You spent millions of dollars on a new data warehouse. It’s been deployed. Things are going swimmingly. Then, a new question that impacts an important business arises. There’s an obvious variable to answer the question. Unfortunately, you don’t track it yet.

Here are some quick steps you can take today to reach an answer faster.

First, relax! The data warehouse is not broken. It is still money well spent. In all likelihood you are starting to delve into harder problems. Remember, if you can predict future needs with perfect accuracy, you are probably in the wrong business.

## 1. Use a Proxy

Is there an alternative variable you know, or believe, is highly correlated to the missing variable?

Here is a very simple example to get you started in this direction: Say you are traveling to a new destination and need to know how much gas you will need. With no trips to the destination, there is no data on how much gas is needed. A good proxy is the amount of gas needed to travel to different destination a similar distance away for which you do have data.

Related Article: You're Drowning in Data — Now What?

## 2. Reframe the Question

Oftentimes our brains get locked into one way to solve a problem, when in reality there are multiple paths to success.

Returning to our traveling conundrum, rather than asking how much gas is needed, ask how many miles you have to travel. There is a subtle difference here with the proxy case. In the proxy case we used the same variable (gas needed) from a different context (alternate destination). In this scenario, we have changed the question allowing a different variable to solve the challenge.

Here are some starter questions that can help along this path:

## Learning Opportunities

• What needs to be true?
• What needs to be false?
• What is the final outcome I want to achieve?
• What is an intermediate step I can solve?
• How have other {people, industries, competitors, friends} solved this?
• Ask why? Then again. And again.

## 3. Run a Quick Online Survey

Sometimes just a small amount of data is sufficient to answer the question with a reasonable degree of accuracy. Tools exist today to quickly and effectively survey the world.

Running a survey of a population as small as 100 folks is likely to return a reasonable estimate of the gas required to travel from point A to point B. Will it be perfect? Unlikely. Will the true value be within an acceptable tolerance? Almost certainly.

Related Article: 9 Things Holding Back Your Data Analytics Strategy

## 4. Make Assumptions

Every once in a while, you just have to guess or plant a stake in the ground for what you believe. There is nothing wrong with doing this. Just be sure to recognize that you are making an assumption, and to the extent possible articulate why you believe it to be true. This will be especially beneficial three, six, 12 months from now when you have actual data to compare to.

A caveat here: This is different from when the data exists and you choose to ignore it. Henry Ford and Steve Jobs are both famous for ignoring the data with strong (true) convictions that their actions would change the data.

## 5. When All Else Fails …

Go get the data. Occasionally, the best answer is just to hit the road.