As soon as my cofounder and I came up with the idea for our startup, my financial instincts took over and I immediately opened Excel to run analyses. I spent the following weeks calculating the unit economics, building a business model, and guessing market size and other data points—like five-year revenue forecasts and customer acquisition costs. I even modeled more than 50 different conversion rates at various prices to confirm our assumptions. And through all of this analysis and number-crunching, I forgot about the product itself.

Naively, we assumed this was the first step in creating the next unicorn. Our thinking was this: If we can’t see an up-front path to clearly making money, our business won’t succeed.

Real life didn’t hit us until our first investor, Mark Cuban, stopped us and said, “Instead of projecting these figures, you should have built your product yesterday.” He was right. Analysis is all well and good, but the product should be the horse leading the cart. Of course, I wish I had realized that before wasting so many weeks. Since then, we’ve learned a few more lessons about running a startup:

Recognize false positives

Friends, family, and even strangers will lie to you in the most altruistic way possible. “What a fabulous idea!” “I would buy it!” “Love the concept!” Fewer will say, “That’s probably going to fail,” or, “I won’t buy that from you.” As an aspiring entrepreneur, don’t be fooled by false positives. The only opinion that really matters is your actual customers’, and even more so if they’ve been able to truly experience your product. Find a way to get your customers to purchase your initial minimum viable product (MVP), and you will reduce the possibility of failure. And believe me, try to tune out your Aunt Sally’s enthusiasm to a degree.

See also: The One Thing That’s Sure to Make Any Business Fail

Be quick to build

Abiding by Mark’s advice, I quickly reached out to an engineer to build our first-generation product. We wanted to get something—even if it wasn’t perfect—into the hands of our customers for one simple reason: to gather purchasing data from our target demographic. Before building the prototype, we sent out online surveys to gauge customer interest and gather more specific data. To our surprise, after rolling out the first-generation product, it became clear that our consumers’ interests were totally contrary to the survey results. In other words, consumers lie. Therefore, testing our MVP gave us vital, irreplaceable, and qualitative feedback that helped shape the customer experience, our hiring needs, and our company itself. Everything we learned from our MVP testing set us on a new course.

Don’t try to be perfect

In order to pull together our first-generation product, we needed to scavenge and be resourceful. If we wanted to create something to gather customer data, we didn’t have time to engineer the perfect machine (at least, not yet). So, we set up shop in our engineer’s garage and fabricated our product with used 3-D printer components, old computers parts, and surplus Home Depot supplies. Honestly, the product looked ugly, but, our hacked-together machine validated our sales and pricing variables, which in turn completely changed our previous unit economics. So be resourceful and don’t try to make everything perfect. Build even if you don’t have an investment or team. All you need is a few contractors to help cobble together an MVP and gauge market demand.

If you’re an aspiring entrepreneur with the next big idea, you need real customer data. Build an MVP and test traction. Persuading investors to invest in your new project or company will be significantly easier when you have actual customer data. And learn from our mistakes: Don’t’ get lost in building beautiful financial models. Most importantly, don’t be afraid to build an underdeveloped product. You can advance the product in the future, but right now you want to aggregate important feedback that is critical in successfully launching your business.

Jackie De La Rosa is an entrepreneur and cofounder of BeautyTouch, a venture-backed stealth company. Before BeautyTouch, she worked in venture capital for Radical Ventures and managed their portfolio valued at $55 million. She is an avid adviser and mentor for founders and their startups in the greater Boston area.