I’ve got this idea for a new word processing client. It’s going to sync between mobile and desktop, and be totally awesome. – Graeme, of Auckland and Wellington

Firstly, and this applies to any new idea, stop and do a competitive scan and check what is out there.

In your case you’ll be joining the queue behind some very well-funded and other well loved products. Some cool cats have just released Quip.com, Google Docs has been around for ages, while Scrivener is a great tool for writers (except the iPad version is taking its time) and then there are niche products like OB3 for medical education and Workflowy for notetaking.

Once you’ve done the scan for competitors, you will start to see gaps. Don’t stint on the Googling – denial of your competition is not a great way to embark on a business. Pick the top opportunities and move on.

It’s too soon to start developing yet, though – now is the time to meet the end users and understand their world. Interview them, hang out with them and observe what they do.

Lawyers, for example, use document change tracking and version comparing in Word extensively for communication between internal and external lawyers and their clients. Designers, on the other hand, like laying up their text into documents page by page, paragraph by paragraph or even letter by letter with kerning.

Now imagine, as I’ve recently experienced, trying to link the two together, with a dynamically changing change-tracked legal document, wrestling with designers wanting consistent text.

We know that the best designs are done with real text, and that the best text is integrated and informed by the design, and so the current process is broken. 

The next step is to quickly prototype a solution and then test it. The prototype can be a model, a physical model or on paper even, and the testing can be as simple as a role-play with each other.

The more models and tests you do, the more quickly you will iterate to a good answer. Once you have something that tests well then just build a very basic product that performs the key functionality and test it again with end users.

Ideally, even if you are not an expert developer or builder, you will do this yourself, and only after it tests well will you need to look for more developers and designers, and ultimately, some funding.