One of the most exciting and frightening things in the business world  is implementing new ideas. You can crunch numbers, consult experts and run simulations forever, but you can never be sure if new ideas are going to work in your business until you try them.

So then, what if your latest brainstorm turns out to be a bad idea? What if you put everything behind it, and it fails? That’s always a possibility. But the thing to do with any new idea is to start small, and limit the risk. There are a few ways you can do that.

Pilot programs.

A pilot program is a short term implementation of your idea on a small scale, to test out the reaction and the results. For instance, say you want to try out a new algorithm for recommending products to your customers based on what they’ve bought in the past, or what they’ve searched for or expressed interest in on your site. Amazon does it. Maybe it’ll work for you, too.

So the thing to do is take a small, representative cross-section of your e-mail list, and target them first. Spend a few months sending them the recommendation emails before you launch them to everyone, and gauge their reaction. This way, you can see how many people open the email, how many click through, and how many actually make a purchase.

You can do the same thing with landing pages. “Create private landing pages that are only provided to a select few. They can’t be searched for, they can’t be reached through general browsing of your website… They’re just a test to see who clicks on them and who doesn’t, in your test group,” says Anthony Trapani, Marketing Director of Palm Springs Golf Vacations.

Once you have this information, you can determine two things:

Is this something that people will even want?

If you get a tepid reaction when you release your idea, then perhaps it’s something that your target audience just isn’t interested in. Maybe it worked for other companies in the past, but they aren’t you, and their customers aren’t your customers. Better you find out at the beginning, so you can just cut the program before it causes too much time and expense.

Is this the best way to give it to them?

Maybe your initial emails did get a strong reaction from your target audience. It looks like your idea is something they’re interested in! But then there’s very little follow-through. Interest quickly drops off, and there’s very little conversion to actual sales.

It may be that your idea still has some bugs in it. People like it, but you’re not presenting it to them in the best way. To take our product recommendation example from earlier, maybe your algorithms are off. People are interested in more products from your company, but the ones you’re recommending to them aren’t in line with what they want. No worries. Root out the problem, figure out how to fix it, try again. When you’re ready to launch on a bigger scale, you’ll know that the idea works the way it’s supposed to.

Always be testing.

You can test any idea, big or small, in this way. Say you want to experiment with targeted ads or promoted social media posts. You want to know what kind of ads your customers will respond to, and what they should look like. A failed ad campaign can cost you tons of money and turn your business into a cautionary tale. So try each new ad campaign or venue on a small scale first, with a limited budget, before releasing it to the public.

No matter how good an idea is on paper, you never really know how it’s going to turn out until you try it. After all, business is all about risk, and if you’re not willing to take a few risks, you’ll never succeed. But just because you’re trying out a new idea doesn’t mean you have to jump headfirst into the deep end. Test the waters first, get your feet wet, and build up to it. That way, when the time does come to go into the deep end, you can be sure of making a big splash, instead of just sinking to the bottom.