Twitter CEO Dick Costolo on Innovation, Start-ups, Google, And More

The oft heard sound bite about Twitter CEO Dick Costolo is that he once was a stand-up comic.  The impression left is that he miraculously went from open mic night at Yuck Yuck’s to CEO of one of the world’s largest social media companies.  Nothing could be farther from the truth.  Costello has a degree in computer sciences, co-founded numerous tech start-ups, all with successful exits (including FeedBurner), and spent time in the trenches at Google before being named COO, and then CEO of Twitter.  And he did improvisational comedy, not stand-up comedy.  Big difference.

Last week Costolo was Chris Hardwick’s guest on the Nerdist podcast and spoke on a range of topics, including….

What he wanted to accomplish when he went to Google;

I don’t remember.  I don’t think about things that way.  The one interview question that I was always horrible at answering was the “where do you see yourself in three years?” Since college, whenever I would get that question, I would always think, “I don’t know.”  I can’t even see myself in three years.  I didn’t think that “now I was at Google I have to make sure ‘X’ happens by 2009.”  

It just so cool, such an amazing company, with so many crazy smart people there, that I was learning a ton, about everything.  About the way Google systems were built, about they way they thought about architecting systems, which I thought was really cool.  The way they thought about the twenty percent program at the time.  It was like ‘oh yeah’ that’s an interesting way of thinking about how you could go create new things.  It was awesome.  I loved working there.

What really goes on in the bowels of Google labs;

You would probably be disappointed to go into the actual lab, or lair, and discover that they are just working on a new type of winter coat.  Just a down coat, but half a gram lighter.

Running a start-up with structure;

I knew a guy who had been a CEO of a bunch of companies, he told me, “You know, once in a while people would tell me, well, I don’t want to have a management structure because the engineers and designers and product people here don’t want to be managed.”  And then he would go out into the hallway and talk to the actual engineers and designers and product people and they would say “I would love to have somebody give us a better sense of how we should be doing this and which direction we should be going. We kinda all have our oars in the water, but they are not going in the same direction.”

I prefer to build companies with more of a structure where people understand how what they’re doing fits into what everyone else is doing, and the context of all that stuff, opposed to, “those guys over there are doing something, and I’m doing another something.”

Is it always a good thing to scale up;

It’s not always a good thing to scale up.  I think that the big mistake people make is, “I have to do ‘X’ because I’m supposed to do ‘X’ and I’m supposed to do ‘X’ because that guy did ‘X’ and ya know, the people who invested in my company like the way that guy’s thing worked.  And they keep telling me, well, that guy did ‘X’ and so you should do ‘X.'”

Especially in Silicon Valley, we lionize these figures who everyone then says, well, that worked because person ‘X’ says “you should always do this, and so I’m supposed to do it that way.” Books are written about these guys, whoever it is; Steve Jobs, or Jeff Bezo’s, or Zucker, whoever.  I think you have to realize that doing what you think you’re supposed to do, instead of doing what you want to, is almost never a good idea.

Keeping innovation alive in a growing company;

That becomes a harder and harder challenge as the company gets bigger and bigger.  One of the things we talk about here as a company, all the time, is trying to make sure we encourage people to take bigger chances, and make bolder choices, and take bigger risks.  Because, I think that people, as the company gets bigger, start to think to themselves, without even asking anyone, “well I couldn’t do that.  If I did that, this might affect this other thing, and nobody would ever let me do that.”  And I try to make sure that people don’t feel that way.  I want people to always feel like they can suggest whatever crazy idea they want to suggest, and whatever crazy design idea they have–and let me worry about whether that is something we should do or not.  They shouldn’t be thinking about, or worrying about that.  I want to make sure that people can suggest whatever they want to suggest; I think you have to encourage that as the company grows, or people start to feel like, “I could never suggest that.  Dick would never go for that.”

 What his process is when an idea bombs;

We get together and look at it and say “why did we think that was going to be so great, and what’s the difference between that and where we are now?”  We try to look at what we planned to do, and what we thought was going to happen, and what actually happened.  Not in a post-mortem way, because that is a “what went wrong” way, and it’s not “what went wrong” it’s why did we think “this” was going to happen, and instead the “other” thing happened?

Human curation vs. algorithmic curation;

I continue to believe that there is a happy middle-ground in that world that we haven’t totally zeroed in on yet.  When we, we as in the “small Twitter,” we tried to really focus on human curation of the previous Olympic games as an event experience on Twitter, and we were only going to curate the best tweets from the best accounts, but it really lost the roar of the crowd that makes Twitter feel like Twitter. If just felt like a sterilized version of Twitter, and so you need some of the volume, and some of the roar of the crowd, and the “way out there,” serendipitous things that only the algorithms would find, instead of somebody just combing through all the best tweets of the Olympics.  But at the same time, it does feel like that right combination of the two, a curation tool plus an algorithm, is probably the right way to present those things to people.

People still, in many cases, can provide the best context to the right kinds of ways to expose information to the user.

Check out the full Dick Costolo interview (Nerdist Episode 450)

Brilliant stuff like this rains down like..well, rain, on my stream during the week.  If you want to get wet, follow me on Twitter and StockTwits.  You can also pick up my book Trading – The Best of the Best: Top Trading Tips For Our Times by clicking here.  All proceeds go to fight pediatric brain cancer.