Are your media dollars telling you a story about your consumers?

When you run a paid media campaign, do you really know what your audiences are trying to tell you? In the process of developing your audience targets, how granular should you go? The more personalized your segments are, the more likely you will be able to drive true insights, simply because it’s easier to act on an insight regarding “women in their 40’s who like to shop online” than it is to act on an insight regarding “women” as a whole. Attribution becomes meaningful because you’ll actually know who is doing the things you want them to do, which makes it far easier to try to figure out why.

When you segment down far enough to get to relevant audience characteristics, you can reach and deliver messages to a specific audience that reacts in a specific way — and then build your marketing to leverage those behaviors. You should construct your segments based on what you’re trying to learn about your customer.

Let’s say you’re advertising a new women’s clothing line. It’s one thing to know that women in a certain region responded best to your message. But what if you knew their town, education level, and propensity to like pop music or country? What if you knew they liked Passion Pit over Taylor Swift? Would that change your campaign? It should. Technology allows us to fine-tune our messages so that each ad reaches the right person at the right time, on a channel where that message is going to be best received by them.

The more granular you make your segments, the more campaign parameters you’ll have to include, and the harder your campaigns will be to design and launch. In Facebook Power Editor, for example, you could potentially create hundreds or thousands of ad variations on a single campaign. If you have a small digital marketing team, this volume could be overwhelming to manage well. Think about it, even if each ad took you a single second to make, 100,000 ad variations would take over 27 hours of mind numbing work to program into your campaign.

The value this type of approach adds, however, is incredible. If you’re a new brand or running a new campaign, intricate segmentation can challenge your understanding of who your consumers are and how they react to your messaging.

“The concept of automated ad optimization works because the process of breaking down audiences into finer and finer target segments helps us find the best possible outcomes for every campaign. More precise targets will find your content more relevant and convert more often than broader categories.” — Jordan Baines, VP of Strategy & Analytics, Accomplice

When building and testing your audience segments, ask yourself the following:

• Which attributes make sense to test against each other or as a group?
• What do you want to learn from this campaign?
• How will your findings impact your creative or targeting next time?
• Which adjustments will help you reach your goals?

Optimization software allows you to automate many of these processes so that you can gain more insights, faster. If an audience segment isn’t producing results, you don’t want to spend money trying to reach them.

This type of granularity used to be unattainable. With new technologies, your team can dig deeper, iterate faster, and reach the absolute best audiences for their messages. If you really want to connect with your customer, it makes sense that you should use every interaction to learn who they are.

This post originally appeared on LinkedIn Pulse.