Lean Startup Interview with Dave Shackelton and Steve Gotz

by Anthony Howcroft

I’m pleased to be interviewing two members of the Frost Data Capital Incubation team, Dave Shackleton, Chief Incubation Officer, and Steve Gotz, Director of Incubation. We’re here to talk about Lean Startups, and to discover how the companies in the Frost portfolio are adopting this approach. The Frost team has launched 17 companies so far, and has just announced a partnership with GE called Frost I3 – Incubator for the Industrial Internet, that will launch up to 30 additional companies in the next 3 years. Lean Startup is a key part of their model.

So guys, Eric Ries published his Lean Startup book in 2011, although he’d been blogging about it from 2008 onwards. When and how did you both get involved?

Steve:  I think Steve Blank really kicked-off the whole push towards optimizing the startup process with his book Four Steps to the Epiphany, which really inspired Eric. At the time I was working in a lab creating new startups and the idea of making the process more repeatable was very attractive

Dave:  I first met Stuart [CEO and Founder of Frost Data Capital] in 2003 when we were both on the advisory board of a company in Carlsbad, and we connected on a similar view of
startups. Many years later in 2012 we reconnected and discussed how Eric’s book was a good articulation of what Stuart had done in his prior successes and would like to replicate in a new incubator he was launching.

And so you’d both been doing startup work in your roles?

Dave:  Yes, I had been involved in startups for many years, had experience with two other incubators (Idealab and Kindling), and had even been an executive in a company focused on software prototyping (iRise). Both Eric Ries and Steve Blank had been inspired by Design Thinking by David Kelley, a professor of mine in college who had a huge impact on me. So my connections with the core philosophy run pretty deep.

Steve:  Yeah… and Blank’s definition of a startup really resonated with me: “A startup is merely a temporary organization designed to search for a repeatable and scalable business model.” It’s basically the scientific method applied to companies

If you had to explain Lean Startup to a friend who wasn’t in business or technology (like an artist) what would you say?

Steve:  I describe it as a systematic way to identify and eliminate the most common types of risk involved in starting up companies, particularly product and customer risk. Most startups fail because they build the wrong product for the wrong customer. Technical risk is usually not a key driver towards failure

Dave:  Startups have historically operated in a way that resulted in a very high failure rate, and The Lean Startup is a different approach that attempts to dramatically improve the rate of success. It does this by focusing on specific things like “validated learning” that reduce the risk of building products and services that nobody wants.

Can I call Lean Startup a methodology?

Dave:  There has been considerable debate about whether it is a methodology or a philosophy. We have our own “Frost Incubation Methodology” which is based on Lean Startup principles.

Steve:  We tell our CEO’s not to hedge…but I’m going to break that rule. I think it’s a bit of both – methodology & philosophy

I know about the Frost Model but I didn’t realize it was a methodology!

Dave:  Ha! Actually, we have been using “Model”, “Process”, “Framework”, and “Methodology” almost interchangeably because very few people have succinct and consistent definitions of each. We do have a process that sits within the Frost Model. Incidentally, I think Ash Maurya was successful with his book Running Lean because he provided more concrete ways of carrying out the philosophies in The Lean Startup. It’s a good read, if you are adopting Lean Startup.

So why is this different from previous management fads like Business Process Re-engineering, or Management by Walking around or Six Sigma? Will we all have moved on to something new in a year or two?

Dave:  We will likely have a new articulation that improves on Eric in the same way that Eric improved on Steve, but it’s an evolution rather than a series of fads.

Perhaps it will be The Frost Model!

Steve:  I think management science is continually codifying readily observable behaviors in business frameworks. The notion of a continuum or evolution is a great way to describe the process.

Dave:  I think it shares certain characteristics with prior trends like BPR in that just checking boxes in a methodology doesn’t create the desired results. Rather it requires diligence in very specific dimensions. The same is true with Design Thinking. There have been criticisms that it didn’t result in what people expected, but those were cases of lip service rather than diligence. Take prototyping, something that is critical both in Design Thinking and The Lean Startup. A failed example might say “but we created 3 prototypes!” where the successful one says “we created 100”.

Yes, I agree. I think a lot of the ‘fads’ have actually contributed to current thinking, via evolution and new combinations of ideas

Steve:  I think one of the great challenges with Lean Startup is that on the surface it seems so simple, almost trivial, but in practice it takes quite a bit of effort to implement with discipline

Yes, harder than it seems! In Eric’s book, a lot of the examples seem to be in B2C businesses – do you think it can work for B2B where the volume of transactions may be much smaller (although higher value)? 

Steve:  Without a doubt….and even more so

Dave:  One of the reasons we say we follow the “philosophy” of the Lean Startup is specifically that enterprise software has different characteristics but the same challenges. For example, we have far fewer data points but our measurement is just as important. Five paid POC’s with enterprise customers might be an equivalent metric to thousands of customers in a B2C

Steve:  Consumer markets can be fairly forgiving but with more protracted enterprise sales cycles… if you miss the mark on either the MVS or MVP, you might be looking at substantial setbacks

Dave:  Great point Steve

So how would you say you’ve adapted the Lean philosophy to fit the Frost Model?

Dave:  The first way is the one I just mentioned – metrics. For example, we have an evaluation of “Resonance” as a milestone for proving out a Minimum Viable Segment. Actually, even the concept of a Minimum Viable Segment is something that doesn’t appear in the Lean Startup. We didn’t invent MVS at Frost, but it has become a key part of what we do

How do you measure “Resonance”?

Dave:  “Resonance” is multi-dimensional and has to do with the validation of hypotheses by key people within target market segments

Is it clear to the CEOs how it works? I mean, in my experience a metric needs to be something everyone understands how to influence, or else it’s just a light on a dashboard

Steve:  A key part of the Frost Incubation team’s job is making sure our CEOs are aware of the Frost Data Capital metrics and what success looks like at each step along the way

How many metrics do you track per company?

Steve:  Classic metrics that we track include problem validation via resonance, then solution validation, and eventually growth metrics. Of course each of these is composed of sub-areas that collectively allow us to assess our portfolio companies

OK, so is that 5 big ones with 20 sub-metrics or…?

Steve:  Within Frost DC we’ve actually come up with a single integrated score that represents company trajectory and the likelihood of a given set of outcomes. Where we have companies with more uncertain outcomes, we invest extra time and effort, using the lean process to convert product or market uncertainty into risk. Risk is something that can be managed, uncertainty is more amorphous. I think another adaptation to the Lean philosophy is having a predefined set of corporate partners who trust us enough to cooperatively work together to proactively identify pain points and validate them in a meaningful way. Just having that dynamic goes a long way towards removing the traditional risk attached to a new venture (e.g. what’s the problem and who’s going to pay for it.)

Well said, Steve. Not sure I remember much on that from my Lean theory, but it is certainly a key part of The Frost Model

Dave:  Yes, excellent point – when we at Frost DC talk about implementing The Lean Startup, we are starting from a fundamentally different place. We have a singular focus on what we know: big data analytics, CEOs with long and successful histories of startups, a financial model that better aligns everybody, in fact all of the strands of our Frost Model. Another critical part is that we drive this from the top. I can’t emphasize this enough. We have structured our board meetings to discuss hypotheses, validation, business model canvases, and so on rather than revenue projections. Every CEO is constantly thinking about it. We use it as a “common language” among all of our CEO’s and Frost DC folks

If all of the team and CEOs are really focused on this, and you have a single integrated score that represents each company’s trajectory, does that not make it a competition between CEOs?

Steve:  Good question. The Frost Model incentivizes our CEOs to collectively support each other.

Dave:  We have also recently organized by clusters or sectors (vertical or horizontal) as we have grown, and we’re finding even better collaboration by sector. The CEOs are all competitive types, so they’re naturally inclined to show great progress, but they do work very well together. This comes back to the benefits of having a defined focus like big data analytics, as our CEOs talk to similar customers and partners, have similar technology challenges, and so on

Steve:  Our companies are not in competition with each other. The score is not about company performance vis-a-vis their peers. We constantly see collaborative problem solving, especially where newer companies are facing issues that have been tackled by some of our more mature companies

Thanks for clarifying that. Changing tack slightly, in his book, Eric says “Startup success can be engineered by following the right process, which means it can be learned, which means it can be taught”. So if I follow the principles, will I end up with a billion dollar business?

Dave:  HaHa! This comes back to our earlier discussion of discipline. Actually, if followed correctly it will help you find out quickly that your first idea wasn’t a billion dollar business but that something else can be a 150 million dollar business within four years. If you are disciplined about the process, you will definitely achieve the “validated learning”

That’s a great way of expressing it. I like the way Stuart Frost also says that while you can make the process much easier, you still need the ‘magic pixie dust’, and an entrepreneurial spirit

Dave:  So many things have to go right for an enterprise software startup to succeed. As Steve mentioned earlier, we are diligent about seeking out areas of risk and reducing those while putting in place things to increase our chance for success

Steve:  Magic pixie dust is a great way to describe it. There is always an element of luck and force of personality… we just don’t want to be solely dependent on it to deliver a successful outcome!

I hear a lot about the importance of hypotheses. Can you define them, and suggest how many one of our startups might have?

Dave:  A hypothesis is an assumption about a business that is required to be true for it to succeed and is clearly falsifiable. There are naturally a huge number of these, but only a limited number will rise to be the highest dependencies

Steve:  Every decision point is an opportunity to express a hypothesis and test it with the market. The more this is done, the more uncertainty that is removed from your business model. For a really diligent lean startup practitioner I wouldn’t be surprised to see hundreds of hypothesis tested over the course of their development. Dave’s point is important though, because you start with really high-level hypotheses that call into question entire aspects of your business and then over time your hypotheses get more focused

How about pivots, where the whole company changes direction – how often do the Frost portfolio companies do them, and who makes the decision?

Dave:  There is an interesting discussion about the difference between pivots and iterations. At the beginning of a business, we change the direction constantly in the face of more information through validation. Later on, it is more of a pivot if originally it seemed like the right path and then new validation threw that into question

Steve:  As much as possible pivot decisions shouldn’t made by people, they should be made by the data

Good philosophy, but someone has to make the call!

Steve:  Agreed

Dave:  It’s generally not a surprise to anyone that a pivot is happening – the validation data shows it so the CEO, the Board, the General Partners, are all normally in alignment

Is there not a slight stigma attached to a CEO who makes several pivots? Even though he is following the data?

Dave:  Keep in mind that CEOs here are more often than not buying into joining the Frost DC family, not just a specific company. So they are more open to making substantial changes based on evolving information. A pivot is only a concern if it took too long to make

Steve:  I’ve been here for several months, witnessing lots of startup validation, and I haven’t seen a single instance of stigma attached to pivots or changes of focus. In fact it’s the exact opposite, pivots are a sign of healthy engagement with the market and active assimilation of data and facts

We’re nearly out on time, and I realize we haven’t spoken much about Minimum Viable Segments (MVS) and Minimum Viable Products (MVP). One quick question – which may seem obvious, but which comes first?

Dave:  MVS. Although we are always thinking about both at the same time

Then if we pivot on the MVS, do we have to change the MVP?

Dave:  We can’t validate a market opportunity and problem without some ideas of innovative solutions, so they evolve together, but we push CEO’s to validate an MVS before developing an MVP

My gut feeling is that most people over-spec an MVP. Would you agree, and if so, any tips on avoiding this?

Dave:  One approach we have found helps is to do more early prototyping and doing an MVP for something much more limited than your MVS. So for example, one of our companies, Jointly Health, did a very early deployment with just five local patients that was nothing like what they would eventually build, but it gave them a lot of great validation and MVP information that they wouldn’t have had otherwise

How about using wireframes, videos, or ppt mockups?

Dave:  Absolutely! We have companies whose first Proof-Of-Concept was with a powerpoint deck and services

Great, I hear you have strong sales & marketing team within the incubator that can help with that :) What happens when you change the MVP based on your customers experience though, and one of your big customers liked it the way it was?

Steve:  Invariably your product will never survive its first meeting with your customers, so don’t sweat it. Rapid prototyping the key aspects of your MVP is a great way to confirm solution hypotheses before investing significant dev resources

You must have to set expectations with early adopters though, that the product will change?

Dave:  At the point in the evolution our early customers are with us on the idea of taking a really exciting and valuable journey, so they understand what is likely to come through

Steve:  As Dave says, we are taking our customers on a journey. We have found that the more involved they are in helping to collaboratively solve the problem, the more understanding they are as the product evolves

Is there a point where Lean finishes? Where you have a product defined, it’s selling well, and the CEO switches focus to the next round of funding and ramping-up sales capacity?

Dave:  No, Lean Startup for us is involved all the way to exit. After the “value hypothesis” is the “growth hypothesis” which CEOs need to validate in order gain traction and growth. Our model though is to exit earlier than normal because that best meets the needs of the partners who acquire our companies.

Steve:  It’s an interesting question… given that we’re starting to see companies like GE implementing Lean Startup practices and even setting up centers of excellence, I have to think that within the right corporate culture there is a place for lean to continue

Guys, thanks so much for your time today. If people want to learn more about our lean approach at Frost Data Capital, are there any resources you would point them at?

Dave:  No, it’s totally secret :)

Steve:  You could try here though: The Frost Data Capital. It’s got some videos of Stuart Frost, the CEOs, and Partners discussing our approach. There are also blog posts that talk in more detail about our approach.