the campaign in real time.

6. Can you tailor and manage data based on the goals of my campaign?

Consumers in each category have a unique buying journey, and the more you present the right message to the right consumer, the faster you'll grow your business. Targeting data should reflect the specific characteristics of your audience -- which requires customization.

I'm delighted that, according to Winterberry Group, almost all marketers will leverage audience segments in just a few years. Now let's make sure they have the quality of data they need to ensure their expenditures reach their intended audience.