1. Find the business KPIs that matter. Reverse-engineer your marketing performance to hit the notes that you know the senior leadership of your company will listen to most closely. Ultimately, most KPIs ought to tie back to revenue in some way — units sold, customer churn, sales performance with a certain demographic and so on.

Speak openly with your company’s leadership to determine which data line they’re tracking most closely, as well as the volatility they see in that data.

Specifically, explore any blind spots around why those KPIs do or don’t do well. Where are you facing a KPI mystery? And which KPIs should you check most regularly to keep an eye on changes?

2. Benchmark and correlate your data against those KPIs. Once you know what KPI you need to impact most, explore the connection between what marketing has been doing and how that KPI has been performing.

What are the most likely causes behind an increase or decrease in that KPI? What are the obvious correlations you can spot, and where do you as a marketer have blind spots?

And go a level deeper than just putting two data lines next to each other. Get curious and courageous.

Make sure you’re using analytics to find correlations that matter, not coincidences that don’t. And if you can’t tell the difference, bookmark those and get help investigating them.

New approaches to “explanatory analytics” can help you explain correlations, not just predict outcomes.

3. Build explicit checkpoints into your marketing plan. While you may not be able to predict what course corrections you’ll need to make, you can safely predict that you’ll need to make them. So be explicit about that.

Plan ahead for what questions you’ll investigate along the way. Where you have hunches, build in ways to test them against data. Where you have blind spots, build in ways to examine your data to fill them in.

And build in explicit permissions with your internal stakeholders so that when the data tells you that you need to course-correct, you’ll do so.

Educate them on what explanatory analytics can do that predictive analytics can’t. Find out what it will take for them to trust the data you bring them.

Changing lanes or making a U-turn is a lot less painful if you signal in advance. Making explicit agreements to course-correct helps reduce the friction and awkwardness of course corrections as they happen.

4. Let the data lead you to better questions. The only thing worse than the wrong answer is the right answer to the wrong question. And some of the most pernicious side effects of unpredictability are the unknown unknowns.

If you don’t know what questions would yield the best insights, then let your data whisper to you and guide you to smarter questions. Explanatory analytics solutions can help you explore how strong certain correlations are.

The simple act of scanning your data for strong correlation signals can sometimes lead you to serendipitous discoveries that make a huge difference. Learning how to let your data lead you to smarter questions is going to create massive advantages for marketing organizations in 2016 and beyond.

5. Act on insights decisively and courageously. According to the same Gartner report, one in four market leaders will be overtaken by a smaller, younger company by the end of 2017. One of the reasons given is that smaller, nimbler companies are more comfortable with digital technologies such as data analytics, social media and mobile technologies.

But another reason is that smaller companies have fewer cultural barriers to overcome when it comes to acting on data. At a recent Techonomy conference, business leaders from multiple companies including PayPal, Hilton and Visa talked about how larger companies that need to compete with smaller businesses must have the courage to change and act on insights.

Plan ahead for whatever cultural barriers you might face when marketing recommends a course correction. It may be as minor as adjusting your creative approach or messaging for a specific campaign. It may be as major as recommending a fundamental shift in your product or go-to-market strategy.

Whatever it is, scan ahead for where you might run into opposition. Bring other players into your process. Walk them through your data, and make sure they appreciate the strength of the correlations you’re looking at.