Will my startup fail? It’s a question every entrepreneur obsesses over.

It’s nearly impossible to give general advice about startups. They succeed or fail for what appear to be extremely specific reasons. It’s maddening, and it makes prediction very difficult. Venture capitalists, who are highly motivated to be good at prediction, generally bet on team first, product idea second, and everything else (market size, competitive environment…) in some order after that. Even among VCs, there is little consensus.

I’ve been thinking a great deal over the last few years about how a start-up — specifically, a tech start-up — should bring its product to market. In what ways are start-ups like traditional, larger businesses, and in what ways must they operate differently? Articles are written in business and tech periodicals every day describing signs you’ll succeedearly signs you’ll succeedcharacteristics of startups that succeeded, and characteristics of entrepreneurs who succeeded. When reading such articles, it’s important to keep a clear distinction in your head between correlation and causation. Entrepreneurs, especially tech entrepreneurs, tend to be cut from similar cloth. They have a lot of similar characteristics. Listing “passion” as a sign of success is unhelpful. Sure, most successful entrepreneurs are passionate, but so are most unsuccessful entrepreneurs and there are many more of them. So passion, along with many other widely shared traits, could truthfully be said to be a much bigger sign of failure.

Other articles try to go beyond prediction and offer actionable advice. Specifically, they try to help entrepreneurs identify the “unfair advantage” that will cause you to be successful. Unfair advantage is an idea that’s been around for a long time. My father used to insist it was an IBM thing (he worked at IBM); he used to say that IBM wouldn’t enter a new business unless they thought they had a clear unfair advantage. More recently, the meaning has been muddied to include individuals rather than companies, and dubious unfair advantages (curiosity is an unfair advantage?!) more akin to personality traits.

Later on in this essay, I’m going to suggest an unfair advantage available to nearly any tech startup that, if not predictive of success, at least suggests a specific roadmap for founders. It’s not just correlated with success; it helps create success.

But first, some fundamentals. For all that’s written and said about it, for all the importance it plays in everyday life for individuals and nations, the science of business is fundamentally pretty simple: create value and capture the value you create. All of a company’s activities ultimately fall into those two buckets. Build a great product and grab as much of the market as possible before competitors do. Create value and capture value. Otherwise known as: product and go-to-market.

Some executives think people can be convinced to buy anything and that product is itself a derivative of go-to-market; it exists only to serve customer acquisition. Companies operating in mature industries, high up on the “commoditization curve,” tend to think this way. The danger is that if you think product innovation won’t be rewarded by the market, you stop looking for innovation. When it finally comes, you missed it entirely. A company culture that celebrates sales, and treats product even slightly contemptuously, will only ever make bad products. American automakers pre-1970s were like this, and it’s the reason they got clobbered by Japanese automakers. Textbook publishers tend to behave this way, and it’s why they have struggled to adapt to digital and largely ignored Open Educational Resources, an existential threat to their current business model.

Even if your product is great, though, it isn’t enough. Is your product finely calibrated for where the market is right now? Or are you too early? If so, your product-market fit will by definition be awry and you will have a hard time acquiring customers. Alternately, perhaps you’re too late? If so your competitors have already acquired a chunk of the market and you will have a hard time acquiring customers. It’s better to optimize product-market-fit with an imperfect product than to perfect your product and sub-optimize product-market-fit.

To be successful as a start-up, you have to be outstanding at both product and go-to-market. Being truly outstanding at even one thing in business is hard; being outstanding at two very different things is that much more so. And that already difficult path is made near impossible by the fact that what it takes to be outstanding for the first 5% of the market is not what it takes for the other 95%.

Even if you nail your go-to-market strategy at launch, you must quickly transition to growth mode. In a really terrific interview on the Andreesen Horowitz website, Marc Andreessen describes this phenomenon:

Every market has early adopters. There are early adopters for everything, and it’s kind of amazing that that’s the case. But there are always people… And a lot of product/market fit is the fit with the early adopters. And so you get these extremely enthusiastic people, who in a lot of cases have sought you out as the vendor, saying, “Wow, your thing is really cool. Can I please use it?” And that’s your sign of product/market fit. The problem is, the early adopters are only ever a small percentage of the overall market. And so a lot of founders, especially technical ones, will convince themselves that the rest of the market behaves like the early adopters, which is to say that the customers will find them. And that’s just not true.

At launch, your product appeals to early adopters, but the rest of the market can’t or won’t use your product until it contains some key feature they want. There are lots of possible features that the market is telling you they’re interested in, so you need pick and prioritize what to work on next based on what will help you expand fastest beyond the early adopters.

Similarly, your customer acquisition process for the first 5% of the market will have to change when you begin to pursue the other 95%, whose behavior will be different. Click rate and open rate numbers won’t be the same for the other 95%. Your entire conversion funnel will look different.

Financing strategy now becomes critical. What if one of your competitors raises a huge financing round and uses it to build features way ahead of you and/or to spend more on customer acquisition and acquire most of the market? Even if your product is slightly better, you can’t overcome that unless your market happens to be characterized by high churn (in which case you have other problems).

To become a big company, both your product and your go-to-market have to evolve, perhaps unrecognizably, which means you have to be outstanding all over again at two more tasks. To grow from a start-up to a large company therefore requires you to be exceptional at four discrete tasks. No wonder so many start-ups fail.

One strategy to simplify the problem space is to sell your company early to a larger company. Let’s call this the “modest exit” strategy. It cuts the number of times you need to be exceptional to two (possibly one), but limits the size of your exit. If you don’t take too much investment and adroitly manage your cap table, this will still be a life-changing exit for the founders, great for early employees, and a solid win for investors but probably not one that “returns the fund.” However, it’s not in your control. There aren’t that many likely acquirers, and the timing may not work out for any number of reasons.

To become a large company, then, start-ups must have a strategy to transition from launch product and at-launch go-to-market to expanded product and expanded go-to-market. That strategy perforce includes financing (or one of your competitors will raise a big round and use it to acquire market share rapidly). And don’t assume that you’re going to be able to grow organically just from revenues. That would require a self-contradictory scenario whereby you grow free cash flows like crazy but attract no competition.

By the time you can raise that big round, you may also be able to pursue the modest exit strategy. The problem is that after you accept that big round of financing, it becomes harder and eventually impossible to pursue a modest exit. If you raise $150m, as I did at my previous company Knewton, then you need to sell for more than that to make anyone happy. The more cash you raise, the more you reduce the set of good potential outcomes. One of the difficulties I faced at Knewton was that I had to raise the cash ahead of clear feedback from the market about our product-market-fit, go-to-market strategy, or even the size of the market. If you’re going to raise lots of money—and at some point you probably must if you want to become a big company—then it’s much better to do so after you’ve already gotten clear evidence that product-market-fit and go-to-market strategy are working, and the market you’re pursuing is certain to be large. Then you can raise with confidence. Even if growth subsequently starts to slow, or it becomes clear that the market is smaller than you thought, at least you have a strong revenue position from which to operate. You can streamline operations, become profitable, and you’re not at the mercy of events.

In addition to all the competition from other start-ups, there is also competition from large established players. They can quickly put 100 engineers to work on replicating your product and they don’t need to raise financing in order to spend money on customer acquisition. Plus, they have either direct or adjacent operations in the market, which will give them additional unfair advantages over you. And while startups can bleed cash to build a brand, big companies already have strong brand awareness.

In fact, large companies have so many unfair advantages over start-ups, and the degree of those advantages is so immense, that it can seem like a miracle when a start-up actually succeeds. Chris Dixon, who is a colleague of Andreessen’s, is quoted often as saying words to the effect of “good start-up ideas initially sound like bad ideas.” (Chris credits Peter Thiel with being the source of this idea.) If it’s obvious to everyone that it’s a good idea, everyone is already working on it, including lots of big companies. So it can help enormously if everyone else thinks your idea is terrible — until such time as you prove them wrong with marketplace traction and it’s now too late for them. An example from the early days of search engines is Google sending all their user traffic away from their site. The big portals like Yahoo and AOL thought that was crazy. Why would you want to drive traffic away from your site? They could have crushed Google; instead they gave them the time and space needed to, well, build Google.

If you should be so lucky as to have an idea that appears to be terrible but is in fact fantastic and you can prove it to be so with just a few million dollars in venture capital, then that’s awesome! You should drop everything and do that. This kind of camouflage from other competitors, large and small, gives you the time you need to develop your product and go-to-market (i.e., succeed at the four tasks) before competitors start to move in.

Camouflage is a nice-to-have. It makes success more likely at the four tasks of launch productlaunch go-to-marketexpanded product, and expanded go-to-market. The must-have is success at the four tasks. Anything that helps that along—like potential competitors thinks your idea is dumb until it’s too late—is a type of luck that helps you succeed at the four tasks. Another type of luck may be that you benefit from a classic Innovator’s Dilemma scenario: your product is at the low value end of the market, where bigger players don’t want to compete for low margins, so they cede it to you and you use that position to climb up the value-added ladder. Or, that other players who could compete are doing well enough at their own thing that they don’t realize your approach is much better until it’s too late (e.g., Myspace vs Facebook).

Every successful company got some luck of some kind or other along the way. For a start-up to become big means that, at some level, the existing big companies did something wrong—something so big that it more than compensates for all their unfair advantages.

So what unfair advantages do start-ups have over big companies? There’s only one: it’s the ability to iterate rapidly. Rapid iteration, based on real time market feedback, can give startups a decisive advantage in the four key tasks. Startups iterate in both product-market fit and go-to-market strategy. Successful start-ups capture a moment when the market is just right and ready for their new product. They iterate and improve the product. They keep adding features that attract ever-widening concentric circles of the market, from the core of the onion to every subsequent layer. They also iterate and improve their customer acquisition process, from optimizing their funnel to building community and building brand. They can do all this incredibly quickly—faster than its possible for big companies to move. They can take market feedback and deploy it to product a week later.

Marc Andreesen says, in the same interview, that customer acquisition can become a moat. Product moats, such as network effects are great. But by the time you have a product moat, you’ve won the market anyway. You win because of the moat you build around customer acquisition. Here’s what that means: you’re probably going to market using some mix of following channels: ad words, search engine optimization, Facebook, Instagram, LinkedIn, Twitter, Pinterest, YouTube, e-mail marketing, direct mail, tele-sales, and conferences. In the year after launch, you should have figured out exactly what works best. You’ll measure things — channel by channel — like customer acquisition cost, customer service cost, churn rate, and viral coefficient. If a channel has higher acquisition cost, but lower service cost and churn rate, it may well produce higher lifetime value customers. You need to know that. If a channel has a higher viral coefficient, (the rate at which customers bring along other customers) — perhaps because it’s a community like Facebook or Pinterest where customers naturally talk to each other about your product — then the lifetime revenue per user may not look better but that channel still grows your business faster than other channels. All of these core metrics, and the conclusions you draw from them, are your go-to-market trade secrets. After you launch your product, you need to test each channel and iterate, iterate, iterate.

Despite their advantages, in the real world big companies are almost physiologically incapable of iterating the product quickly. Even technology companies struggle to iterate product quickly once they’re large. This is because when change affects either large enterprise accounts or large numbers of users, companies become extremely sensitive to existing users. Big companies will almost always prioritize existing users over potential future users. Startups, since they’re growing from such a low base, intrinsically prioritize future users. This gives startups a structural advantage in product iteration.

And big companies don’t generally authorize departments to spend $100m acquiring customers at some new initiative. If anything, they’re more conservative than start-ups. They’ll allocate a few million to see what’s there. Big companies often have useful adjacencies — assets in other divisions that could help this initiative succeed — but they general don’t share those assets easily between divisions. By the time adjacent assets are deployed, it’s usually because the big company is already visibly losing. So big companies don’t efficiently iterate customer acquisition any more than they do product.

Rapid iteration is an unfair advantage in the tech world, but it is unimportant in a great many industries. It doesn’t matter a whole lot if my supermarket is rapidly iterating the product experience based on customer feedback. Customer acquisition doesn’t change much year-in and year-out for many bricks and mortar industries. But rapid iteration is everything in technology. Technology creates new industries all the time, where whoever is first to get product-market-fit right will win. There are no switching costs from one technology solution to another—at least, not until one company becomes dominant, and sometimes not even then (e.g., Match.com to Tinder). Many real world industries have switching costs due to location — like supermarkets and clothing stores.

If a core part of your product experience is your physical location, and if you are in a low-innovation industry, then rapid iteration is not very important and large incumbent players are protected. This is true of many small companies that are not technology companies. However, tech start-ups tend strongly to have the exact opposite of those conditions: physical location is irrelevant to the product experience, and innovation is rapid and meaningfully improves the product experience. And if it’s a new industry or a new take on an old industry, as tech is wont to do, then customer acquisition is a brand new process just waiting to be iterated and optimized.

So, while big companies have nearly all the unfair advantages, start-ups have one crucial unfair advantage of their own: the ability to rapidly iterate product and customer acquisition. For other industries, it may not help that much, but for tech start-ups it’s more important than all the other unfair advantages combined. The decisiveness of that advantage is proven again and again, whenever a start-up turns into a technology giant. Ultimately, whether they think of it this way or not, this is why founders and venture capitalists are so attracted to technology. They think of it as “fast growth.” But what it really is the ability for a small company to iterate its way into becoming a big company, because for one reason or another the already-big-companies couldn’t optimize either product-market fit or customer acquisition.