I think the best approach is just to name your personal biases and be aware of them. I don't believe it is possible to Go Clear and remove all biases by eliminating the Reactive Mind. And it's good to surround yourself with folks who have slightly different but compatible biases. Which is a good non-altruistic argument for heterogeneity.

I like to build products that I want to exist in the world—first, for myself. I've failed when I tried to build stuff that I thought "the market" would want. Maybe that's cuz I'm bad at user research.

But more seriously, I think most people are bad at user research. And the beauty of the Internet at this point is that, unless you're *really* weird, there a still lots and lots of people like you.

So, I'm biased toward using one's personal bias as much as possible. I really think that's where the best stuff comes from.

Lastly, I'll add the caveat that when it comes to growth and new-user experience sort of things, it pays to be much more scientific—since you can't recapture what it's like to be a new user of your own product.

I would argue that building for yourself is a luxury that exists only in the earliest days of a consumer based product. After the initial idea don't you start building for users more than yourself?

Having said that... I also believe that that the best products/companies have this common trait...

The CEO is the most demanding customer of their own product.

For me it's been a phased approach. You build for yourself and then realize there are different ways people use your product as it grows. At that point, you (and your team) still possess the 'true north' of the product, but you're not building just for your own person specific use cases.

As Goldman points out, i think this is one reason that diverse teams are important in consumer products.

Instead of 'personal bias' you might say 'use case understood very well' and the potential negatives of your proposition disappear.

I dont see the bias. Bias suggests you need objectivity when creating something -- but creating something new is by its nature subjective. Once something has been created, once it is formed then objectivity -- data driven or qualitative objectivity is critical. But doing market research on a pile of wet clay is a waste of time -- the clay has to be formed, shaped first. The market is an average, a statistical average -- sampling it prematurely wont yield something new. Shape the clay first and then use data to refine, adjust, listen, pivot. etc.

I would replace the word "biases" with "instincts". A big part of what makes the greatest products great products is that the creators of them have unique instincts about what will make a product special and that those characteristics will appeal to the broadest possible set of users. I think being able to read data when something is working and isn't working (as in case of Odeo or Instagram) still relies on founder instincts. Per @ev's point about growth being more scientific, I think that is right at the optimization level, but still requires a lot of instinct and biases to come up with the right ways to introduce a product to users. Does instincts sound better than biases? Does to me...