Eric Reis on Why Continuous Deployment?

Under continuous deployment, as soon as code is written, it’s on its way to production. That means we are often deploying just 1% of a feature – long before customers would want to see it. In fact, most of the work involved with a new feature is not the user-visible parts of the feature itself. Instead, it’s the millions of tiny touch points that integrate the feature with all the other features that were built before.

This seems to be the key to continuous deployment.  You can still make huge changes while working in small batches by flipping a bit when you are ready to deploy the UI changes.

Eric Ries does a fantastic job of sharing his experiences on his blog: Startup Lessons Learned.  Following his example, I'm going to start regularly posting a "Lesson Learned".  At the advice of Hiten Shah, I'm going to structure the lesson into a template that others can follow to avoid re-inventing the wheel.

Hypothesis:

People are more comfortable submitting information to a website that has a smiling face on it.

Rationale:

Websites frequently use stock photos to prevent bounces and increase sales.  People are more likely to trust a person with their email address than a faceless website.

Methodology:

I split tested two versions of the landing page for Lean Startup Circle, one that included my bio with a picture, and one that did not.  To encourage traffic, I posted a tweet that linked to the landing page.

Results:

Within 24 hours, 274 unique people had visited the website.  Of those, 151 were presented the landing page absent the bio.  The other 123 were presented the landing page with the bio.  46% of the people that viewed the bio-less page submitted their email address.  Only 35% of the people that were presented the bio submitted their email address.

Bio w/ Portrait 35%

Bio w/o Portrait 46%

Analysis:

We can say with over 99.7% confidence that people presented the landing page with my bio are less likely to submit their email address than people who aren't.  Some possible explanations:

• The bio was too detailed and removed focus from the proposal.
• Smiling faces work well for some audiences but not others.
• A friend suggested that people might have assumed that the site was about me, rather than about something relevant to them.
• A more handsome picture (highly unlikely) would have performed better.

Lesson Learned:

The assumption that faces improve conversion isn't always valid.

I encourage you post your own "Lesson Learned".  The faster we learn, the faster we will create things that people love.

Discuss At the Lean Startup Circle

The minimum viable product is one of the more interesting concepts I've come across in the Lean Startup community.  I've had some lively discussions with my co-founders about what would constitute the minimum viable product for Stylous.  Our vision for Stylous is to create a product browsing experience that lets you focus on the design of the products without being distracted by irrelevant information.  Against conventional wisdom, we've removed prices, buy buttons, brand names and almost anything not related to the actual look of the products.

Instead of building and testing something that meets only these goals, we didn't launch until we additionally had:

• 25 categories of products
• Price information
• Ability to favorite products
• Tracking of previously viewed products

Before we had any promising results we also added:

• Filtering of products based on brand, price, sale price, color, store and previously seen status
• An innovative navigation UI

We added these features because we felt that the site wouldn't be very useful without them.  All of this took months to build.  As a result, I often wonder what form the MVP could have taken for Stylous.  Would one category of aggregated products  with very simple navigation have been useful to people?  Would there have been any evidence that they loved the site enough to tell other people about it?

I find the concept of the minimum viable product to be very intriguing as a way to limit the risk of spending time building something no one wants.  In practice, it seems very difficult to decide what form a good MVP will take.

I would like to hear about other MVPs that people have built.  If you haven't created one, then how would the MVP look for your product?

Join the Conversation