7 years of teaching lean startup!
As Many of you know, I’ve spent a fair amount of my time since leaving Microsoft teaching at Northwestern University in Chicago. I taught software engineering my first year; and then at the urging of the dean of engineering, Julio Ottino, I worked with the director of the Farley Center for Entrepreneurship to develop the NUvention Web course. Our mission was to create a multi-disciplinary course, that mixed grads and undergrads from different schools who would work on teams to define and develop a software product concept over two quarters. We’ve had over 300 people take the course; and while not an explicit goal, the course has spun out a few startups over the years. Most notably, Adaptly, started by Nikhil Sethi in our inaugural year; but also Groovebug, Tradeup.io, Waddle, and Sweetperk—all of which raised rounds of funding or attended good accelerator programs post class. More importantly, the feedback we have is that the experience had significant impact on how they approached their careers and helped them establish initial mentor relationships with each other and with alums and our advisory board that have persisted at whatever ventures students did next. I’m teaching again this year along with Mike Marasco, Chris Riesbeck, Rich Gordon and Rich Padula. A few years ago we added Medill journalism graduate students to the mix to think about content oriented startups (and content marketing) and renamed the classs NUvention Web+Media. For this year, we have done a complete revision of the curriculum.
How we taught the course in the past
Our initial curriculum was a mash up of Steve Blank’s The Four Steps to the Epiphany (recommended by me when doing customer development for the class with a then new alum Alex White who founded Next Big Sound) and a software engineering curriculum Chris Riesbeck and I put together. Entrepreneurship education has evolved rapidly in the last 7 years. The most significant event being the establishment of NSF iCorp and the leanlaunchpad curriculim supported by VentureWell, which was developed by Blank and Jerry Engel; supported by a good udacity course, Launchpad central software, and the excellent Startup Owner’s Manual and Business Model Generation: A Handbook for Visionaries, Game Changers, and Challengers. We have taught the course using the straight leanlaunchpad method, after experimenting with approaches like Ash Maurya’s Running Lean.
Starting immediately after our 6th class ended last June, we began an overhaul of the class based on lessons from those first six years. We consulted with Blank and Engel, where we shared some of our reactions in our experience of teaching the course and using some of the leanlaunchlab material. Steve rightly pointed out that because we have the luxury of two quarters; we go a bit beyond customer development; in fact in recent years, most teams have users they are serving and even actual revenue with their products by the time the class ends. Also, critique hasn’t worked as well for us as it has for Blank and Engel. It’s a very important part of the course, still the center of what we do, but for us we decided we want to use more tools in the teaching toolbox.
NUvention Web 2016 Curriculum
As we put the curriculum together, we sought to address a number of goals:
- Better ideation more grounded in user needs. An important lesson for our students is going from the notion of their startup dream, which is usually grounded in a solution, to identifying specific customer audiences of sufficient size and identifying the high value problems that the team can address.
- Increase the software engineering depth as well as the business customer development depth. CS students who take our class can do so as part of a software project requirement. Our goal is to have them exit the class with grounded experience in architecting and writing software as a team. We also wanted to develop a better community of developers who could help each other. We were doing ok on this; but felt we could do better.
- Increase class engagement. This was to address issues we found with class wide critique mentioned above as well as working with the multiple disciplines. Our feedback was that business people to a certain extent would tune-out some of the software development topics; and hackers would tune out the business pieces. The benefit of the class is clearly exposure to both working in their team. We thought we could do better.
We also wanted to build on many aspects of the class that worked:
- Pushing teams to launch products to real customers before the end of the class
- Using four panel storyboards and scenario canvases to drive MVP development and backlog prioritization
- Using the business model canvas as a framework for thinking through the business.
- Linking to an advisory board of entrepreneurial practitioners to strengthen student networks and provide a source of ongoing feedback and coaching beyond the teaching team.
- Pushing teams to deliver from the first class to accelerate learning and team bonds.
- Using participatory teaching techniques that drive class engagement.
Changes from leanlaunchpad model
Teaching technique: call and response
We have been moving to a situation where there is more ‘call and response’, something we stole from Chris Riesbeck. For classrooms with good internet connections (note: this method did not work for me when at Ashesi because at the time connection and reliability of the internet was too unpredictable), googledocs live editing is a revalation. Essentially, we introduce a concept, provide a google docs template for teams to explore it specific to their business/product, have them work on it in a single consolidate deck (students learn from seeing live what other teams are doing); and then doing spot follow up when complete. Time wise, this takes about 4x the time of just presenting the material but I would assert it is also more than 4x effective. It’s also less time than a traditional team by team critique.
Critque Optimization: Split class
If you have more than 6 teams in a class, a combined critique consumes most of a three-hour class. Class wide critique has value for sure; but we found doing this every week gave us lower marginal return over time. We are blessed with the luxury of a large faculty team (4) and often have advisory board guests that help out from time to time. If you have 4 faculty and 10 teams, you can split into two rooms and cut elapsed time in half. Mixing combinations of teams helps broaden team learning and in our experience keeps it fresh. Lessons learned across rooms can be shared once the class is brought together.
The new NUvention class now has four five week phases:
Using Dan Olsen’s product market fit pyramid: From customer needs to value proposition
This summer like usual I searched for techniques to help us fill the identified gaps in the class. I came across Dan Olsen’s The Lean Product Playbook: How to Innovate with Minimum Viable Products and Rapid Customer Feedback. The book is an excellent guide for product managers in organizations of all sizes on developing and testing your value proposition. We connected with Dan (He’s a Northwestern Alum which was a bonus), and added his book to our reading list. He’s going to come talk in our class and join the advisory board. We are centering the early phase of the class around the bottom of the period that addresses identification of a customer segment and a set of user needs. This gets teams focused on identifying a key need/pain point. From there we focus teams on developing a scenario canvas that lets focuses the teams on developing their initial product slice as part of development of their Minimum Viable Product (MVP). Then students in parallel flesh out key personas and work on sizing the market.
Once these steps are complete, we move to creating a hypothesis business model canvas and analyzing the competitive landscape. Teams iterate over this concept while developing the MVP in the second half of winter and the first half of spring; at which point they launch their product into their initial market.
Separate customer development and software engineering tracks
We usually open our 3-hour class with either a critique or a specific topic we cover using another teaching method. This year, we decided that sometimes we would split out the ‘product development’ and ‘customer development’ groups. We did this to cover some topics in more depth. Many are meant to then be taught by student peers to the other parts of the team. On the product development side, it lets us go more in depth technically without losing the rest of the class. For example, on Tuesday we went into more depth on the pros and cons of different starting architectures for web apps in a startup. In the customer development section, we talked in more detail about how to interview. We used some material from Running Lean in addition to our own experience,
The class is a 3 hour block of time. We work to give each team some time that they can use to work through the team issues while the faculty wander between the tables to give feedback and offer assistance. Another terrific thing that’s happened since we started in the creation of the NU garage space. This gives our teams a meeting place they can use that is in a single open layout where we can interact with each of the teams.
Our first session went well, and I’m looking forward to seeing how the teams evolve.