@AndrewChen

(if you don’t see the slides, go here to Slideshare)

Recent slides for a talk in Steve Blank / Eric Ries’s class on High-Tech Entrepreneurship
Yesterday I had the pleasure of giving a talk at Steve and Eric‘s class at Haas on the topic of Minimum Desirable Product – if you haven’t read the original article, it provides some useful context. I included an set of slides above on the topic, updated from my talk yesterday, which you can peruse at your convenience.

After you’re done, you can read my extended remarks below on some stuff I learned along the way. Frankly, any of these could probably be its own blog post but I’ve been feeling lazy lately so you get a couple sentences apiece instead

“Viable” means different things to different people – my usage is meant to be pretty specific
Eric noted during my talk that I use a very narrow definition of “viable” within Minimum Viable Product, which is true. I believe in his usage of it, the focus on viability is actually a conglomeration of IDEO’s concept of desirability, feasibility, and viability. It’s frankly a coincidence that IDEO and the Lean Startup use a common term, though I believe they mostly overlap. I prefer IDEO’s framework because it allows a bit more precision in describing the class of issues you’re concerned about, but frankly there’s a ton of gray area. (Is a low-priced X a desirability thing or a viability thing? Honestly, both.)

Viability-first strategies do work, and may be the right thing for you
Many companies have come and gone that make products that aren’t that great, don’t generate a lot of consumer value, and yet still pull in a lot of money. It’s a strategy that can work, and I’m not arguing the opposite. However, I’m convinced that if your goal is to make a mainstream web property that has daily engagement, starting with the goal of creating lots of user value is probably the way to go. Similarly, if you have a highly transactional business like ecommerce, designing for daily engagement is probably overkill – in that case, reducing your cost of customer acquisition might be the right way to go. So it’s all very situational, and frankly, very personal based on how you want to run your product.

Minimum Desirable Product is just a starting point – you still need to figure everything else out
I also want to note that my message isn’t just to build for any random group of users and then the rest will take care of itself. That’s far too idealistic. Instead, it’s just a starting point for how you think of the problem. Ultimately, all your product ideas still need to be filtered through the lens of whether you can market them, that the market is big enough, and that the technology issues aren’t insurmountable. There was a recent Times interview with Steve Jobs on the iPad that illustrates this perspective:

… surely Apple stands at the intersection of liberal arts, technology and commerce? “Sure, what we do has to make commercial sense,” Jobs concedes, “but it’s never the starting point. We start with the product and the user experience.”

Metrics can be oriented towards user value
I’ve written before on some of the short-comings of using metrics-driven product strategies, such as here and here. An analytics dashboard is ultimately just one tool out of many that help you optimize whatever goal you want to set. If you are very focused on validating your business model and spend all your time tracking metrics such as viral factor, ARPU and conversation rates, then you will make those go higher. If you use your metrics to define user benefits and optimize those (I’ve begun calling this “Metrics of Love”) then you’ll make your value proposition go higher. So depending on your perspective and where you want to start, you’ll end up in different places.

Highly desirable consumer products also have minimalist featuresets
In consumer products, unlike some enterprise products, there’s a big focus on simplicity and immediate value. In some ways, the idea of a “minimum desirable product” is kind of misleading because highly desirable products may also have minimum featuresets also, perhaps even more minimal than an MDP. The important part is that they are the right features, and in fact, it often takes a longer time to simplify your product and boil it down to the core value. I think that’s an interesting paradox that exists in consumer products, and one that I didn’t grasp for a long time.

Learning about your business and learning about your product desirability are different things
One of the interesting points that came up yesterday was that if you view your company as a learning machine to validate your business before you run out of money, then you may see that worldview clash with wanting to deliver maximum product desirability. In many cases, shipping a 50% done feature may teach you a ton about the market, and very quickly you will learn what you need and want to move on. The problem is, it may turn out that going from 50% to 100% in user experience actually continues to increase value to the user, by making things more refined and more compelling, even if you stop learning about your business. This is a hard thing to trade off, and requires situational judgement. As Steve noted during yesterday’s discussion, deciding when you stop and just consolidate and refine what you have, versus packing in new features – well that’s the place where entrepreneurship is an art and not a science

OK! Back to blogging vacation See you guys later.