40 tips for building a startup after 19 years of success and failure

I’ve been working on startups in one form or another since about 1999. I started hanging out in Silicon Valley informally in 2008, and moved formally at the beginning of 2009. After building a number of startups, products, brands and ecosystems my journey so far has culminated in heading up product for the Uber Developer Platform and now, helping growing startups as a Strategic Advisor.

In Silicon Valley there’s a particular flavor of startup that gets built. One where you “raise money, grow fast and make a big impact”. “Success” is about being validated by a large market (most important), investors (less important) and “making the world a better place”. By this measure, I feel like I’ve had some share of big wins and great outcomes. More importantly, though, I’ve had the opportunity to make and learn from a lot of mistakes. Along the way I’ve also advised many startups, as well as payed close attention to the broader startup ecosystem.

With the benefit of hindsight, I’ve observed a number of “obvious” mistakes that I and others have made — so I thought I’d catalogue as many as I could here so that others might benefit.

Before I get started — a few things to keep in mind.

Perspective: What follows are some of my personal, high-level and incomplete tips resulting from this experience. Although I’m very technical, I am not a practicing engineer, so these tips are written from the perspective of a “non-technical co-founder”.

Ordering: I’ve written the advice more or less in the order in which you might need it, and in a fairly stream of consciousness way. Some of the tips might seem cliche or obvious to those “In the know” — but in my experience they are easy to ignore, misunderstand or underestimate when you’re in the thick of following your dream and “thinking different”.

Exceptions: As always, there are exceptions to every rule and many of the individual tips in this list are not mandatory on their own. However, if you start skipping too many of them, it will really undermine your chances for success.

Crypto: This is from the perspective of a more traditional VC backed Startup. Crypto startups and ICOs tend to play by a different playbook.

History: This is a repost of an older post. I’ve updated it to reflect the latest reality and my most recent learnings.

Before you even get started

Be in a major tech hub

Yes you can succeed in other places, but the chances of any business succeeding are so small, why start with a disadvantage?

Some argue that the ecosystem in their area is “getting better” or they’re “going to help build the ecosystem along the way”. Trust me, as someone who’s tried, It’s hard enough building and driving the train, you can’t lay the tracks at the same time!

Silicon Valley is the obvious choice — but it’s getting increasingly expensive and the US political and cultural climate is spinning out of control. More recently China is becoming very interesting as well. If you’re doing something in Crypto then somewhere like Singapore is an option too.

Have a big vision

It’s your master plan for how you want to change the world

Now forget your big vision

Or you’ll risk trying to boil the ocean. Instead…

Figure out a very specific, painful, real problem

Find a problem where there’s a significant number of people or companies that would materially benefit from a solution. It’s essential that the problem is one you are personally experiencing and/or are super passionate about solving. Once you’ve figured out the big problem, figure out which very specific part of the market (people, place, situation) is experiencing this problem most acutely. That’s where you should start. (Check out this great post about defining your problem from my friend Marius)

Find people who are uniquely qualified

This is one of those cliche pieces of advice that’s easy to ignore as you band together with your friends to just build something. The fact is there’s a huge gap between great people who win and mediocre people who don’t. Find great people to execute on your idea with you. Great is hard to quantify here, but it includes curiosity, humbleness, hunger, high-EQ, team spirit, self-motivation, discipline and, ideally, high levels of skill and experience in the given problem domain.

A big, sobering question to also consider is: Why are you qualified?

Make sure you have a technical co-founder who knows what they’re doing

If you’re building a tech/software product (and you should be) you can’t outsource the core technical function. You must earn (not find) a technical co-founder by doing lots of wireframes, mockups or even having a minimum viable prototype built from an elance contractor. But before you kick into high gear, you need to have a technical partner on-board who has significant buy-in.

The word ‘earn’ is important. There are lots of posts that cover this in more detail but to summarize: Since great engineers are so sought after in the Valley and around the world, you need to essentially treat them like investors — prove you know what you’re doing and earn their investment in your startup through as much supporting material and momentum as you can.

You’re going to have high, highs and low, lows. It’s a roller-coaster

As the sunscreen song says “Whatever you do, don’t congratulate yourself too much, or berate yourself either. Your choices are half chance. So are everybody else’s.” — stay calm, focused and deal with the most pressing issues one by one. You set the tone and tempo of your team with every word, action and attitude.

It’s likely going to take many, many years

A startup is not a get rich quick scheme. Succeed or fail, It’s likely going to take about a decade of your life. Thats why you should make sure you’re working on a problem you’re passionate about. All of the “overnight successes” you hear about likely…

  • Took many, many years toiling in obscurity before they ‘took off out of nowhere’
  • Had a team with a strong background in previously successful startups
  • Followed every tip in this list (and many more) flawlessly
  • Got very lucky building the exact right thing for the exact right market at the exact right time (e.g. lightning strike)
  • Some combination of all of the above


Figure out a very specific, simple, discrete solution (read: product) for a specific, well defined customer

Ideally your product should start out by providing a few simple interactions/behaviors. Ask yourself “What is the ‘press a button to create joy’ version of this?”

Your solution should be software

If you are providing a service selling your billable hours then you are not a Silicon Valley style tech startup; you are an agency of some kind. Uber, for example, provides logistics software and a two sided marketplace. The drivers are partners.

If your product is B2B, try to avoid heavy, high-touch sales processes

Provide an easy way for employees to onboard themselves and then up-sell them through the UI or a bullpen of interns dialing for dollars — See Slack

If your product is B2C you may still need a B2B plan

Even B2C startups need to make money (shocking I know!). So remember that once you hit scale, you will need to do some combination of…

  • Selling to brands, distribution partners and advertisers (See Twitter, Facebook, Snapchat etc — in this case just note that you can’t avoid a ‘sales’ team)
  • Charging for the app (easier for mobile apps)
  • In-app ads or purchases (easier for mobile apps)
  • Charging a fee for ‘pro’ features
  • Charging transaction fees

Avoid propping up incumbents

Uber didn’t focus on helping the existing taxi system, instead they disrupted taxis by finding their own pool of underutilized drivers (in the form of black cars and every-day people). Facebook and Twitter didn’t work with media companies — they built their own media companies and then forced traditional media brands to play by their rules.

Ask yourself how you might build a full stack startup and disrupt existing slow, fat, inefficient and ineffective vendors…

Choose your platform carefully

By platform I mean companies with the vested interest in the end-user’s eyeballs that you might be using for data and/or distribution. Think about the Twitter API (remember all those Twitter clients that got disrupted by Twitter’s developer rules, acquisitions etc), Facebook Pages (Remember Causes, Zynga etc) and even MySpace back in the day (remember all those crazy widgets?). Building on top of someone else’s (very successful or rapidly growing) platform might be ok as a way to bootstrap — but look out for cannibalization from the platform. Ensure you have alignment, access to their team and — better — a way to syphon value to your own property

Mobile first

Mobile is very important right now — start there. Mobile apps offers easy access to distribution, monetization and the end-user who will be carrying you around in their pocket.

Try to avoid any initial reliance on network effects

Your idea will need a great ‘single-player mode’ to avoid a ‘cold start’ catch-22 problem. This doesn’t mean your product can’t benefit from network effects and viral mechanics, but it must (ideally) provide enough value without them so that it’s immediately useful and sticky for your initial set of users.

Make sure you have enough money and a guerrilla marketing campaign

You will need to find and activate passionate groups of users that take notice in a hungry and addicted way (passing interest isn’t enough). Particularly if you can’t avoid the need for network effects, it will be essential to find clusters of users to onboard in large enough groups to break the catch-22. Workflows that start with on boarding a user’s address book can work, but often presents too much friction for an unknown new app.

Your first, specific, simple, discrete solution should not be a “platform”

If you think you’re going to start by building a container into which others pour their creativity — you’re likely very wrong because…

  • Platform suppliers (e.g content creators, app developers, drivers, sharers etc) only pour their creativity into platforms that provide massive utility (e.g. make it 100x easier to create something beautiful) and/or provide massive distribution to an existing audience (e.g. make it 100x easier to reach large numbers of people)
  • If there’s an existing competitor with massive market penetration, even significant utility may not be enough to attract suppliers to your platform since they tend to gravitate towards momentum, winners and large existing audiences
  • Platforms most often emerge from massive companies OR from great products with passionate users that are then carefully opened up as a platform by exposing specific APIs, points of extensibility or supplier participation.
  • If your idea must be a platform on top of which 3rd parties will play, plan on doing ALL of the initial work yourself or have a pool of capital to bootstrap things because…
    a) You need to kick start the platform with great ‘seed content’ so that end-users (demand side) can start using it before you have suppliers contributing 3rd party value on their own
    b) Suppliers need amazing examples to be inspired by
    c) Most suppliers suck so you need large numbers of crappy 3rd parties in order to get just a few great ones

A good example here is AirBnB. They had to manually go and get the initial listings. Pintrest had to create the initial pinboards. To make this happen you will need people (possibly you!) dedicated and passionate to do this obsessively well. It costs money and takes a lot of time — so plan for it.

“There’s only two ways to make money — bundling and unbundling”

On mobile, bundling multiple features into one app is a bad idea. The current trend on Mobile is unbundling features into stand-alone apps. This is because mobile users use the OS itself as the ‘integrated environment’ and their home-screen as the place to choose from a palette of beautiful best-of-breed features (aka Apps). Instead…

  • Unbundle— Do one thing really well.
  • Do an old thing really beautifully. This alone can be a big differentiator.
  • Make a really complicated thing really easy by implementing better operations and smarter technology. See Uber.

Make it easy to adopt

Really easy. Self-serve. Up-front pricing. Free-trials. A big ‘go’ button.

Make it easy to understand

Ideally it shouldn’t take more than a one page website to explain and convert users for your first version.

Get users to the magic moment as fast as possible

Getting users to that moment where they say to themselves “Ooooh… that’s awesome!” when using your product is the most important thing you can do. Even better, get them to that moment and then provide a meaningful reason for them to share. This is essential. And it’s also very hard to do.

Wire your product for metrics

How many page views, how many downloads, how many signups, how many users tried feature 1, 2 and 3? See: Mixpanel and Intercom to make this as easy as possible.

Product/Market Fit

Ship it.

No seriously. Ship. It.

Get it into customers hands

I don’t mean do a big launch. I mean 1:1 user research and hand holding. Lots of it. Manually if you have to. Go meet your users. Learn from their confusion, feedback and advice. If they all start saying “Oh, I need this right now” then you’re onto something. If they say “Oh, maybe [insert some other customer/market here] could use this”, then they’re just being polite. They can’t figure out why they would use it and they are trying to be helpful by suggesting a home for idea.

Look at the metrics

The initial product and metrics will suck. Refine. Ship it again.

Keep iterating

Keep refining until your initial user acquisition and retention rates trend up and to the right. Keep going until more and more people in your target market say “I need to use this right now” and actually use it.

User support is massively important

Listen carefully, look out for trends, respond quickly (both to the specific inquiries and with broader product changes).

Learn how to say ‘No’

Or ‘Later’. Because customers, partners, employees and your own entrepreneurial spirit are all going to demand you do a lot of different/related things and you need to understand what’s inside and outside of scope for your product at this stage of its development. Focus.

Iterate, iterate, iterate

And then iterate again.

Have a plan (and enough money) to promote the hell out of it

Once the metrics (acquisition and retention rates especially) are trending in the right direction (and ONLY when they are), you need the resources and skills to do a big launch and promote in a sustained, long-term way (this is why you need Angel and Series A fundraising). Promotion tactics might include PR, Sales, Bizdev, Content Marketing, Influencer outreach, Events (Sponsorship and speaking), Social Media Marketing, SEO, SEM, Inside Sales etc.

Avoid just dabbling a little with each of these and then giving up. They can feed off each other. Each channel takes perseverance and it own iteration and experimentation. None of them are a silver bullet and you will eventually need to ramp up to doing a mix of many of them at the same time.


Raise an Angel round. Don’t skip it

Angels can provide a strong foundation of contacts and peer influence. Don’t skip them because…

  • Angels invest in great teams. At this stage your idea is just an idea — so they need to know they can bet on YOU to figure it out. So lead with your team in the pitch. Why are you uniquely qualified to do this?
  • Angels also want to be part of something that’s got early glimmers of traction that has the opportunity to become the next big thing
  • Don’t pretend like you have it all figured out or you’ll just come across as full of shit
  • You just need to show that you’re asking the right questions and have concrete ways to go find the answers with their money
  • You should always raise at least 18 months of runway — because you will need to start fundraising again about 6 months before it runs out
  • You will need more money than you think — don’t be too conservative in your expenditure projections. Don’t be too optimistic in your user and revenue projections.
  • If you’re really good, a clever trick is to include Angels and Advisors in your Series A rather than splitting them into two separate rounds — less dilution.

Unless you’re a rockstar, you can’t raise an Angel round until you’ve done all of the above

YC, 500 Startups, Techstars are also fantastic options but don’t let them be gatekeepers to your success.

Treat the Angel round like a roadshow

  • Meet all the Angels you can. Get commitments. Getting a lead investor is the hardest step here.
  • Angels often invest in packs — try to find a great group of Angel friends who will make intros to each other.
  • Make sure you define the metrics that matter in your business with your Angels so that you’re both paying attention to the same numbers. Focus on them. Move them.

Make sure you have plenty of time to raise your Series A from a VC

Start about 6 months before you run out of money. Once again you will need to raise enough money for at least 18 months of runway. Once again don’t underestimate your costs and don’t overestimate your growth/revenue.

VCs are people too

  • Your Angels can/should help make the intros
  • Talk to them early
  • Build a relationship with awesome partners and great firms
  • Listen to their objections carefully and find a way to address them
  • Ask about the metrics/proof points that would matter to them for a Series A (learn more here)
  • You will need to talk to a LOT of them
  • Remember they are not just ‘bankers’ or ‘money men’. They want to feel qualified to help you beyond just the money

Most of their objections will be around Team, Technology and Traction and Market

  • Team: Why is your group uniquely suited to this.
  • Technology: Why is your stuff better/faster/cheaper.
  • Traction: You need to show that you have Product/Market fit by showing your User Acquisition Costs and Retention Rates. They should feel comfortable that pouring money on the ‘problem’ will result in rapid growth.
  • Market: How does this become a billion dollar business?

Hiring/Growing your team

Hire slow and fire fast

This is a cliche/classic recommendation — but it’s serious.

Don’t just hire whoever’s in-front of you. Cast a wide net. Actively reach out to amazing people. Build your dream team.

Vet potential hires thoroughly for culture, expertise and vision alignment. Of course this doesn’t mean that you hire ‘yes people’ — you want people who are smarter than you who can help make big strides in their respective roles.

Don’t avoid firing employees because it feels bad. Fire them QUICKLY if they are not fitting. The wrong employee can burn a LOT of time and capital and destroy morale if they are a bad fit for your team.

Hire more women!

I’m not a believer in quotas or shoehorning people into roles just for the sake of appearances, but watch out for subconscious biases and actively work against them.

Get great advisors

This one might sound self-serving, but a good advisor can help end a debate that could have lasted weeks or months and/or could save you from weeks or months of going down a blind alley learning that something doesn’t work the hard way. That being said — avoid people who’ve spent their entire career being advisors and never actually built and ran something themselves. Plug for the StartupSuperSquad here.

This is, of course, not the end of the journey — but it is the end of my personal experience. Next steps involve further fundraising, IPOs, Acquisitions etc. More for me to learn!