You can't build a company all by yourself, much as you might wish you could. Cloning's not really an option yet, either. That means most entrepreneurs will eventually turn to outsourcing.

Which, as we all know, can be a train wreck. That's why a recent presentation at the Columbia Startup Lab, by serial entrepreneurs Laura Smoliar and Ted Hou, was so valuable. In a 40-minute talk, they presented their six-step process for outsourcing without going insane or losing your shirt. (The official title was, "Outsourcing: Managing in the Discomfort Zone.")

The duo put a particular emphasis on protecting your intellectual property even as you rely on others to build your product. They broke the process down into six steps:

1. Define the product

This sounds straightforward enough. But Hou used the example of his own fledgling wearables company, Neem Scientific, to show that defining your product can get complicated quickly. At first glance, his product would seem to be a device. But the product could also be the data collected by the device. If Neem were to develop a platform to help users make better health decisions, that would be a different product altogether. So the first step is to define your product, and, by implication, the business you're in.

2. Map your product to a modular design

Break down your product into its component parts, and start to figure out which companies should make each one. By breaking up production in this way, you ensure that no single vendor knows everything about your product. Instead, each vendor gets the necessary information about the particular part they are building, as well as a solid understanding of the interfaces that connect their piece to the others. If you have modularized things properly, this will work.

3. Decide what stays in-house, and what you can outsource

Anything involving your core competency should stay in-house. The rest of it can be outsourced--although Smoliar and Hou joked that no one in Silicon Valley says "outsource" anymore, because it has too many negative connotations. Instead, you "create an ecosystem around your company."

4. Find your outsourcers (or, ahem, identify your ecosystem)

This means building out your network, and just as importantly, using it. There are certain vendors that are well-known and used by a wide variety of startups, including Engine Room, which can handle finance, accounting, information technology, human resources, and facilities management; High Tech Connect, which handles marketing and communications; and SoloPoint Solutions, which provides as-needed engineering talent.

If she doesn't have expertise in a certain area, Smoliar says she's occasionally asked her advisors to interview vendors for her.

5. Write out a statement of work, or SOW

The statement of work (sometimes, scope of work), is a document that sets out what you expect your new partner to do, when it will be completed, and how much it's going to cost. Basically, you need to get agreement on six things:

  • Specifications, or, on a high level, exactly what you're asking your partner to accomplish for you

  • Timeline, or when certain pieces of the project will be completed

  • Milestones, or what goals need to be achieved before a portion of the product is considered finished.

  • Deliverables, or the chunks of work your vendor will provide you with. You need to pay particular attention to these, because this is where arguments over money frequently begin.

  • Frequency of regular meetings, or how you'll keep things on track. This is extremely important. If you have a weekly call scheduled with your vendor, it makes it easier for them to let you know when there's a problem. They don't have to worry about whether or not it's big enough to bother you with--they can just tell you on your regular call. Engineers, especially, tend to go quiet when things head south, Smoliar says. Sometimes there's a technical problem. But maybe they just took a full-time job, and they thought they could do your job and the day job, but it's not working out and they don't know what to say. A regularly-scheduled call makes it more likely that they'll tell you what's up. 

  • Payment terms, or when and how you need to pay. Generally, there are two ways you can pay contractors: a fixed bid for a given project, or "time and materials," which is an hourly rate plus expenses. Time and materials is often best for initial exploratory work, but can balloon if not managed carefully. Mostly, Smoliar prefers fixed bid payments, especially if you're managing work outside of your area of expertise. A fixed bid is also more likely to force the difficult-but-necessary conversations about milestones and deliverables. Even with a fixed bid, costs can shift. Sometimes there will be situations in which you'll need to sit down the contractor and negotiate a change to the SOW. It often makes sense not to negotiate too hard on price, because you probably won't save that much, and what you really need are iterations. Instead of trying to knock the price down 10 percent, ask for five iterations rather than three.

A lot of entrepreneurs want to skip these steps and get straight to work, says Smoliar. Bad idea. If you can't get through this process, it's a huge red flag.

Post mortem

How'd it go? Once the work has been completed, you want to sit down with your staff and your vendor and talk about what worked, what didn't, and what you can improve upon next time.