Slava Akhmechet

I am a cofounder of RethinkDB — an open-source distributed database designed to help developers and operations teams work with unstructured data to build real-time applications. We're hiring.



57 startup lessons

13 Aug 2013

There are already very good lists of startup lessons written by really talented, experienced people (here and here). I’d like to add another one. I learned these lessons the hard way in the past four years. If you’re starting a company, I hope you have an easier path.

People

  1. If you can’t get to ramen profitability with a team of 2 – 4 within six months to a year, something’s wrong. (You can choose not to be profitable, but it must be your choice, not something forced on you by the market).
  2. Split the stock between the founding team evenly.
  3. Always have a vesting schedule.
  4. Make most decisions by consensus, but have a single CEO whose decisions are final. Make it clear from day one.
  5. Your authority as CEO is earned. You start with a non-zero baseline. It grows if you have victories and dwindles if you don’t. Don’t try to use authority you didn’t earn.
  6. Morale is very real and self-perpetuating. If you work too long without victories, your investors, employees, family, and you yourself will lose faith. Work like hell not to get yourself into this position.
  7. Pick the initial team very carefully. Everyone should be pleasant to work with, have at least one skill relevant to the business they’re spectacular at, be extremely effective and pragmatic. Everyone should have product sense and a shared vision for the product and the company.
  8. The standard you walk past is the standard you accept. Pick a small set of non-negotiable rules that matter to you most and enforce them ruthlessly.
  9. Fire people that are difficult, unproductive, unreliable, have no product sense, or aren’t pragmatic. Do it quickly.
  10. Some friction is good. Too much friction is deadly. Fire people that cause too much friction. Good job + bad behavior == you’re fired.


Fundraising

  1. If you have to give away more than 15% of the company at any given fundraising round, your company didn’t germinate correctly. It’s salvageable but not ideal.
  2. If you haven’t earned people’s respect yet, fundraising on traction is an order of magnitude easier than fundraising on a story. If you have to raise on a story but don’t have the reputation, something’s wrong.
  3. Treat your fundraising pitch as a minimum viable product. Get it out, then iterate after every meeting.
  4. Most investor advice is very good for optimizing and scaling a working business. Listen to it.
  5. Most investor advice isn’t very good for building a magical product. Nobody can help you build a magical product — that’s your job.
  6. Don’t fall in love with the fundraising process. Get it done and move on.


Markets

  1. The best products don’t get built in a vacuum. They win because they reach the top of a field over all other products designed to fill the same niche. Find your field and be the best. If there is no field, something’s wrong.
  2. Work on a problem that has an immediately useful solution, but has enormous potential for growth. If it doesn’t augment the human condition for a huge number of people in a meaningful way, it’s not worth doing. For example, Google touches billions of lives by filling a very concrete space in people’s daily routine. It changes the way people behave and perceive their immediate physical surroundings. Shoot for building a product of this magnitude.
  3. Starting with the right idea matters. Empirically, you can only pivot so far.
  4. Assume the market is efficient and valuable ideas will be discovered by multiple teams nearly instantaneously.
  5. Pick new ideas because they’ve been made possible by other social or technological change. Get on the train as early as possible, but make sure the technology is there to make the product be enough better that it matters.
  6. If there is an old idea that didn’t work before and there is no social or technological change that can plausibly make it work now, assume it will fail. (That’s the efficient market hypothesis again. If an idea could have been brought to fruition, it would have been. It’s only worth trying again if something changed.)
  7. Educating a market that doesn’t want your product is a losing battle. Stick to your ideals and vision, but respect trends. If you believe the world needs iambic pentameter poetry, sell hip hop, not sonnets.


Products

  1. Product sense is everything. Learn it as quickly as you can. Being good at engineering has nothing to do with being good at product management.
  2. Don’t build something that already exists. Customers won’t buy it just because it’s yours.
  3. Make sure you know why users will have no choice but to switch to your product, and why they won’t be able to switch back. Don’t trust yourself — test your assumptions as much as possible.
  4. Ask two questions for every product feature. Will people buy because of this feature? Will people not buy because of lack of this feature? No amount of the latter will make up for lack of the former. Don’t build features if the answer to both questions is “no”.
  5. Build a product people want to buy in spite of rough edges, not because there are no rough edges. The former is pleasant and highly paid, the latter is unpleasant and takes forever.
  6. Beware of chicken and egg products. Make sure your product provides immediate utility.
  7. Learn the difference between people who might buy your product and people who are just commenting. Pay obsessive attention to the former. Ignore the latter.


Marketing

  1. Product comes first. If people love your product, the tiniest announcements will get attention. If people don’t love your product, no amount of marketing effort will help.
  2. Try to have marketing built into the product. If possible, have the YouTube effect (your users can frequently send people a link to something interesting on your platform), and Facebook effect (if your users are on the product, their friends will need to get on the product too).
  3. Watch Jiro Dreams of Sushi, then do marketing that way. Pick a small set of tasks, do them consistently, and get better every day.
  4. Reevaluate effectiveness on a regular basis. Cut things that don’t work, double down on things that do.
  5. Don’t guess. Measure.
  6. Market to your users. Getting attention from people who won’t buy your product is a waste of time and money.
  7. Don’t say things if your competitors can’t say the opposite. For example, your competitors can’t say their product is slow, so saying yours is fast is sloppy marketing. On the other hand, your competitors can say their software is for Python programmers, so saying yours is for Ruby programmers is good marketing. Apple can get away with breaking this rule, you can’t.
  8. Don’t use supercilious tone towards your users or competitors. It won’t help sell the product and will destroy good will.
  9. Don’t be dismissive of criticism. Instead, use it to improve your product. Your most vocal critics will often turn into your biggest champions if you take their criticism seriously.


Sales

  1. Sales fix everything. You can screw up everything else and get through it if your product sells well.
  2. Product comes first. Selling a product everyone wants is easy and rewarding. Selling a product no one wants is an unpleasant game of numbers.
  3. Be relentless about working the game of numbers while the product is between the two extremes above. Even if you don’t sell anything, you’ll learn invaluable lessons.
  4. Qualify ruthlessly. Spending time with a user who’s unlikely to buy is equivalent to doing no work at all.
  5. Inbound is easier than outbound. If possible, build the product in a way where customers reach out to you and ask to pay.


Development

  1. Development speed is everything.
  2. Minimize complexity. The simpler the product, the more likely you are to actually ship it, and the more likely you are to fix problems quickly.
  3. Pick implementations that give 80% of the benefit with 20% of the work.
  4. Use off the shelf components whenever possible.
  5. Use development sprints. Make sure your sprints aren’t longer than one or two weeks.
  6. Beware of long projects. If you can’t fit it into a sprint, don’t build it.
  7. Beware of long rewrites. If you can’t fit it into a sprint, don’t do it.
  8. If you must do something that doesn’t fit into a sprint, put as much structure and peer review around it as possible.
  9. Working on the wrong thing for a month is equivalent to not showing up to work for a month at all.


Company administration

  1. Don’t waste time picking office buildings, accountants, bookkeepers, janitors, furniture, hosted tools, payroll companies, etc. Make sure it’s good enough and move on.
  2. Take the time to find a good, inexpensive lawyer. It will make a difference.


Personal well-being

  1. Do everything you can not to attach your self esteem to your startup (you’ll fail, but try anyway). Do the best you can every day, then step back. Work in such a way that when the dust settles you can be proud of the choices you’ve made, regardless of the outcome.
  2. Every once in a while, get away. Go hiking, visit family in another city, go dancing, play chess, tennis, anything. It will make you more effective and make the people around you happier.


Thanks to Michael Glukhovsky for reviewing this post.