Data is reinventing the online advertising market. Today, algorithmic media buying is one of the fastest growing areas in online advertising and without data, it wouldn’t even exist.

This isn’t just a consumer advertising trend, increasingly B2B advertising is being driven by data as well.

However, the problem with online data is it is inherently dirty. Collected from registration forms, browsing behavior or appended offline data sources, collecting online data is an imperfect process at best.

So what is the smart way to use dirty data to improve your advertising? Glad you asked.

Here are three steps that will guide both media and creative.

If you want a better understanding of the data that is available about you, you can view your profile on BlueKai or Bizo. In my case, only 70% to 80% of the data points from Bizo are correct.

1. Media Buying: Start Testing It

If you already advertise online with targeted publishers and relevant data is available, test it. Remember that your test covers both the data and the space where the ads are running. Many networks use low-quality inventory, be careful not to dismiss the data too quickly.

Yes, some of your ads will reach the wrong person, but they will on trade publication sites as well. Trade sites reach reach students, marketers, vendors and journalists and draw search and social media traffic from a broad audience. Depending on your audience, you may find using data to target is significantly more cost effective.

For B2B marketing organizations, Bizo continues to develop more robust data. Providers like Madison Logic, Demandbase and D&B are worth watching in the B2B marketing space as well.

2. Creative: Create Serendipity

Serendipity isn’t a banner ad or landing page that knows your name, company or industry. That is merely personalization and a mechanism to capture your attention.

Serendipity is when you deliver exactly the information needed, without it being requested or making the recipient feel like you have hacked into their email.

In B2B marketing, it is using information about industry, role, seniority, and more in order to establish immediate relevance for your message or offer, not to outright personalize it.

Yes, it is a high bar for marketing, but if you don’t aim high, you will never get close.

3. Caution: Protect Your Backside

This is the step everyone misses.

As you strive to create serendipity on the back of data, also consider what happens when the data is wrong.

Does your message mislabel and potentially offend your audience by implying you know who they are when you really don’t? If so, you have a problem on your hands.

Instead, highlight what the data indicates may be most interesting to them, but leave them in control.

Your Turn

What tips do you have for effectively using imperfect data? Share them in the comments below or with me on Twitter (@wittlake).

Photo Credit: Bill Hails via Compfight cc