It’s the one question always asked of startups – why won’t (insert market leader here) simply implement your great idea themselves, wiping you out in the process? A reasonable enough question given the billions of dollars and thousands of engineers available to them, compared to the typical startup’s resources.

It’s worth outlining the reasons why in fact Goliath has fewer advantages than you might think, as compared to David.

1. Innovation is a crap shoot

No one really knows that your great idea is great. That’s only found out when the revenue comes in, and by that stage, you’re well on the way to becoming the next generation incumbent. Venture capital firms (VCs) don’t invest in companies. VCs invest in markets, and people; specifically CEOs. They understand that in a hot market, they have a shot at the super-unicorn which will become the market leader, and a very good chance of a unicorn 2nd to 5th place player, that will deliver the returns needed to keep the limited partners in order. If the market is hot enough, you won’t lose your shirt even on the also-rans. Failures are buried, and history rewritten to emphasize the foresight of picking the 1-2 mega hits that pay the returns and the 20-30 ‘failures’ (and remember some may still be pretty successful by most people’s reckoning) simply forgotten.

It’s a portfolio business (the technical term for ‘crap shoot’), and one most VCs only play with a fund size of $100-$500M. You can’t ‘just’ invest $10M in Facebook circa 2004 and be done – it’s not that easy – you need to invest in a wide portfolio, and that takes serious amounts of money that even market leaders baulk at when trying to run their own innovation program.

This model also requires Darwinian culling – something VCs can do easily since all they have to do is simply stop returning calls when the cash runs out. It’s much harder for a large company psychologically, or practically, to fire an under-performing R&D team. Google is often quoted as a counter example, and certainly they end-of-life projects with impressive regularity, but even Google have shown far more patience with Google+ than any VC would. Large companies are like mammals with their young. VCs parent like fish, and whilst it’s cold-blooded (no pun intended), it’s more effective.

Large companies are not competing with one $5M startup, they are competing against a $B ecosystem, which is why so often it’s the startup that comes out on top. Large companies cannot run projects as portfolios like VCs can.

2. Everbody is loss-averse

Nobody wants to cannibalize revenue. Startups don’t have any revenue, ergo, they don’t worry about cannibalization. It’s not easy to walk away from stable, predictable (even if declining) revenues. By comparison, this latest new-fangled widget may or may not deliver in the market. Better to invest in an upgraded line of mainframes…

3. Large companies are process bound

Everyone can say no, and no one can say yes. Or more charitably, there is a need to get ‘buy-in’ across the corporation, where ‘across’ can be departments scattered over the globe. There is an inherent inertia, built around middle management more focused on managing their own pension trajectory. Upper management is often too distant from the market realities to identify up-and-comers. And people at the bottom just do as they are told. Or go and form their own startup…

4. Large companies need large projects

All large companies fundamentally become conglomerates. When you are the market leader, the only place to grow is another market. If you are delivering several $B/quarter, then only large projects can move the needle. Sure, a startup growing at 100% pa looks impressive, but the additional $1M revenue would be just a rounding error.

Where large companies do focus in innovating is in the area of large capital intensive bets – think Google StreetView, something which would be hard for a startup to compete with.

This also impacts growth strategy. From the early 90s Oracle stopped being a database company, transitioning to the full-suite enterprise IT provider we know today, delivering hardware, software, and services. People often confuse the fact that SalesForce is the incumbent market leader in CRM, with the belief that SalesForce is still a CRM company. Salesforce today is a Platform-as- a-Service company looking to expand out their footprint with big data and everything else. They want to be Oracle for Cloud. With 85% of IT spending coming from enterprise customers, Marc Benioff could not care less about smaller customers nor the CRM startups earning their stripes with them, any more than Oracle gives a monkey about MongoDB.

5. Not invented here

Google offered themselves to Excite in 1999 for the princely sum of $1M – but Excite po-poohed the idea, since it was so simple to implement themselves. The funny thing is, they were right, but they just never got around to doing it. The rest, as they say, is history.

It’s hard as a big company product manager with access to 100s of developers to accept they could have missed something a couple a geeks with barely a pot noodle between them have managed. So usually they don’t.

6. Large companies aren’t large

Large companies have lots of departments. Each of which may or may not be that big. “But they still have lots of resources” some will say. Yes, but there are all doing things, and in terms of the actual skill sets needed, no, they don’t.

I joined Sybase (1500 employees at the time) from IBM (300,000 employees) in the 90s and was surprised that up-and-coming Sybase, which only did databases, had twice the number of database-focused staff than IBM – which did everything including project manage the delivery of military helicopters. I wasn’t surprised then when joining the startup Illustra (<80 staff) that we had more web database expertise than Informix, Oracle and Sybase put together. Informix eventually coughed up $400M for Illustra.

7. Time is money

Why, on earth, did Facebook not simply reproduce the 24 man-month effort that Instagram represented, rather than pay an eye-popping $1B???

It wasn’t an acqui-hire for 13 staff. Nor was it really about market share, since almost every one of Instagram’s 30M customers would already be using Facebook (845M at that time). And it certainly wasn’t about acquiring revenue – Instagram had none, nada, zero, zilch, $0.00.

The reason was that Facebook was convinced that this was the future of photo-sharing and would put their own business at risk (since Facebook, reduced, is really about photos). For all the reasons outlined above, they knew that it would take at least 12 months to replicate it, and in that time there could have been a significant reversal of fortunes between the incumbent and the startup. Buying rather than building means additional revenue, but at levels commensurate with the acquiring company’s existing sales and marketing channels – not the startup’s. This can easily justify apparently silly revenue multiples, since the comparison point is not what Instagram is currently doing, but rather what Facebook would be missing out on, were it not. With Facebook doing almost $3B in 3Q14, then paying just one month’s revenue for Instagram, it now looks like a bargain given its importance to the Facebook value proposition.

You can’t reliably grow an oak tree by picking just one acorn. You need a full tray of seedlings, to nurture them all, and hopefully at least one will turn out well. It’s actually the entire premise of the VC model, and it’s also the reason big companies don’t typically stamp on startups, since they don’t know which seedling will be the one to finally overtake them, and there are too many to snuff them all out.

Like all generalizations in business there are exceptions. Developing a feature for a market leader typically does not end well, but these should be clear to identify. The general rule though, is clear. Few startups are actually at risk from the incumbents they seek to displace. The greater risk is does any part of the market actually want what you offer?