Why Google can’t build Instagram

12 Nov

Tonight I was talking with an exec at Google and I brought up the success of (they’ve gotten more than 500,000 downloads in just a few weeks) and asked him “why can’t Google do that?”

I knew some of the answers. After all, I watched Microsoft get passed by by a whole group of startups (I was working at Microsoft as Flickr got bought by Yahoo, Skype got bought by eBay, etc etc).

I told him a few of my theories, and he told me back what they are seeing internally. Turns out he was talking to me about these items because Google, internally, knows it has an innovation problem (look at Google Wave or Buzz for examples of how it is messed up) and is looking to remake its culture internally to help entrepreneurial projects take hold.

1. Google can’t keep its teams small enough. Instagram was started by two guys who rented a table at DogPatchLabs in Pier 38 (the first time I met the team was when Rocky and I did this video on Dogpatch Labs). The exec I was talking with said Google Wave had more than 30 people on the team. He had done his own startup and knew the man-month myth. For every person you add to a team, he said, iteration speed goes down. He told me a story of how Larry Ellison actually got efficiencies from teams. If a team wasn’t productive, he’d come every couple of weeks and say “let me help you out.” What did he do? He took away another person until the team started shipping and stopped having unproductive meetings.

2. Google can’t reduce scope like Instagram did. Instagram started out as being a far different product than actually shipped (which actually got it in trouble with investor Andreesen Horowitz, according to Techcrunch). It actually started out as a service that did a lot more than just photographs. But, they learned they couldn’t complete such a grand vision and do it well. So they kept throwing out features. Instagram can do that. Google can’t. Imagine you come to Larry Page and say “you know that new social platform we’re building? Let’s throw 90% of it out.” Google has to compete with Facebook. Instagram had to compete with itself. As to Andreesen: this is why lots of my favorite companies like GoPro or SmugMug never took any VC. The pressure to “go for the home run” destroys quite a few companies.

3. At Google, if a product becomes successful, will get tons of resources and people thrown at it. Imagine you’re working at Google and you have 20% time. Will you keep spending that time on a boring project that isn’t very cool? No, you will want to join a cool project like Instagram that’s getting love around the world and getting tons of adoption. If the Instagram team were at Google they’d have to deal with tons of emails and folks hanging outside their cubes just to try to participate. I saw exactly this happen at Microsoft when a small team I was enamored of started getting tons of resources because it was having some success.

4. Google forces its developers to use its infrastructure, which wasn’t developed for small social projects. At Google you can’t use MySQL and Ruby on Rails. You’ve gotta build everything to deploy on its internal database “Big Table,” they call it. That wasn’t designed for small little dinky social projects. Engineers tell me it’s hard to develop for and not as productive as other tools that external developers get to use.

5. Google’s services need to support every platform. In this case, imagine a Google engineer saying “we’re only going to support iPhone with this.” ( is only on iPhone right now. They’d get screamed out of the room) and they need to support every community that Google is in world-wide. I remember at Microsoft teams getting slowed down because they’d need to make sure their products tested well in every language around the world. Oh, some screens didn’t work because some languages are read right to left? Too bad, go back and fix it. Instagram doesn’t have those kinds of problems. They can say “we’re English only for now, and heck with everyone else.”

6. Google’s engineers can’t use any Facebook integration or dependencies like Instagram does. That makes it harder to onboard new customers. I’ve downloaded a few iPhone apps this week and signed into them, and added my friends, just by clicking once on my Facebook account. My friends are on Facebook, I don’t have a social graph even close to as good on Google. Instagram gets to use every system it wants. Google has to pay “strategy taxes.” (That’s what we called them at Microsoft).

7. Google can’t iterate in semi-public. Weeks ago Kevin showed me Instagram and loaded it on my phone. He asked me to keep it somewhat quiet, but didn’t ask me to sign an NDA. He also knew it would actually help him if I did leak something about Instagram (I didn’t). What he really needed at that point was passionate users who would try it out and give him feedback about what worked and what didn’t. Bug testing. Now Google will say “we eat our own dogfood” but the reality is that you need to get people outside of your company to invest some time in you. Google can’t do this, because it causes all sorts of political hell. Instagram has no political problems to worry about, so was free to show it to dozens of people (when I got on it there were already hundreds of people who were using Instagram and I had it weeks before its official launch). I saw tons of bugs get fixed because of this feedback and those early users were very vocal believers in the product.

8. Google can’t use Eric Ries-style tricks. Eric’s “lean startup” methodology advocates making sure that customers want something, before going on and building infrastructure that scales. Google, on the other hand, has to make sure that its services scale to hundreds of millions of people before it ships a single thing. Google Wave failed, in part, because it couldn’t keep up with the first wave of users and got horridly slow (and that was even with an invite system that kept growth down to a reasonable rate).

So, how does a big company innovate? Well, for one, Google can innovate by buying companies like Instagram. For two, Google can use its strength in places where small companies can’t dare to go. For instance, building autonomous cars (I have a video with Stanford’s Center for Automotive Research that shows how these cars work and you can see that building stuff like that takes teams bigger than two people. Although to demonstrate that Google gets the power of small teams, Google’s car’s algorithms were mostly approved by just one person, I’ve learned).

Another way? How about open source? Build a system so anyone can code and add value without sitting in meetings and things seem to take off. At Rackspace (the web hosting company I work for) we’re noticing that with OpenStack, which is already seeing some pretty cool new innovations (coming soon) added by people who aren’t even working at Rackspace. As I look around the coolest companies in the valley, like Cloudera, I see the same mentality in place: they know they’ll get slower as they get bigger, so they are trying to build systems that let innovative, entrepreneurial, developers add value without getting caught in the politics of a bigger company. Take it outside of tech, look at TEDx. There they’ve enabled thousands of conferences around the world to use the TED name, but in a way that doesn’t require a lot of approvals from the mother ship. That keeps them innovative, even if they stop innovating at their core (everyone outside continues the innovation).

Sachin Agarwal, one of the founders of Posterous, echoes these comments in a post about what he learned working at Apple (Small teams rule).

Some of these lessons sure seem counter intuitive. Remove people from a team if you want to make it more productive? But I have heard this over and over again in my journey through the world’s best tech companies.

So, how about you? Are you seeing the same problems at your work? When I do I point them out and we try to fix them.

By the way, you can see my Instagram photos done with my iPhone on Tumblr and I’m “Scobleizer” on that service, if you want to follow me.