23 August 2015

One in what will hopefully become a moderate size collection of reviews for books, videos, courses, and other reviewable things.

UX for Lean Startups by Laura Klein

This book was published in 2013 by O'Reilly.

This is an excellent reference, and a nice introduction to the Lean topic as well as UX. The text provides a nice introduction to User Experience from a Lean perspective. Extending from Eric Ries' Lean concepts of iterating, GOOTB, hypothesis testing, and other techniques, this book focuses on how to apply the concepts more specifically from those general principles. In addition to advocating for use of these techniques, each one comes with caveats and cautionary tails of what can happen if the “big picture” is not considered and too much focus is placed on a single technique.

The book is an easy read for someone with an engineering background (such as myself). It presents the design, product, and business topics at a level where they can be understood without needing to have decades of experience in the techniques used to gather user inputs and develop user validation testing.

While the title appears to restrict the audience to startup companies, this is not true. Existing organizations of any size can benefit from using the principles and techniques here. They are used by many startups since they give the greatest chance of success using the least amount of resources, which is ideal for a startup. But any organization can benefit from lower cost, shorter time frame developments. Startups also have less organizational baggage to content with. To use some of these techniques in large existing organizations may take some skill to separate the new activities from the existing processes. But structuring a new process does not need to impact existing processes, so it is still possible to apply these techniques.

The book is organized in three major parts: Validation, Design, and Product. The majority of the text is in the Design section with practical advice on how to apply Lean principles in practice. The initial Validation section provides a solid grounding in the reasons for using the techniques, and the Product section leaves us with several points of advice on how to manage products that have been fielded using the same short cycle iterations and testing that were used to create the product in the first place.

This book serves well as a reference — the book does not need to be read in order. Each chapter, even sections, can stand alone and do provide some practical advise. So if you have an immediate need, go to that section and start working away.

While this includes a very good treatment of Minimum Viable Product (MVP) and what that really means, it was the combination of the descriptions in this book along with those in Jeff Patton's User Story Mapping book that finally allowed me to grok that concept well enough to be a useful one. Hopefully I can add a review of Jeff's book here soon as well.

As presented here, the topic of User Experience is described as a combination of several elements. These include:

Validating Hypotheses

Hypothesis validation is central to UX design. It means that you should start by breaking your assumptions and product into small, individually testable, hypotheses. Then validate each hypothesis in as short a time (and as small a cost) as possible.

Hypotheses replace features in the traditional product development flow. Instead of determining what features should be built, then building them, using hypothesis testing allows us to determine if a feature is even required or might be counterproductive.

Hypothesis testing uses the Lean cycle advocated by Eric Ries that includes: ideas (hypothesis) that are then built, followed by a (minimum viable) product which is then measured, and the gathered data is then used to learn about the original idea (hypothesis). This is a cycle that is never ending, and by keeping the individual hypothesis small, we can learn about – and pivot away from if necessary – what is useful and what is not useful in the end product.

User Centric

This typically means prototyping and talking to users. And indeed it still means that here. It adds to these product feedback gathering mechanisms. Lean UX also includes the iterating, agile, hypothesis testing mechanisms of the entire UX design cycle. But using the existing product feedback techniques is still part of the overall flow.


Yep. Lose the waterfall. Include a cross-functional team in the development, time box it, and use modern Agile development processes. There are some additional considerations to include here, but the existing Agile process is used to good effect.

Data Driven

The Lean build/measure/learn cycle includes “measure”. This means that we need to include data generation and collection into the process. We also need to include analysis of the data that is collected. If we simply collect data and generate reports, we haven't really ensured that we will benefit. The ability to select what data is important and to explore and understand the results to gather meaningful information is central to UX design. This means combining good design with good testing, and then gathering and analyzing the results. If we do not learn from a hypothesis, then we have not completed an iteration.

Fast (and Cheap)

Lean does not mean “as cheap as possible”. However, the use of fast iterations and the efficiency of hypothesis testing along with the agile development process often lead to lower overall costs for product development. Indeed, by providing the ability to identify and avoid features and elements that will not improve the user experience early, no investment in these features needs to be made.

Even so, good user experience design and product development is not cheap. The benefit of Lean UX is that it keeps the costs low while still providing for the best user experience and development of only those features that are needed.


Being iterative, with all of the above attributes (testing, user centric, agile, data driven, and fast) is the core of UX design. This is realized in the creation of a Minimum Viable Product (MVP). This is the smallest complete product that can be made. Iteration is what will move the initial MVP to the final product.

It is important to realize that a MVP is not simply created once and discarded. It is continually used as the platform to validate the next hypothesis on and to converge on a better product. You must keep iterating and building on what you already have. Keep trying new things, retain the ones that are good and discard the rest. Don't forget to retain the features that work. Sometimes if many hypotheses are tested at once from the same point, it is easy to forget that more than one of the hypotheses was a good one and needs to be retained.

As you iterate and improve the product, you should be retaining the best and discarding the rest. You should be improving the product and not simply testing and adding features. The user's experience is what you want to improve, after all.

blog comments powered by Disqus