Derivatives

When we try to measure our startup ecosystem we love to count the number of new companies that are started. Or, as they progress, to count how many dollars of capital these startups raise.

But we are poor at reporting on how many of these young companies survive beyond the early stage and go on to contribute anything back, either to their individual investors or to the economy as a whole, via jobs created or tax paid.

We forget that a startup is a phase, not a destination.

Because the failure rate of early-stage startups is so high, it’s tempting to treat them, in aggregate, as a lottery, and believe that to increase the chances of a win we just need to buy as many tickets as we can afford.

But that’s not how it works. It’s not a game of pure chance. There is a pattern. Increasing the number of random bets we make doesn’t increase our likelihood of success. It just increases the amount we’ve spent.

It’s true, we need more startups. But support for startups in New Zealand is seldom centred on individual founders and their teams.

It’s much more likely to be what I called derivatives:

Innovation hubs and clusters; shared working spaces all around the country; accelerator programs of different flavours; networks of angel investors pooling their resources and investing in a portfolio of ventures; public and private organisations that provide advisory services to startups and mentor founders; countless competitions or networking events hoping to flush out promising new business ideas; numerous business awards celebrating entrepreneurship, growth or innovation; various initiatives to commercialise research done at universities; and, last but not least, millions of dollars of local and central government funding including direct grants to companies, subsidised professional services and advice, tax credits and co-investment.

Many people have put significant time, effort and money into creating all these things. A large and growing industry has sprouted, all ultimately trying to increase the number of startups. Startups, and especially early-stage startups, are suffocated with support. And still we continue to invest in more.

Is it working?

What is a “derivative”?

In calculus, a derivative measures how much one value changes in response to changes in some other value.¹

For example, as an object moves we can measure its speed (the first derivative of its movement) and its acceleration (the second derivative of its movement). The higher the derivative the more abstracted we are from the underlying object.

In finance, a derivative is a contract whose value is based on the performance of an underlying asset.²

For example, a stock option is an agreement that provides the option to purchase a share in a company at some date in the future for a pre-agreed price. The profit or loss on the option depends on whether the actual price of the stock on that future date is above or below the pre-agreed price.

Without the underlying company the option is worthless. As investors discovered to their cost in the Global Financial Crisis in 2008, when risky assets get packaged up, given interesting names, and abstracted enough, it’s easy to completely lose sight of what we’re actually investing in, and indeed if there is any value in that bundle at all.³

Abiotic eh?

So what are startup derivatives?

All the people and organisations that are part of the startup ecosystem, but which are not themselves startups. This concept of derivatives gives us a useful way to categorise these things. We can simply consider how close they are to the underlying ventures they depend on.

Some examples…

Founders and their teams are the underlying ventures. They are the people in the arena, dust and sweat and blood on their faces etc.⁴

Incubators, shared working spaces, innovation hubs and accelerator programs are first derivative. Their success is a simple function of the subset of ventures they work with.

Active investors are first derivative too. They are like a tender to a steam train — providing the fuel and getting pulled along at the same time. Passive investors, or anybody investing indirectly via a venture fund, syndicate or angel group, are second derivative, since they are an additional step removed from the underlying ventures.

Advisers working directly with founders are first derivative. Agencies that connect founders with advisers are second derivative.

Business awards are first derivative. Sponsors of awards, or judges, are second derivative.

Networking events are second derivative. The purpose of these is not to create ventures directly, but to help connect people who might with each other and provide some experience in a controlled setting.

Dragons Den (just humour me here!) is second derivative, but only if we very generously assume that some people watching might be inspired and encouraged to become a founder or investor.

Government funding adds an extra layer of abstraction.

Funding NZTE to provide in-market support for startups is second derivative.

A fund-of-funds, like the government venture fund, is third derivative.

Funding Callaghan Innovation to run a program to develop accelerators … we are up to four derivatives at that point! ⁵

In case you think I’m just throwing shade at others here, writing a book about the startup ecosystem is third derivative at least.

(The name given to a third derivative in physics is “jerk”. The fourth derivative is called “snap” or “jounce”, the fifth is called “crackle” and the sixth is … of course … “pop”).

Signs of life

Categorising the different things we do to support startups like this can help us to separate the signal from the noise.

Remember, there are four simple questions that anybody promoting a startup derivative should answer in advance, to understand the impact it might have:

1. Who does this help?
2. What constraints do they have?
3. How do we hope to reduce or remove those constraints for them?
4. How will we show that it’s working?

We should begin with the founders and teams who are working on real ventures and work backwards from there to the constraints that they have.

If we can’t say in advance how we’re going to prove that we’ve made a positive difference to them, we’re very unlikely to be able to later.

By doing this, we would eliminate a huge amount of the distraction that contributes very little to the ecosystem, and focus our limited time and money and energy where it makes the biggest difference.

Let’s consider two common startup ecosystem derivatives, and apply these criteria to each of them…

Incubate

In a hospital, an incubator is a safe place to put a sick baby – somewhere they can recover without worrying about the stresses of the outside world. In 2007 our son spent the first week of his life in an incubator and it was not a time we look back on fondly!

For a brief moment in time, business incubators were the hot new thing. The idea was to create a protected and supportive environment for fragile young companies. That should be familiar to anybody who has spent time in a hospital neonatal unit.

Typically they offer:

1. A funky office space (often designed to within an inch of its life with bean bags, pool tables, tables of Lego etc);
2. Other early-stage companies to share it with; plus
3. Access to advisors and mentors etc.

These days, perhaps because of the negative connotations I described, very few business incubators are brave enough to still use that label. They prefer to be called innovation hubs or technology clusters. More accurately they would be described as co-working spaces or serviced offices. However, the services they provide to startups haven’t really changed at all. So for our purposes, let’s call them all incubators.

How do they stack up against our impact assessment questions?

Who are these spaces set up to help and how?

There are three obvious challenges when we look at this from a founder’s perspective…

1. The value of the solution is limited by the size of the problem

Firstly, and most importantly, those three things incubators offer startup teams are not usually big constraints, so immediately the benefits are limited.

It is good to have a comfortable desk to sit at and to be surrounded by supportive and like-minded people who can advise and support. But this is very easy to arrange when we need it, especially in a place like New Zealand where the business community is small and well connected and people are generally very happy to help if asked.

Plus, the optimal amount of time spent networking with other founders is “occasionally and intentionally”, not “all day, everyday”. If you like the idea of incubators because you prefer the company of like minded people, seek out non-like minded people. You’ll learn more, and faster.

If you think finding a desk and some sympathetic friends is the hard part of starting a company, wait until you try to find your first customer!

2. Who incubates the incubators?

Secondly, incubators are expensive to create and run. This means the help they offer comes at a price. There are four possible business models incubators use to cover these costs, none of which are founder-centric:

User pays. This effectively makes the incubator a real-estate venture, renting office space one desk at a time, meaning startups pay more than they should. Paying market rates for a desk and access to a fancy coffee machine isn’t a good deal for cash-constrained teams.

Investment. This doesn’t work for either side: founders need to give up a percentage of their company in return for the limited benefits of the shared working space – in the short term equity costs nothing to give away, but in the long run a “free” place in an incubator can end up being hugely expensive if things go well; on the other side of the equation, any incubator taking an equity position in the companies they host quickly ends up with a large portfolio of small stakes in ventures into which they are reluctant to continue investing time and money. As a result, very few incubators contribute more value as shareholders than they extract.

Corporate sponsorship. There is no such thing as a free lunch. Large companies are always keen to be associated with helping startups, especially when they can also supply products or services to the companies they host or when it gives their own staff something interesting to work on occasionally as mentors. So founders need to be aware of the strings that come attached to “free” support.

Government subsidies. Who incubates the incubators? We all do! Remember Hugh from Timaru? Given that many incubators already use all three of the methods described above, it’s slightly depressing that local or central government funding is required at all. On the other hand, if the incubators had to fund themselves as purely commercial businesses very few would survive.

Finally, from observation over the years, the advice that startups get while in an incubator is often terrible. We need to separate the wheat from the chaff.

This is a pattern I’ve sadly seen repeated too many times: a promising startup has built a good enough product, and needs to get out of their own heads and start talking to potential customers, but instead wastes its time working on crafting an investment pitch deck or market positioning document or equivalent nonsense, all because the incubator has a methodology and their mentors need them to progress through each of the steps in the prescribed order.

When we’re at the very beginning it’s not always obvious if the advice we’re getting is smart, or something that might only make sense for a larger company. When an incubator comes with advisors on tap, it’s easy to assume, incorrectly, they are the people who are best qualified to guide and coach.

The problem is that so much of the startup thinking fed to naïve founders in incubators is the equivalent of a backcountry survival guide written by somebody who has never spent time in the bush.

3. Who are incubators really for?

It’s easy to understand the allure of incubators. From the top-down, a building feels like a very tangible thing we can invest in to help startups. Big companies have their own offices, but startups can’t afford that, so perhaps if we make it cheaper it will help startups become big companies.

The real value of these spaces is they allow people who want to encourage startups to feel like they are contributing to the ecosystem. It seems that every town around the country has decided it’s important for them to have a space to host startups. However, if they took the amount they have “invested” so far in incubators and applied it directly as seed capital it could have completely funded a large number of startups.

If we consider the startups that have gone on to become successful high-growth companies, almost all did their hard yards, in the beginning, in a dingy flat or borrowed corner of an office, rather than emerging from the safety and comfort of a business park. Trade Me and Vend both started by sharing desks, Xero started in an apartment,⁶ Timely was a fully remote company for the first few years, so the team were all working out of their bedrooms.

Most founders don’t really need a special desk in a special room and a poorly qualified grown-up in a nice suit or branded hoodie to hold their hand or to be constantly surrounded by other founders.

In a hospital there are babies that really need the protection of an incubator and qualified specialists when they are fragile. But for most newborns getting home as soon as possible is preferable. Likewise, there are some startups that benefit from having a safe space in the very early stages too. But we need to be much more particular about who and really understand how to help them while they are in that mode, and always with a view to graduating them out.

There is an easy way to measure the effectiveness of an incubator:

How many babies make it out alive and go on to thrive?

As an aside, while we’re talking about real estate interventions, here are two suggestions that could make a difference for high-growth companies (as opposed to the early-stage ventures we currently target):

As teams grow they quickly get too big for the dingy flat or borrowed office I mentioned above. There is a valley-of-death moment I’ve seen repeated a few times, where they get to a size where they really need their own office space, but they are not yet steady-state enough to take on a long-term lease. A solution that gives these growing companies dedicated space (big enough to host the whole team away from distractions) but doesn’t require directors to provide personal guarantees or commit for multiple years would be a much better use of the office spaces we’ve created and funded.

There is also a gap for distributed teams who mostly work remotely (which, over time, will become most teams). For them, there is a lot of value in getting the team together from time to time. Much larger fully-remote companies often have their own drop-in spaces for this purpose. The gap is a airport lounge-like space where smaller high-growth teams can book meeting rooms, get good coffee and overlap with each other (not necessarily all day) on an ad-hoc basis.

Accelerate

Accelerators are the opposite of incubators. They are all about speed!

The intention is to take a promising idea and brand new team and attempt to fast-track the venture to a funding event – ideally in just a few weeks or months.

Over recent years accelerator programs have proliferated all around the world, mostly copying the model pioneered by Y Combinator in Silicon Valley.

Accelerators try to select the best founders from a big group of applicants, surround them with the best available mentors and advisors, invest a modest amount (enough for “founders to be able to run their company and pay expenses for around 5-6 months”) and take a small shareholding in the new venture (typically somewhere around 6% or 7%, although some programs take more and others have some flexibility for ventures that might be slightly further along and so justify a higher valuation). The goal is “demo day”, which comes at the end of the program and is promoted as a chance for the startups to pitch themselves to investors.⁷

Again, let’s think about this in terms of our impact assessment questions:

Who are accelerators trying to help and how? And does it work?

Mass assembly

To start with, we can consider this at a program level, and ask what it takes for an individual accelerator to be successful…

The goal of an accelerator is to mass assemble startups the same way as high-tech manufacturing companies mass assemble electronics. It’s not enough to build a startup, we need to build a machine that builds startups.

To make this model work efficiently and cost effectively we need a lot of volume.

At the top of the funnel we need a large pool of capable founders applying to be part of the program.

Y Combinator now has tens of thousands of companies from all over the world applying to be part of each intake, making it extremely competitive. They can accept the very best founders into their program.

By comparison it’s common to see local accelerator programs pleading for applications and taking more-or-less everybody they can get.

We also need a big group of mentors and investors who can add value to the ventures during the program and fund them once they graduate.

The mentors at the established programs overseas are predominantly alumni, who have deep relevant and recent experience, paying it forward. And investors are normally queuing up to attend demo day.

In New Zealand the people involved tend to be corporates, angel groups or people whose main experience is running accelerator programs. These programs promise founders “unparalleled access” to the “best technical and business talent in New Zealand”.

Is it, though? ⁸

Anybody applying to these programs should look closely at the people involved and make sure there is an overlap with the skills they need and the problems they have.

To make it even harder for the local programs, they are competing in a global market. The best accelerator programs overseas are happy to take founders from anywhere. The best founders are naturally going to be drawn to the best programs - and there are many examples of Kiwi founders who have taken advantage of that opportunity.⁹ We’ve also recently seen Australian programs such as Startmate run dedicated intakes for New Zealand founders.¹⁰ This further reduces the size and quality of the pool of founders available to local accelerators.

Unfortunately, we can’t just replicate models that are successful overseas and expect them to produce the same results, unless we can identify our own competitive advantage. Comparing our local accelerator programs to the international equivalents is a bit like comparing Rainbows End with Disneyland.

Perhaps a better model for us would be one inspired by the philosophy printed on the back of every early-model iPhone:

Designed in California. Assembled in China.

Our strength is not mass assembly!

Rush to pitch

We can also think about accelerators at an individual company level.

Great companies take a long time to build. That’s actually a good thing, because we learn as we go. Too much attention too soon can ruin a new venture just as badly as no attention at all.

Watching accelerator programs rush to make early-stage companies investment-ready often makes me feel like one of the old guys in the vintage Mainland cheese ads, sitting bemusedly as younger cheese-makers lose their patience and quit.¹¹

I’m not convinced it helps founders at all to compress the development of a startup into an arbitrary period for the benefit of advisors and investors.

Many of the startups I’ve seen coming out of local programs chasing investment way too soon remind me of young kids whose parents push them into competing in pre-pubescent beauty pageants.

They have spent much more time polishing and rehearsing their pitch to investors than on building a great company worthy of investment. It’s like the difference between cramming for exams and real learning. Great pitches start with the words “we realised”, and those insights always take time.

Founders are encouraged to be “global from day one”. Most early-stage companies who boast about this are dead on day two. By comparison, our most successful companies have often expanded from market-to-market rather than trying to boil the ocean all at once. There is a big difference between having a smattering of customers in many countries (as is common for early-stage SaaS companies, for example) and having solid sales and distribution teams in multiple markets around the world. Those teams need to be built carefully and methodically. Again, that takes time.

We need to think about what it costs us to optimise for speed.

One shot for glory

Many years ago, before drones were as reliable as they are today, I was taught a great mindset by a remote controlled plane enthusiast. He explained to me,

We fly three or four mistakes high …

In other words, when (not if) something unexpected happens and the plane suddenly drops, the plan is to be flying high enough that this doesn’t immediately cause it to crash. There is tolerance for mistakes.

This is not how accelerators are set up. For founders the “demo day” is often a do-or-die moment. This begs the question: who are these programs really benefiting?

People running accelerator programs will generally explain this away by referencing the benefits of failing fast.

The whole idea behind “validated learning” (first popularised by Eric Ries in his book The Lean Startup, now a bible of accelerator programs) was to encourage founders to create feedback loops in order to avoid failure. The most vocal adherents of this philosophy get it completely backwards.¹²

If we bring this back to our impact assessment questions…

The same people who promote these programs to founders by saying “in just 12 weeks you will prove if your venture works or not - fail fast!!” will typically squirm when asked how quickly they can show if their accelerator program itself is working and will make the excuse “it takes 10+ years to build an ecosystem”. That’s an excuse masquerading as a boast.

It’s not enough to justify these programs on the basis they train a small group of inexperienced founders in a customer discovery process. The cost per person is too high and the collateral damage is too great.¹³

At some point every founder needs to meet the dangers of the real world. Eventually plans need to get punched in the face.

But any accelerator program encouraging founders to be cannon fodder and fly one mistake high just because investors are waiting to pass judgement in a few weeks’ time is a bad deal.

Just as with incubators, the measure of an accelerator program is also easy:

How many startups achieve with escape velocity and go on to thrive?

Velocity made good

By funding incubators and accelerator programs like this for many years we have thrown a lot of spaghetti at the wall. But how much has stuck? ¹⁴

We don’t really know, and need to do better.

Perhaps we can take inspiration from the sport of sailing. One measure they use when comparing performance of different boats is velocity made good (VMG).¹⁵ This reports the speed of the boat, but only counts the progress towards the actual destination. How much of the effort we’ve put in has moved us closer to the desired outcome?

We need to be honest about where our best startups originated. When we consider all the companies that have achieved the scale of success we’re hoping for, none of them started in an incubator or accelerator. It’s disingenuous for people promoting derivatives to shine the spotlight on these successes without acknowledging their own contribution.

We need to understand the reasons why some startups are successful and others are not. We should be able to prove that the specific things we’ve done have made the difference. We should be able to link the inputs with the outputs, and attribute success.¹⁶ This will require better feedback loops. In some cases this will be the first time these have existed, and that will be a shock.

When we promote derivatives we need to focus on the constraints and problems that people working on startups actually have, and work backwards to the interventions that might help them. Too many derivatives have been designed from the top down. Far too often, the primary benefactors of these derivatives are the people running them or funding them, rather than the founders and teams they are intended to support.

When we ask how the government can help, we should prioritise companies at the high-growth stage rather than at the early-stage. This is when startups stop having an innovation challenge (needing to get one new thing right) and worrying about their survival, and start having an execution challenge (needing to get thousands of details right at the same time) and worrying about how to scale. This is when system-level support can help. There are repeated patterns and it’s crazy to have founders each make the same mistakes as they go through this phase.

When we design programs to incentivise fast growing companies they need to be as transparent and as low friction as possible. They should reward the things we want directly (e.g. export revenue and productivity, measured by high revenue per employee) rather than factors we hope will contribute to these outcomes in the future (e.g. R&D spending). Life is hard enough as a founder without having to deal with cumbersome grant applications and slow processes overseen by people with little or no direct experience.

(A simplified system of startup welfare would free up many hundreds of people who currently work in the administration of that system to move into roles working directly on startups)

We have to be honest about how our activity translates into progress and stop funding programs that generate lots of new early-stage companies without ever converting them into high-growth companies. More and more derivatives pumping out low quality startups don’t get us any closer to the goal, and burns through too many people. We need to remember that the ecosystem grows one company at a time, when individuals working together come up with a new idea and grow that into a great business.

The constraint is fundamental, not abstract.

We now have too many preachers and not enough prophets. Too many people raising awareness, too few working miracles.

Remember, there is a method to scale, from first experiments to genuine system level growth. First we crawl; then we walk; then we run.

When we try to create an ecosystem by investing in derivatives it’s even more important to understand this progression. First we create one great company; then we do it again; then we do it multiple times at once. The sequence is important, because the lessons from each stage provide the foundation for success in the subsequent stages.

Nearly everybody who talks about wanting a larger, more vibrant and more successful ecosystem misses this. They aspire to create multiple companies at once but don’t really know what it takes to create one. The companies that are created by this system are exactly what we would expect as a result: fragile, precarious, and underwhelming.

We need to let each company grow at the pace that is best for them. We need to stop pretending that we can accelerate the growth of the ecosystem, just by being impatient. Growing a large, successful company takes a long time, and nearly always longer than founders and investors optimistically predict. We should be suspicious of anybody who has a fast-track or short-cut to sell.

But we should stop using patience as a convenient excuse to continue with current initiatives, hoping that it’s only a matter of time before they start to work.

A dangerous idea has recently taken root: because it will take a long time for the startup ecosystem to grow there is no easy way to measure in the interim how we are tracking, even though in some cases we’ve been doing these things for many years now.

I reject that. Sooner is over. It’s later already!

Just because it’s hard to measure doesn’t mean we shouldn’t try. It’s lazy to assume that because measurement is hard and attribution is even harder that the question is impossible.

If there is no way to show it’s working then it’s not an experiment. It’s just wishful thinking.

Are we too scared to look too closely at results we’ve gotten from all these derivatives in case it forces us to admit that we’re not really making as much difference as we hope?

The reality is many of the initiatives we’ve tried have not worked. That by itself shouldn’t surprise us. We’re dealing in uncertainty, and that requires experimentation. The only way to know if something might work is to try it. But that approach only works if we are clear in advance about our thesis and identify the measures we will use to demonstrate that it’s working (or not).

This is really not that hard:

1. Have a hypothesis,
2. Test that hypothesis,
3. Update the hypothesis based on the results.
4. Repeat, until we get the outcome we want.

Remember, there are only three ways to be wrong. We have to admit when things have not worked, and to stop and try something different.

Real failure is not learning and improving on the next iteration. So at the moment, we’re mostly failing.

We don’t only need more startups. We need a higher conversion from early-stage to high-growth. Of course we need to be patient. But we also need to create much shorter feedback loops, honestly measure our progress against regular milestones, and stop repeating things which haven’t worked, hoping that next time will be different.

Integrate

You might read this and conclude I’m exceptionally negative about the startup ecosystem. A number of people have said that to me over the years.

Actually the opposite is true. I couldn’t be more optimistic about the progress I’ve seen first hand, especially in recent years. I’ve had the opportunity to work with exceptional founders and invest in some amazing companies through their early stages. Some have thrived and gone on to become successful high-growth ventures. It has been massively rewarding, both personally and for many other people who worked on and invested in these ventures. It’s how I’ve filled my days for many years.

I’m also very excited about what comes next. The multiplier effect from successes like Trade Me, Xero, Vend and Timely is huge and probably still underestimated. Unprecedented amounts of capital and expertise are recycling directly into the next wave of ventures. From my perspective, the future is bright.

Which just makes me doubly sad that so many people continue to put so much effort into various derivative programs, and then wonder why their results don’t quite match.

In our desire to create a vibrant ecosystem of startups in New Zealand we forgot that what this really means is creating more successful startups. We can layer on as many derivatives as we like, but it will make little difference otherwise. The underlying ventures and specifically the people working on them are the limiting factor.

We should pay attention to people talking about a startup they are working on. Think about what we can all do to help them with that specific venture. This could span from very high-fidelity things such as working with them directly as an employee or adviser or investing cash right through to lower-touch but sometimes still impactful things like giving them our perspective on a problem they currently have, based on our experience.

But if anybody wants to talk about the ecosystem as a whole … run!

The biggest contribution we can each make to building an ecosystem of great startup companies in New Zealand is to focus on creating one great company. That’s hard enough and will probably require our full attention.

If you’re a founder, try not to be distracted by the ecosystem too much. Put on your own mask first, and then you’ll be in a much better position to help others. Don’t worry that there are not yet enough startups in your city or in your country. The best way to solve that problem is to try to make your own startup one that counts.

Be aware that everybody who participates in the startup ecosystem has a business model. Just as your own venture has a business model. It’s okay to stop and ask if others’ business models make sense for you and your venture. Often they won’t. You should not feel obliged to spend your precious time or money on anybody else’s venture, just because you are both part of the same ecosystem.

If you are taking investment from a venture fund, think about their business model. If you are applying to an accelerator program, think about their business model. If you have a “cheap” desk in a shared working space, think about their business model. If you are working with advisers or consultants (especially people who are inclined to peddle the idea that failure is a glorious and exciting step on the path to inevitable success), think about their business model. If you’re pitching to a room full of investors, think about who is getting paid (i.e. who sold the tickets, or who is clipping the ticket on any capital invested), what they promised you and what their business model is.

There are two famous rules of thumb which most likely apply:

1. If you are not paying for the product, you probably are the product; and
2. If you look around the table and you don’t know who the sucker is, then it’s probably you!

If you’re an investor, try to contribute more value than you capture. Stay humble about how much difference you can make. Show you can do it once. Then do it again.

If you are currently involved in a first, second or even third or fourth derivative capacity, try to reduce the abstraction. Consider instead working directly for one of the companies you’re supporting. It makes me sad to see smart people trying to “build a startup ecosystem” when existing successful startups within that ecosystem can’t hire people they need. Failing that, be more willing to investigate if your derivative actually assists.

Don’t kid yourself into thinking you can accelerate multiple new startups every few months. Don’t pretend creating safe spaces for startups to start themselves is sufficient. Don’t just raise awareness and hope that inspires somebody else to do great things. Get as close to the source as you can.

If enough of us work directly on one company that achieves its potential then the meta problem will solve itself and we’ll all be much better off.

Related Essays