Lessons Learned Launching a Lean Startup Accelerator

Year One Labs is a Lean Startup accelerator. I believe it’s one of the first of its kind, with a very heavy focus on customer development and a rigorous Lean Startup process. I co-founded Year One Labs with three partners (Raymond Luk, Alistair Croll and Ian Rae) in April 2010 and launched officially in September of that year.

Since our launch we’ve co-created and invested in 5 startups (which was our target number). We now work with those startups on a daily basis; they have up to 12 months in Year One Labs to get to product-market fit (or as close to it as possible) and move on to raise additional capital, become self-sufficient, or potentially fail.

We’ve learned a lot launching a startup accelerator, and particularly one focused on a Lean Startup methodology. Year One Labs is a startup in and of itself, one that went through many of the customer development and lean startup cycles that we now put our startups through.

Iterating our Own Model

When putting together the first ideas for Year One Labs we had our own hypotheses around what would work and what wouldn’t, and iterated our way to a model that we were willing to go with and test. Our ultimate goal (and hypothesis) is that we can provide meaningful financial returns to our investors (including ourselves.) But there are many steps and hypotheses throughout the process. Here are some examples:

  • Hypothesis #1: We’ll attract very young, inexperienced entrepreneurs (think: just out of university). We assumed this would be the case because our model ($50,000 for 12 months inside Year One Labs) would lend itself to people with minimal expenses, just starting out in their careers.
  • Result #1: We didn’t limit our recruiting efforts to this target audience, but we did make efforts to reach them as much as possible. In the end we found that this demographic was not the most attracted to Year One Labs. I think this has a lot to do with the lack of entrepreneurship being fostered in Canadian universities.
  • Hypothesis #2: It will take startups ~12 months to “graduate” out of Year One Labs. “Graduation” in this case means they’re self-sufficient (with revenue) or they’re raising additional capital and scaling their businesses.
  • Result #2: None of our startups have graduated yet, but some of them are very close. And as we work with them and track progress, it becomes clear that they don’t need 12 months to get to the level of traction we’re looking for. It’s closer to 6 months. This is a significant data point for us because it impacts how we proceed with some of the newer companies in Year One Labs.
  • Hypothesis #3: Startups will come into Year One Labs with a fully baked team and fully baked idea.
  • Result #3: We invested in most of the startups at the “pre-concrete” idea stage. In some cases we put teams together, finding individuals that we liked and matching them up. The realization is that Year One Labs is ideal for investing at the earliest possible stage, and focused almost entirely on the people (the idea and market are very, very distant points of evaluation.) Our model allows for a team to come into Year One Labs with a vague idea, start the customer development process and either validate something in that space, or move into completely new areas of interest.

Our Lean Startup Methodology

It’s important in this context to understand how our model actually works. As I mentioned, we provide $50,000 in funding. That money is tranched into three stages:

  1. $10k is provided immediately for the Probe Stage
  2. $20k is provided to build the Minimum Viable Product (MVP)
  3. $20k is provided to build traction

During the Probe Stage, the entrepreneurs are focused entirely on customer development and validation. They’ve identified a problem of interest, put together their hypotheses and they go out and do as many Problem Interviews as possible. If they find something worthwhile, they move to Solution Interviews as well as surveys (for collecting quantitative data). If things don’t go as planned, they either look at doing something completely different, or they find something worthwhile within their area of interest and pivot. Then they’re back to interviewing people and learning as much as possible.

Passing the Probe Stage is fairly consistent for everyone. They need to have a clearly defined problem (that’s worth solving), a solution (that’s doable), a market (that’s interesting based on size, opportunity, access) and a definition of their MVP. The definition of their MVP needs to include an explanation of their experimental design. It has to be clear to everyone how the entrepreneurs will validate or invalidate the success of the MVP. We then provide the second tranche of funding.

Once a startup gets to the MVP stage they’re in their sweet spot of expertise. Everyone at Year One Labs knows how to technically build products. The challenging part is moving from the launch of the initial MVP into the third stage for building traction.

We’ve learned now that you can’t think of the MVP as a launch, or as a single product release. The MVP is a process in and of itself. That means we don’t provide the third tranche of $20k immediately after a startup releases its first MVP. The companies are now focused on the Measure and Learn stages of Lean Startup. So they put their MVP into people’s hands, measure response (qualitatively and quantitatively) and iterate as quickly as possible.

The third tranche of funding comes at the point when the startup has validated that it makes sense to scale user acquisition. This could take months, but it’s important for them to go through as many cycles as possible, iterating the product, validating or invalidating their hypotheses, and giving themselves enough time to find the right product for the right market. It’s difficult to put specific statistical milestones on the gate between MVP and traction building; it’s proving to be somewhat of a case-by-case situation.

The Localmind Example

localmind logoLocalmind is a mobile (and Web) platform for asking questions of people that are checked into locations. The first MVP was launched in December 2010. It was only Web-based. We immediately started seeing interesting patterns and numbers. Surprisingly, people were very willing to answer questions. But given that it was Web-only, the volume of questions was low. So was user acquisition. At that point we had already proven some of our hypotheses, but not all of them, and agreed that it wasn’t ready to move from the MVP process stage to the traction stage.

The team launched an iPhone version at SXSW in March 2011. That immediately gained incredible traction. Thousands of users signed up and started using the app.

On April 20th, Localmind got more buzz at Where 2.0, winning the Startup Showcase event. Again, this drew in tons of new users and activity. The buzz and attention are great, and by acquiring more users and increasing activity we were able to collect enough data to validate more of our hypotheses. That doesn’t mean we have all the answers and Localmind is skyrocketing to millions of users in an instant, but the acceleration of the process allowed us to quickly figure out what was working and what wasn’t. This also provided enough proof for Year One Labs to provide the final tranche of funding and give Localmind the ability to focus on traction building.

Lenny Rachitsky, co-founder of Localmind, will tell you openly that Localmind hasn’t hit product-market fit. But that’s not the goal of the MVP process; the goal is to get far enough along to validate/invalidate your hypotheses around the value of the product. The traction stage is for growing the user base and the business, but also for continuously iterating and learning towards product-market fit.

7 Lessons Learned Running a Lean Startup Accelerator

We’ve learned a lot launching and running Year One Labs. And we’ll continue to learn. Here are 7 lessons learned that I hope you’ll find useful:

  1. People talk Lean but don’t really do it. A lot of people talk about Lean Startups and customer development, but very few people really do it. There are some fundamental reasons for this. For starters, it’s a lot harder than people realize (which I’ll cover in more detail below). It also forces a lot of people (particularly developers) out of their comfort zone. When we suggest to a technical founder that the first step for him is to cold call 50 people, the fear is obvious.
  2. It’s a lot harder than people realize. Running a startup is incredibly hard. Running a Lean Startup is harder. The methodology forces things that require a lot of effort: intellectual honesty, rigorous processes, skill development outside of one’s comfort zone, etc. It’s easy and natural for people to fall back into bad or old habits and go off course. It’s easy to lie to yourself as a founder. People don’t really understand good experimental design or good metrics. As often as we talk about the dangers of vanity metrics, they’re so sexy! It’s hard to ignore them.
  3. The DNA has to be set early. One of the key values Year One Labs provides is the early setting of Lean and customer development into the very DNA of the founders. Once a founder sees the results of his efforts (even if it was a painful process) they usually buy in and become believers. Until that time, they wander, argue and don’t fully commit.
  4. Writing things down is critical. It’s amazing how valuable it is to write things down as a way of structuring thoughts and actions. We’ve used the Lean Canvas before, but we’ve also designed our own Problem-Solution Canvas (which we’re still learning from and iterating on!) that’s focused at the very earliest stage of validating an idea.
  5. Business model hackathons work. We’re big fans of 1-day developer hackathons as a way of simulating what it’s like to create a startup. But we’ve also done a number of business model hackathons that really force entrepreneurs to think creatively, let go of any baggage they have, and explore ideas. A business model hackathon might be a 1-day session, but it could also be something that spans a longer period of time. With one of the Year One Labs startups we had them work on 5 separate ideas, do some customer development and analysis on each, and pitch us on them. The goal of the exercise wasn’t to have the startup build 5 different things, but to put them into the right mindset about how to evaluate ideas, identify problems and solutions, and think openly, creatively and seriously about opportunities.
  6. There are very few absolutes. It’s true in life and it’s true with startups. It’s also true that there are very few absolutes when it comes to Lean Startup. You can’t treat every situation, person and startup the same way. There’s going to be some give and take, and some variance in how you do things. It doesn’t mean the process implemented for everyone at Year One Labs is any less rigorous, but we’ve definitely recognized that you can’t impose absolutes and expect it to work. In some cases the Year One Labs partners don’t agree, and we know that we have a tendency to provide disparate feedback and ideas. It might feel like a runaway train at times, but by having an overarching process and guidelines, we maintain the necessary anchors to keep everyone focused.
  7. Focus, focus, focus. Perhaps the biggest value that we provide at Year One Labs is focus. It’s so easy to get distracted when running a startup. There are constant highs and lows, pretty shiny objects to chase, partners to talk to, investors, thousands of product ideas and more. But focus wins the startup marathon, and so we work on a daily basis with entrepreneurs to maintain as much focus as possible.

It’s Really About Acceleration to Value Creation

Earlier I said that Year One Labs would be a success if we return meaningful money back to investors. That’s absolutely the case, although even with a model that focuses on early exits, we expect it to take a good number of years to see results. In the meantime we have to judge our own progress, iterate and improve, and figure out how things are going. I look at this as “acceleration to value creation.” Our goal for each startup is product-market fit, and we’re accelerating them as quickly as possible to that goal. But value creation can also be measured in part by other criteria: follow-on funding raised, user acquisition, revenue, etc. These are all criteria we’re evaluating and will continue to evaluate as our startups progress through the process. Beyond metrics, we have to fundamentally believe that we’re creating value for and with the startups, and doing so at a constantly accelerating pace.

June 6, 2011 Posted in Startup Accelerators by

  • http://twitter.com/tydanco Ty Danco

    If your future classes are anything like your current class, you folks will be retiring soon (and that’s in the good way, not the bad way!) Congrats on a job well done.

  • http://dmix.ca dmix

    > A lot of people talk about Lean Startups and customer development, but very few people really do it.
    This is very true. 

    It’s difficult to not become attached to your idea. It’s a natural psychological trait to rationalize everything once you’ve convinced yourself. Super-critical mentors make this easier.

    Also it’s much easier to quickly develop products in a vacuum than it is to slow down and validate constantly.

  • Arrow

    Just curious, has it’s been decided if you will accept more startups once the original five has graduated?

  • franchise for sale

    Business for sale actually resulting from the amount of responses you will be entertaining as buyer’s perception of value rises. Nevertheless, you can be pleased with the great understanding of a good deal of online marketplace spots provide some tools to support you to definitely determine a fair, normal offering price point.
    Regards,
    franchise for sale

  • http://www.instigatorblog.com Benjamin Yoskovitz

    Not yet.

  • Pingback: Launching a Startup Accelerator « Ivy Plus Alum

  • http://webbroi.com Casey

    Awesome post. Like the honesty.

  • http://twitter.com/kevindewalt Kevin Dewalt

    Awesome, guys.  Your observations perfectly parallel what I have experienced and what I’ve seen with startups I’m helping.  

  • Anonymous

    This
    post was very nicely written and it also contains many useful facts i enjoyed your
    way of writing this post thanks you have made it very easy for me to understand…

  • Pingback: How To Run a Business Model Hackathon

  • http://twitter.com/evanpro Evan Prodromou

    Ben, you say, “I think this has a lot to do with the lack of entrepreneurship being fostered in Canadian universities.” I don’t remember any kind of entrepreneurial classes or emphasis at my university (UC Berkeley, back in the 80s). I think there’s a very natural outcome of specialization in university thinking — everyone has a major — that suggest that CompSci students should learn about computer science (and nothing else) and business students should worry about business (and nothing else). I doubt if you polled CS students today, if you’d get more than 5% that knew they even *could* start their own business. I don’t think that’s Canadian universities — I think that’s *any* university.

  • http://www.instigatorblog.com Benjamin Yoskovitz

    Could very well be, although we’ve seen examples at Stanford w/ the FB app class that show an appetite for doing things a bit more entrepreneurial in nature.

  • Ted Benson2011

    There are very few absolutes…boy, not a week goes by that I don’t say that. I never speak in absolutes. reality changes as we perceive it, even as we do.very nicely written article!

  • http://www.essayprovider.com/thesis-dissertation-writing-services.aspx Thesis Writing

    Great post and some really useful tips there. I love resource lists like this. Have social bookmarked it in the hope that others can also benefit.

  • Ayhan KARAMAN

    Thanks backlink

  • Pingback: Finding Your Startup’s One Thing

  • Pingback: So You Want to Run a Startup Accelerator…

  • Pingback: Year One Labs Completes its Incubation Phase

  • Pingback: This Much I Know is True

  • Pingback: Lean Startups, Word of Mouth Marketing, and Email | Firepole Marketing Blog

Ben Yoskovitz
I'm VP Product at Codified (makers of VarageSale).

I'm also a Founding Partner at Year One Labs, an early stage accelerator in Montreal. Previously I founded Standout Jobs (and sold it).

My bio »

or

Follow me on Twitter

Get updates and special content
When I publish new content, get it directly in your inbox. Subscribers will get special stuff as well not available on the blog (but I promise it will be infrequent + high quality.)
Get the Lean Analytics Book!
Awesome Jobs
Check out the job opportunities at my portfolio companies.
Startup & Investor Resources
Find Stuff
My Photos