Creating a new feature or product that results in deep customer engagement is hard to do!

Jared Spool (a thought leader on UX and user research) explains why in his tweet…

Testing usability is a known and well documented process. But are there ways to test demand for a new product or feature before you build it?

Yes! So long as you do it properly.

Here’s how..


Asking “Would you?” is NOT Enough

Let’s start by highlighting a common testing mistake.

Unfortunately, asking a test respondent if he would use a product does not produce a reliable response. Jared’s above tweet partially explains the reason why. But the underlying reason that this question is an unreliable predictor of demand is that there is no cost to the test respondent to saying “yes”.


You MUST Make “Yes” Cost Something:

In normal, non-testing life, saying “yes” is a tradeoff decision. To say “yes” we have to first determine, “Is this worth it to me given my constraints?” Since this is the same question that your users will ask themselves when your product (or feature) is released, you need to force that decision during testing too. Tweet this


How to make “Yes” Cost Something:

There are three ways (in escalating order) to make saying “yes” to a qualitative question cost something during a test.

1. Time. Ask the test respondent to give you more of her time. If she is truly interested in the product (or feature), she will say “yes!” Here is an example question:

Time question

2. Reputation. Ask the test respondent to share your product amongst his friends/peers without compensation? If she is truly excited about your product (or feature), then she will gladly share with others. Here is an example:

Reputation question

3. Money. Ask the test respondent to pre-pay for the product (or feature)? Even charging a nominal amount is telling. Here is an example question:

What these tests tell you:

To be clear, this is NOT a 100% guarantee of building a product (or feature) that customers use heavily. After all, I am sure that you have downloaded apps that you don’t use. Or canceled a pre-paid hotel reservation. But these tests are a great way to avoid a clearly bad product/feature implementation. Simply put, if a test respondent is unwilling to give you her time, reputation or money, then she doesn’t really want to use your product. Tweet this


Past Behavior is a Good Indicator Too.

Another good validation of future demand is whether a test respondent has used a comparable product/feature in the past. Jared suggests this question:

Past behavior tweet

The goal of this question is to better understand why a user “rented” a competing product. The “tell me about it” part is trying to dig into what would make them switch to your product. What do they like or hate about the other product? Jared’s point is that you are unlikely to convince someone who doesn’t currently do an activity to start doing it. But you can cause them to switch from another product if your product satisfies an unmet need. Tweet this

And of course, you want to use Minimalist testing for these tests to validate what feature set a customer is saying “yes” to.