« A Democrat's Defense of Romney | Main | Taking People With You - Book Review »

January 28, 2012

Steve Blank vs. Steve Jobs


The Four Steps to the Epiphany: Successful Strategies for Products that Win

I am co-teaching a class at Harvard Business School on entrepreneurship called "Launching Technology Ventures" along with my friend and colleague, Professor Tom Eisenmann.  The class kicked off this week with two cases:  Dropbox and Aardvark.

As I reflect on the class discussions, one of the interesting tension points that arose is the challenge an entrepreneur faces in selecting their primary product design approach.  Should they follow the Steve Blank, Customer Development Process school of product development or the Steve Jobs "vision" school?  In other words, should they pursue a user-centric design paradigm -- setting priorities based on rigorous tests and listening excercises that determine what users want -- or should they pursue a more top-down approach akin to Steve Jobs, who famously said: "It is hard to design by focus groups because most of the time people don't know what they want until you show it to them. "

Steve Blank's book, Four Steps to the Epiphany, has become an instant classic in Start-Up Land for good reason.  Along with the complimentary book by Eric Ries, The Lean StartUp, it provides an incredibly useful guide for starting companies, testing hypotheses and creating products that users love.  Dropbox and Aardvark were terrific first case studies for the HBS students -- both adhered to user-centric design principles quite religiously, but sprinkled a little founder vision in for good measure.

In the case of Dropbox, founder Drew Houston was brilliant in developing an MVP (minimum viable product) that was no more than a simple prototype and then used a rudimentary online video to test user reactions to the prototype.  Houston kept focusing on a test and learn approach to product development, event creating a "Votebox" feature that allowed users to vote for the product changes they wanted most.  But Houston did not strictly follow the Blank/Ries paradigm religously.  For example, after launch, he ignored the most requested feature that users asked for:  enabling the service to synchronize files outside outside the Dropbox folder.  In ignoring his customers' top request, Houston was exerting a Steve Jobs-like, top-down vision in order to stick with the focus on simplicity.

In the case of Aardvark, a social search start-up that was later acquired by Google, co-founder Max Ventilla, was obsessed with following user-centric design principles.  At one point in the case, Ventilla notes:  "We were wary of relying too much on vision and intuition in developing a product."  Yet at the same time, the company refused to provide an archiving capability in the early days of the product, focusing the service on a conversation paradigm rather than Quora's reference paradigm.  Again, the insertion of a Jobs-like product vision.

So in both cases, founders adhered to the Steve Blank school of product design, yet allowed their vision and instincts to overrule user feedback.  What's going on?  When should you choose between the two?  

First, I would observe that the dichotomy may not be as stark as it seems.  Blank is careful to point out in his book that when a company first begins, "there is very limited customer input to a product specification."  Therefore, "start development based on your initial vision."  Yet, in both the Dropbox and Aardvark cases, the founders ignored their customers well into the development cycle.  

I would submit that there are two guiding principles that founders should use when considering overriding their users.  First, when the feedback is in violation of a coherent set of product principles.  In the case of Dropbox, this was an unwavering focus on simplicity.  In the case of Aardvardk, a focus on social search being a conversation.  Second, founders should only have the confidence to develop these principles and  override their users when they possess very strong domain knowledge.  When product-centric founders deeply understand their customer's viewpoint and have tremendous customer empathy, they have the right to make hunch-based product decisions rather than data-driven.

That said, founders should never let themselves off the hook to applying the test and learn principles of Steve Blank to monitor their decisions and continuously validate them.  And the bar should be very high for such overrides.  As the 19th century German philosopher Arthur Schopenhauer observed:  “Talent hits a target no else can hit.  Genius hits a target no one else can see.”

Founders who override their users are betting on genius.  Steve Jobs and Drew Houston have proven that genius pays off.


TrackBack URL for this entry:

Listed below are links to weblogs that reference Steve Blank vs. Steve Jobs:


Feed You can follow this conversation by subscribing to the comment feed for this post.

Interesting commentary... I happen to agree with Steve Blank's approach more than Steve Jobs (it's easier to teach Customer Development as a learned behavior than Steve Jobs personality as a learned behavior). At the end of the day, startup founders are making a set of educated guesses and hoping they move the company in the right direction. For every Dropbox and Aardvark, there are I'm sure plenty of startups that executed in the same way but did not get the same excellent results.

BTW - what happened to the comments on your Romney post?

Very interesting post. I was thinking of suggesting Dropbox on my own blog, as a great tool to further time management. Didn't realize it was a heavyweight in the world of product design!

Thanks, Scott.  I think your point about learned behavior is a good one. As for my Romney post, the comments are there.  Got to the home page - www.SeeingBothSides.com.

[off topic]
Schopenhauer is 19th century...

You note that Steve Blank leaves room for vision in his customer development process. Eric Ries does too. Ries says that a founder should start with a vision for her startup, then translate this vision into falsifiable hypotheses that can be subjected to testing that reflects what users do (i.e., how they respond to MVPs in real world conditions) rather than what they say (i.e., their feature requests). Ries also says that with the lean startup approach, effort is pulled not by users’ stated needs, but rather by the need to test a hypothesis (note the contrast to lean manufacturing, where activity is pulled by a customer’s order).

Viewed in these terms, the issue with Dropbox and Aardvark is not whether they are departing from user-centric design principles by ignoring user feature requests, but rather, whether they have subjected their original vision to rigorous testing. Dropbox’s team has received lots of validation for its vision of simplicity in the form of strong adoption of its products by a range of different types of users—both early adopter/power users and mainstream consumers. It’s less clear whether Aardvark’s team had fully validated their vision of Q&A as private conversation between parties in an extended social network, as opposed to the traditional approach of a query posted to a community where strangers can offer a archived response for all to see.

TypePad HTML Email

Good catch. Born in the 18th but wrote in the 19th. I'll fix.

From what I can tell, Aardvark was more or less a pity acquisition and has since been shutdown by Google.

Dropbox on the other hand has hit it big and will have a >$1B exit.

In your course, did you come up with a sense for how correlated these outcomes are to the principles espoused by their founders? Or did you just assume both companies were equally good outcomes?

I'd double down on Tom's comment. Lean startup is not about "doing what customer tell you" but rather running experiments and reality checking your actions.

Good post.

A similar debate to the one in your post is often cast as "incremental innovation" vs. "disruptive innovation", the idea being that "visionary" Steve Jobs types perform disruptive innovation by ignoring what customers say (examples being the ATM and the iPhone), while others at best can only grind out incremental innovation by listening to customers.

For me, both types of innovation have their place depending on the situation, and the two perspectives are reconciled by distinguishing between the "problem space" and the "solution space", which I've discussed in several talks I've given on Lean Product Management, available at http://www.slideshare.net/dan_o/lean-product-management-for-web-20-products

Most customers can only give you good feedback in the solution space, so asking their opinion about a new product before it's launched can be misleading. Visionaries such as Steve Jobs succeed because they master the problem space; therefore, even without feedback on their products, they are confident that their solutions will do well.

Also along these lines, Greg Cohen at 280 Group has a draft "Lean Product Management" book categorizing different product situations and best practices for tackling each one. It's worth a read, especially since you can download it for free at http://280group.com/blog/?p=1454


The most successful founders see their vision-based decisions in the same light as their customer-driven decisions: as experiments. When founders bet on genius, it doesn't have to be an end-game strategy where to win you have to be right. The question is not whether the source of a decision is founder vision or user testing, but whether the founder is willing to accept that he was wrong and react accordingly.

Drew Houston's decision to ignore power users' requests can be seen as another hypothesis test within his lean approach. He believed complicating the product would make it more difficult for people to adopt it into their lives, and adoption/retention was the whole game for Dropbox at that point. But if Dropbox's analytics (which have the advantage of tracking user behavior rather than articulated requests) showed virality slowing and user retention dropping, I bet Drew would have reconsidered his vision. Continued success is a valid data point, as long as it doesn't blind you to changing conditions and customer behavior. As such, I'm not sure that Drew's decision to keep the product simple is any more "genius" than the decision to abandon SEM because the data showed it to be an ineffective customer acquisition method. It just turned out he was right, which makes it harder to scrutinize--and attribute causality to--his process.

As with Dropbox, Aardvark's success or failure stemmed from the desire and ability of users to adopt and use the product. The team had great data, so they knew things were not going as planned, but amid their efforts to change the trend the "conversation" paradigm was a sacred cow. It's not clear that archiving was the answer, but faced with flagging adoption rates (in contrast to Dropbox) Max and his team probably should have given it a shot. I think Drew Houston would have.

(Full disclosure to the community: I'm a student in the LTV class.)

TypePad HTML Email
Framing the decisions around experiments is dead on, Matt.  And your contrast between the two is interesting.  I wonder if a founder can improve their odds of conducting experiments that are going to turn out right?   

TypePad HTML Email

Great point, John. We did explore this question deeply. A few observations I shared with the class, from my point of view: 1) Dropbox tested its initial hypotheses rigorously, particularly the fundamental ones, Aardvark less so. 2) Aardvark didn't identify a "must have" pain point, but rather a "nice to have" novelty. 3) Dropbox religiously designed the service to be inherently viral, Aardvark never managed to hit true virality.

Jeff - maybe it's just me, but I cannot see the Romney comments - in both Chrome and Firefox - even if I hit the Comments link (and I see there are 8 comments), it brings up your post with no comments.... not a big deal, but just FYI.

Just curious about exactly how much presumably insanely accurate consumer studies and data the Apple guys have when designing (i.e. hardly designing in the dark) vs our lean startups and whether their design processes have something that differentiates them ?

Debate not so novel if we stick to the incremental vs disruptive line (back to henry ford's line about people wanting faster horses then :-)) --> do we know more about the inside working of Apple to dissect this further than the "Jobs Exception" ?

TypePad HTML Email
Great question, Fred.  I don’t have an inside view here, but we have a few students in the class from Apple so I will push them on this.  And with the end of the Jobs Era, I wonder if they aren’t becoming much more data-driven.    

4 Steps and www.custdev.com both spoke loud and clear to me, I feel like I have been often told to go forth and sell, when the founder or CEO really means, please go find us some customers. Typically that involves a very consultative sales process with an early stage company or product, folks that didn't have a process or experience don't know the cost of the pain they are trying to solve and therefore don't really know the value of the solution. That said, I think that is isn't so much Blank v. Jobs, but how to know more about the user than the user does him(her)self. This might have been part of Jobs magic, know what the user want before they know it themselves. Assuming most of us don't have Job's genius then we need to fall back on user-centric design and persona to help drive the product development cycle. Oh yeah, and don't forget to talk to 30 or 40 prospective customers or buyers. "Talk" meaning, don't pitch, show demos, wire-frames, screenshots or rest, just ask them about their problems or pain points. What would a good solution look like, would they be willing to use or even buy a solution like that...

Jakob http://www.useit.com/ and the usability guys get a lot of it right: you don't always give the user what they want, sometimes we should to give them what they need.

Very true. Also, a good online presence counts a lot since it makes advertising easier. Your exposure to potential customers is also higher as the world wide web continue to expand. The possibilities of your small/starting/growing business are endless.

Aardvark is often cited as a shining example of lean startup principles as well. I've seen the founders talk about the subject more than once and the focus is always on the fact that they did not build much actual product in the beginning. Instead, as a method of doing valuable customer development and avoiding building less than viable product, they kept the service's technical components to a bare minimum and handled a great deal of the work by hand.

As a case study, Aardvark may not be the best way to determine whether there is validity in the approach. But, there are certainly lessons to be learned.

The comments to this entry are closed.

Social Media

Portfolio Companies