Starting is Hard To Do

Mandrill (our new email as a service product, for those that don't know) has had an interesting history. It started as an idea about two years ago, and its path of becoming a product is an example of a growing company trying to figure out how to start something new.

tl;dr As a company grows around a product, inertia from past success can kill short-run productivity on something new. A way to get around this is to create a whole new team isolated from the core to break the inertia and force them to bet big on the new risky product.

Mandrill Skunkworks

Why Do Startups Win?

To a naive observer, it seems insane that tiny companies bootstrapped with no money and made up of a bunch of friends from college could ever build products that rival the output of huge companies. The huge companies have easy access to large amounts of money. They have built-in pools of trained experts that they've hired over the years and can afford to pay well. They have a brand that people recognize. They have man-millenia of experience building products and systems with a proven track record of success. (they're big for a reason, right?) Tiny companies have none of these things.

So why aren't big companies exploding with innovative new products? Why do they pay huge premiums to acquire tiny startups? It can't be for the idea - everybody has ideas. It would have to be for the team or the execution of the idea. With all of these advantages, why does the big company need to pay a premium for that? Why doesn't the big company just create new products from within?

Now, MailChimp probably isn't large enough to qualify for everything I mention above, but I think that our experience trying to start and launch Mandrill may help to illuminate some of the reasons why it's so hard to get a big company to create something new.

Big Companies Aren't Generic

One huge problem with the advantages listed above is that big companies are not generally built around generic ideas. MailChimp is not a software company. Yes, we do make software and we care a lot about being good at it, but it's much more specific than that. MailChimp is a web-based large-scale UI-driven email software for businesses company, and the way the organization is put together reflects that. MailChimp is our core product, and all the people and departments we've created are built around sustaining and expanding on MailChimp's success.

On top of that, we have all these departments and processes that make sense when they're applied to MailChimp, but could be totally wrong when talking about Mandrill. Should the MailChimp support team handle customer support for Mandrill? Should we be using the same sort of design and UX process for Mandrill? Should the anti-spam systems be the same in MailChimp and Mandrill? The answer to all of these questions is "We don't know - it doesn't exist yet." Sure, we have a decade of experience making a successful MailChimp, but the farther the product gets away from MailChimp, the less that experience translates.

Success Builds Inertia

This uncertainty and risk is natural to the process of starting something brand new. If you're in a startup that's the fun part, and the risk of failure and bankruptcy provides plenty of incentive to make bold choices. The problem is that larger companies are generally built to avoid those kinds of risks. When you already have a successful product, most radical changes are probably worse than what you have - the landscape of failure is much bigger than the landscape of success. You start worrying more about protecting the core from risky disasters than you worry about missing risky opportunities.

In day to day operations, this is easy to see. When working on Mandrill while still a MailChimp developer, I frequently had a choice to spend the day building out features that might be useful to some theoretical customer for Mandrill if it ever launches and has customers, or I could spend the day writing a feature for MailChimp that I know would help customers that I know exist. It's not surprising that when given that choice I usually went with the MailChimp feature.

On top of the internal issues with risk and uncertainty, there's the problem of the new product maybe reflecting poorly on the existing successful product. If a startup makes mistakes or misses the market early on in product development, it doesn't really matter. Sure, the few customers you have might think poorly of you, but almost nobody's heard of you yet - you don't really have a brand to destroy. This obscurity gives you time to make wild and drastic changes to the product without the risk of getting a bad reputation in the market.

Big companies can't do this - they already have a recognized brand. The MailChimp brand means that Mandrill had more customers on day one than it ever would have without the MailChimp brand. That's a good thing - customers are great! But it also means we have less flexibility. We can't change Mandrill around too much without hurting all the new customers and maybe even hurting the MailChimp brand by association.

Small Company People

Another thing that changes for bigger companies is the general makeup of their employees. It's pretty obvious that as your company grows, most of the folks that you hire are going to be people that prefer to work in larger companies. The people that prefer working in small companies generally leave or burn out. Just think of how long the original teams at startups stick around once they're acquired by a huge enterprise - generally not very long.

There are a lot of reasons why someone might prefer smaller or larger organizations, which I'm not going to get into here. The important thing is that the appetite for risk is much more prevalent in small company people. Since big companies have fewer of these people, it's harder for everybody to get behind taking a big risk. After all, if they wanted big new things, they would have joined a startup. This one probably matters less at MailChimp since we've only gotten big very recently. It's more noticeable in companies that have been big long enough for big company to really soak into the DNA.

Research and Labs Products Don't Count

There is one type of new product that big companies can generate in droves without too much effort, which is the experimental "labs" or "research" product. If you were to ask a big company why they aren't building big new things, this is probably what they'd point to. MailChimp is the same way - we have tons of labs projects and little products. But, Mandrill (or any other big, risky bet) is different than these things. The difference is that research products are generally free, not promoted, not aggressively developed, and easy to kill. The typical labs project will never be a risk to the company. If it doesn't work out, just kill it or stop working on it; no harm, no foul.

The success rate of these types of non-core products is laughably low. Almost no product actually succeeds in the real world without focused development, pivoting, or at least active promotion. By not aggressively putting the brand or huge resources behind a new product, the company can remove all the risk of starting something new. It just so happens that putting resources and brand power behind a new product is also necessary for it to succeed. You can't have most of the organization ignore a non-core product and expect it to succeed. You need to bring it into the core, and that involves risk.

A Hopeless Task

So while big companies have access to money, talent, a good brand, and experience with scale; all of these things also make the company both tightly coupled to one particular core product or goal, and more reluctant to try new, risky products. Maybe this explains why big companies might prefer to let smaller guys build the new ideas.

That wasn't good enough for Mandrill. We didn't want to acquire a new product just to get around organizational challenges. We wanted to build something new that would succeed, and we wanted to build it ourselves. So how could we possibly fix these issues?

Visionary Dictatorship

One possible solution to this is to have a visionary executive at the top forcing the organization to take a risk - call it the Steve Jobs model. I can't really speak to this too much beyond mentioning that it might work, but it requires that you have a trusted visionary dictator at the top of the company. Most companies I'm aware of don't work like that, and MailChimp definitely doesn't work like that. We're not a top-down authority-driven organization. We're a bottom-up creatively-driven organization. If we were going to be able to start something new, it was going to have to be based on people getting themselves excited about the idea - not because the CEO sets a direction and everyone falls in line.

Create a New Core

Thinking back on it, the path we ended up taking seems like an obvious solution. If your organization is built around a single product but you want to build something new, then just build a new core. Turns out that's really easy to say but really hard to do. First, you need to get everyone to agree that you need a new core in the first place. This isn't easy, since people will be afraid that the new team might screw up and reflect poorly on the main brand. You need the complete trust of the entire leadership to make it happen. Fortunately, we didn't really have a problem with this at MailChimp. Once the idea was floated that we needed a totally separate core team for Mandrill, it was quickly understood and accepted. It other, more dysfunctional companies, I imagine this would be a much bigger challenge.

Next, you need to build the team. Naturally, that means taking people from the core product and moving them to work on the new product. It's not just any people either. If the new product is going to succeed, they need to be your best people. Your best people, by virtue of being your best, are probably relied on heavily for building and improving the core product. They may be integral to many essential processes, so just reassigning them will be a big challenge, but it doesn't stop there. The inertia of success isn't just organizational, it's personal. You need to isolate the team totally from the core product. They need to have no option but to build the new product, even if they feel like the existing core product needs them (and it will). If you don't, they'll take care of their existing baby before creating the new baby, and the new baby will never get made.

And that's about where we are today. It's still early in this experiment, but the Mandrill team is happy and productive (so far). We've really only just begun the process of seeing if we can turn MailChimp into a company revolving around one core product to being a company revolving around multiple cores that may end up looking very different from each other. The future is uncertain, but we're excited to see what might happen and we hope that you'll stick around for the ride. Even if we fail completely, it'll probably be fun.