Big data isn’t just a buzzword–or the basis for understanding customers’ proclivities. It can help you hone your entrepreneurial path, too.  (This first appeared in Inc.com here.)

The Lean startup methodology popularized by Eric Reis and Steve Blank leans heavily on making product decisions using customer feedback. To most, this appears to be obvious. Gather copious amounts of relevant data from your perfect, targeted customer segment and they will, without much work, provide you insights that steer your product roadmap.

Sounds easy enough, but good luck. I’m pretty sure it isn’t that easy for a majority of the companies in their first year(s) of operation. Why? Let’s break down the previous statement.

First, “copious amounts of relevant data” are hard to come by. The companies I work with atThe Startup Factory don’t have customers yet, or only have a limited number of customers–certainly not enough to provide copious data. Without that, it’s understandably tough to build out a product customers want.

This is a big deal. You read the books on startup dos and don’ts. You listened to your mentors–the ones who really understand the needs for copious amounts of relevant data. But you can’t easily find or get enough customers to satisfy the development team–least of all your mentors or any self-respecting investor.

So what are you to do? This is the hard part. The easy way out is to build a product based on your gut instincts (they’re usually correct, since you have done this successfully many times already, right?) or with the feedback of a handful of random customers (usually friends). All this does is prolong the real questions at the heart of a startup’s real mission: What do your customers want and how do they want to get it?

Here are four steps that you can apply to just about any business:

1. Identify how many target customers you need to reach relevance. It depends on the type of business, of course. I would posit that you should talk to at least 10 percent of the number of customers you would like to have in your first year, with the caveat that there is also a minimum floor. If you have an app and want 1,000 customers, you should target, say, 200. If you’re a business utility and hope to reach 50 customers in the first year, you should talk to at least 10 target customers.

2. Figure out three ways to address your target customer segment with the questions that you need answered. A one-on-one interview may be obvious, but many times it only delivers the answers you want to hear. Interview bias will affect outcome. You need actionable evidence that your targets support your thesis. Do the targets want your product or service? How does your target want to engage in the product? Formal surveys work if you ask the questions correctly. Landing pages with a call-to-action prompt work for many Web-oriented businesses, especially if you drive some paid traffic to the landing page. Utilize Facebook, LinkedIn, and Twitter to reach and find potential and relevant targets. Don’t rely on just one channel; use at least three to create balance.

3. Calculate how much time and/or cash you need to bring these target customers into the fold. Just starting out? Then allocate $1,000 to build the landing page and drive $500 of traffic through Google Adwords to see if your target audience reacts. Can’t decide between two messaging approaches or two target audiences? Then run an A/B test using Google’s Experiments tool over a few weeks. In the iterate and discover mode, you only need to calculate to the set of answers you need to figure out if you have anything. The funds allocated at this stage are a much better use of cash than hiring a developer to code a product nobody wants. Also be ready to spend twice the time and amount of funds to get to statistical relevance.

4. Gather the data (more than you need) and analyze. If you thought you needed 100 targets, get 200 if you can. Then, sit with your team (advisers, co-founders) and pick through the results. Did your thesis prove out? Are you sure? Did new issues present themselves? Were the results action-driven and quantifiable or is the data a little soft, coming from what you think you heard in the interview? Pure, un-adulterated honesty is critical at this point.

If this takes weeks instead of days, live with it. This isn’t the time to cut corners. This should turn out to be the foundation for your next year or so, so get it right. You are now running a data-driven company.