What are the most important things to keep in mind when scaling a startup? originally appeared on Quora - the place to gain and share knowledge, empowering people to learn from others and better understand the world.

Answer by Rand Fishkin, founder of Moz, author, blogger, speaker, on Quora:

Given how challenging startups are, the high failure rate, and the fact that I've only ever built one personally, I'm going to include the important caveat that this list is definitely not comprehensive, and almost certainly not in order of importance. That said, I'll take my best stab:

  • Build a startup around your strengths, not your weaknesses. If you're a phenomenally talented engineer, but have never been great at the product or marketing side, you can bolster those strengths by hiring great people into the roles, but in my experience, there will always be weakness there. Rather than trying to fight it, work with it and be an engineering-centric company whose unique value proposition matches up with what you can do better than anyone else in the field (maybe that's better data or speed, reliability or infrastructure prowess, better features and more of them, etc). Companies that try to make a weakness into a pillar of why customers should choose them or why employees should join tend to struggle mightily.
  • Don't try to be all things to all people. It's very, very tempting as your startup has success to think "hey, there's a big market over there that's similar to ours, I bet we could serve them, too!" Then you go spend years building new products and features, doing market research, talking to those customers, launching and trying to get traction, only to find that you've been outgunned by competitors in your primary market and are still struggling to get traction in the new one. It's far easier to pick a target market and audience, delight them, and continue to provide more value to them in exchange for more revenue, rather than expanding into adjacent markets because you're bored or think there's more opportunity. I'm not saying it never makes sense, but most every startup does this way too early. Doing too many things fails. Focus wins.
  • As you grow, you need to hire more people faster... which sucks. Bringing people into an organization at a rapid pace is a recipe for cultural struggles, for slowing down, and for a high degree of tension and unhappiness among your team (even if everyone agrees that you need to hire and they're all excited to do so). A few suggestions:
    • Have people join the company in cohorts (e.g. all new hires in February start on Monday the 18th and go through two days of orientation together), which creates camaraderie between those folks and can build inter-disciplinary relationships (a powerful thing).
    • Build a cultural screening system into your interview process, and don't be afraid to reject great candidates because they don't fit with your organization's values. We do this at Moz, and have folks from a different team interview for culture/values fit (e.g. a new engineer might get interviewed by someone on the customer success team).
    • Anticipate and plan for new people to slow down the teams they join, before they speed them up. This means you need to create realistic estimates for project completion that factors in onboarding and ramp-up time (which can be months).
    • Fire poor cultural fits fast. Fire poor performance fits with strong cultural fits much more slowly (by which I mean -- give them more chances to turn things around). Most companies incorrectly reverse these two and miss out on great potential future talent, while keeping around jerks who drag down everyone's productivity.

There are hundreds more items that belong on this list, and I apologize that I don't have time to share them all. Hopefully these three will at least be useful to you.

This question originally appeared on Quora. -the place to gain and share knowledge, empowering people to learn from others and better understand the world. You can follow Quora on Twitter, Facebook, and Google+.

More questions: