I am delivering this talk at the Lean India Summit, Bangalore, Nov 15-16. Ahead of my talk, this Q&A captures some of the key ideas behind the talk.
Abstract of the talk: Startups that operate under a stealth mode achieve over 90% failure rates. While they might have bright ideas, access to top talent, adequate funding, etc., they typically fail to accomplish their original objectives. A key reason behind such spectacular failure rates is premature scaling at each stage of the startup. In this talk, we will examine the mistakes that startups make, and what we can learn from the Customer Development model proposed by Steve Blank to improve better chances of survival and growth.
Q1. You mention that premature scaling at each stage of the startup leads to failure. What is the right time to start scaling for the startup? How can the startup course correct if they realize that they have started scaling prematurely?
A: As Steve Blank says, a startup exists to search for its business model and not to execute it, I think the right time to scale is when you have ‘discovered’ the ‘business model’. What that means is that unless you have gone through Customer Discovery and Customer Validation process and turned your hypothesis into facts based on actual market feedback, you don’t really have any need (and hence the justification) to scale up.
If you have started scaling prematurely, I think you should stop and introspect. Ideally, stop all other marketing and sales and other activities of scaling up the product development organization, and just focus on validating your initial hypothesis. One important point to remember is that sometimes having excess cash could force one to scale up, so avoid getting more cash than you really need!
Q2. There are couple of questions we asked one of our other speakers Wes Williams, which we think is relevant to your topic as well. Would like to hear your thoughts on this – a) How does the size of an organization impact its scalability? Would you need to think of a different scalable model for large sized organizations?
A: Large organizations have typically been cash-rich and hence were more eager to scale-up in the past. However, the harsh realities of today’s economic environment have forced them to be more conservative. I believe even the largest, most cash-rich organizations are thinking twice before scaling up prematurely. I know how hard it is in most product companies to get an additional resource even, so clearly, the old rules don’t apply so well anymore.
I would build a much shorter runway where I can demonstrate tangible results in terms of customer discovery and customer validation that allow me to pitch my case for incremental funding that is backed by solid data from the market. That way, my expenses and promised revenues are growing in tandem rather than making a wild promise about the revenue without a scientific basis and piling up a large expense plan in anticipation of revenues.
Q2. b)What do you see as the difference between scalability, flexibility and agility?
A: Agility is the ability to achieve (or exceed) the set goals by being flexible about how to accomplish them. Flexibility is the means to accomplish agility as an end-goal. Even there, let me correct myself – agility is not an end-goal. It is once again a means to accomplish the next higher-order goal, viz., create a successful business.
Scalability is very different, though an effective scalability might require them both. In traditional manufacturing, it was all about building something at a large-scale so that economics of manufacturing or production could become much more affordable. The problem is that accomplishing scalability invariably requires lots of sunk costs and irreversible decisions and might be fraught with risks if the underlying assumptions turn out to be false. Henry Ford was able to reduce car prices by creating a process that eliminated user customization (“you can have any color of car as long as it is black”) to create a standard one-size-fit-all car that could be mass produced to achieve unprecedented economics of scale. However to achieve such economics of scale, he had to pump in millions of dollars in manufacturing plant and a car design that might hopefully sell in the market. So, in manufacturing, mass production brings down prices, and hence, scaling up is good, but one needs to make sure they have the product that the market wants, lest they be stuck with deadwood!
In software, scalability needs to be understood differently from manufacturing. Presently, there are two key manifestations of scalability – one is scaling-up product development process across the organization, e.g., going outside a single scrum team and aligning to the strategic portfolio, etc. One such framework that attempts to integrate all organizational functions together is Scaled Agility Framework. The other one is scaling-up your software product or services, e.g., offering the app on all major OS versions of mobile or tablets, or offering the software in all languages, or offering the product in all geographies, etc. The first one is all about being internally efficient and reduce friction and hand-offs across different departments in achieving a seamless supply chain to ‘scale up’ your agility throughout the organization and not just limiting to software teams alone. The second is all about scaling up your product and services from the market coverage or product capability standpoint and is conceptually more akin to traditional notions of scalability. One needs to be pragmatic about market potential before committing to any such expenses that involve sunk costs or are irreversible in nature. To that end, the principles of agility are a must to establish early proof points about the real assessment of market potential.
Q3. The Customer Development Manifesto suggested by Steve Blank mentions that “A Startup Is a Temporary Organization Designed to Search for A Repeatable and Scalable Business Model”. In your experience, how much of scaling is good for the organization? At what point does the startup need to pause and introspect on whether to scale any bigger, or to look for some other way to scale?
A: There is some very good data from Startup Genome Project that captures such data points. While I think such guidance is very comprehensive and highly relevant for all startups, it must not be construed as universally applicable. Take the data as a guidance and question before committing any fixed expense item whether you really need it at this point? To me, it is not any different from how one would commit to such expenses in their home or even as a group manager in a company.
Customer Development model provides a framework to systematically validate all hypothesis before going out and scaling up the startup prematurely
I meet so many entrepreneurs who build a large team – VP of Sales, VP of Marketing and even 8-10 developers before they have built a serious product to market! If you follow Steve Blank, he is very unambiguous in his guidance – during customer discovery phase, it should only be the founder of the company to validate the ‘leap of faith’ hypotheses, and no one else! I also find that sometimes getting funding early in the day could make people scale-up prematurely and that’s just as bad.
Is there an alternative to scaling up? Yes, first off, build an MVP and keep your expenses low. A key thing about the MVP is that you must aim for selling the MVP as a means to validate your Business Model – freebies don’t give the proof point that you need to turn your hypothesis into facts. Once you have a reasonable confidence backed by solid data about customer validation, then you might be better-placed scaling up.
Q4. Could you talk a bit more about the Customer Development model proposed by Steve Blank and how you are going to extrapolate from this model in your talk?
A: Traditional startups, for example during dotcom, were known for the so-called ‘stealth mode’ development where they could work for an extended period of time, often away from public glare or any meaningful form of real-world customer feedback. However, we now know such a model is fraught is too many risks. It is the Hail Mary pass of entrepreneurship! On the other hand, Customer Development model is a highly scientific approach to entrepreneurship that recognizes the need to have real-world market data before going out and building the product. In today’s time and age, technology and consumer preferences change at warp-speed, and it is no longer realistic to build a product with a long-lead time. The worst thing that can happen is to build a product that no one wants! This might appear as poetic exaggeration, but the world is literally full of such experiments. Think of McDonalds Pizza or Colgate Breakfast Entrees or Frito-Lay’s lemonade or hundreds of such products – they were all created by highly successful companies that definitely knew what it took to built new products. And yet, they all failed. Why? Because essentially, they built something that the market didn’t need. Surely, they must have done market research throughout the lifecycle, but clearly they missed something much more fundamental – product development is not simply developing a nifty idea but also entails a parallel cycle of customer development, especially in new markets. Customer Development model is one such model that helps figure out what and how to do to do customer development – whether one is a true-blue startup or a brand-new idea in an established company.
Q5. New research by Harvard Business School’s Shikhar Ghosh shows that 75% of all start-ups fail. In your experience, generally what must the remaining 25% be doing right?
A: One thing we could be safely assured they are doing well is that they are listening to early feedback and adapting to it rather than going out and building the complete product or service based on untested hypothesis. And to be able to do such adaption both effectively and efficiently, they are using short iterations where they are able to give functional products in the hands of the customers so that they can get some valuable customer feedback. Popularly known as an MVP (or a Minimum Viable Product), it allows a high degree of prioritization of features from users point of view, and an agile development process that allows creating rapid iterations of high-quality and well-test features.
Q6. If I am part of a start up like organization, operating within a giant organization with its set business models and processes (that are also required to keep the big brother organization up and operational), how does one navigate the plethora of subsidiary business groups like facilities, infrastructure, HR, etc. to enable a lean start up mode for my particular group?
A: No easy answers here! In fact, history has repeatedly shown us that the more successful an organization is, the more it is difficult to get a radically new idea going. No wonder why Sony, the undisputed leader in Trinitron technology, completely missed the Plasma and LCD display bus (before eventually joining the LED display party, even if if a bit late) or the Internet leader Google who has never quite figured out its social strategy (well, some might argue that with Google+, they do have it, but that’s one point of view). So, traditional methods of pitching an idea and expecting the upper management to bless might work in theory, but I wouldn’t rely on them.
Over year, I have found one approach that perhaps works much better compared to anything else – skunkworks. If you are truly passionate about the idea, let that passion help you ‘recruit’ other co-creators (always remember – it takes a village to raise a child) and then put your spare cycles to build a prototype to demonstrate your idea. If you are not willing to take risks and spend days and weeks in even trying the idea, why should you expect that the management will give you budget for it? It’s like going to bank asking for money, whether to buy a house or to build a company – they want you to put down some money first in the form of your contribution or a guarantee, and the logic is same – if you are not willing to take any amount of risk behind your idea, why should others do it?
Java language came out of skunkworks at Sun. And so did the first Apple Macintosh. I have seen several good ideas come out like this in large companies, though I don’t think anyone can guarantee skunkwork projects. However, on a relative basis, it is much better demonstration of the power of your idea (compared to a powerpoint deck) and your passion that might improve your chances of getting the necessary organizational support to take your idea to its next leg of journey.
Q7. Finally when is it the right time for an organization to realize that it has grown too big to be in a startup mode anymore, and might need to revisit its way of working to adapt to its current size?
A: This might be the next million-dollar challenge and I wish I knew the cookbook answer! Unfortunately, we build the bureaucracy one day at at time and before we realize, we have built an impregnable wall of self-preservation that blunts any attack on status-quo. We don’t realize it because it is like slow-boiling the frog. Initially it looks innocuous and we tend to trivialize it, until one day it becomes so huge that we can’t get rid of it even when we try so hard!
I think an organization has stopped being a startup when people stop complaining and simply give up on you, when no one volunteers and there is generally a diffusion of responsibility, when people build territories and silos, when we are intolerant of failures and anything perceived disruptive, when people want to join because of the perks of the company rather than the challenges they can work on and the impact they can create, when more time is spent in waiting for decisions rather than learning from actions following implementing those decisions, and so on…
Q8. Who is the target audience for your talk? Only startup organizations?
A: Anyone who is undertaking a new endeavor under conditions of extreme uncertainty.