Validate your assumptions from the very beginning.
Validated learning is the key to progress. And the sooner you incorporate feedback into your development process, the better your product will ultimately be.
One of the core ideas behind the lean startup methodology is that validated learning is the best measurement of progress for new products. It's critical that you learn who your customer is, what they want, whether your product addresses their need, and if the solution commands a price that makes the product profitable.
You don't know what you don't know.
We've been involved in dozens of projects in the past where we were handed a 50 page specification document and given marching orders to build it as specified. It usually takes a couple of conversations to explain why this is a bad idea.
We've learned that on any given project, customers really resonate with one or two features and ignore the rest. The problem is you don't know up front which features will fit into each category.
Building everything in your spec document before showing the application to customers is a sure-fire way to waste a considerable amount of money and time. Nothing is more demoralizing for a client (or your developers) than spending months building an application only to find out that users don't care about 90 percent of it.
A better way.
The good news is that you don't need to build your product before you start getting validated learning. At every step of our product development process, you can build in systems for validating your product and market assumptions.
Validated learning can take a number of forms:
- Problem and solution interviews Before you create a single line of code, you can interview potential customers to see if you're solving a problem that even exists. In the process, you can uncover many things - the features they really care about, the language they use when describing their problem, some early clues on pricing, etc.
- Landing pages As soon as you have a domain name secured, you can put up a splash page with an email signup form. You can then use AdWords to test your headlines and key benefit statements. You can also test feature and benefit statements on the landing page itself, using email signups as a metric. And if you actually get folks signing up to receive updates, all the better - you now have a group of early adopters who've raised their hands, ready to help you refine your offering.
- Testing mockups and prototypes Showing users mockups and HTML prototypes can help you identify whether you're executing on the product vision, both visually and functionally.
- Get feedback from frequent, iterative releases. As you ship features in your Minimum Viable Product, ask your customers how well the interface reflects the vision outlined in your solution interviews. Iterate and improve on those features as many times as necessary to get it right, shipping regularly and asking for feedback each time.
Learn early, learn often.
You don't have to spend months of development and thousands of dollars to learn what your customers want. And you shouldn't. Get in front of your customers from day one and you'll have a more successful project and better results.