May 16, 2006

How vertical-market software firm Duck Creek grew 1,131 percent in three years

by Adam Stone, Senior Writer, SoftwareCEO

What is it with software CEOs? Just when you think they're going to retire, they keep coming back for more.

And sometimes they do it even better the second time around.

Take Doug Roller. Classic case. He built a business doing software for the insurance industry, sold it, then stayed on to help run the place for another decade, before finally calling it quits in 1996.

"My oldest daughter was in middle school, my son was getting ready to be there, and I felt I was missing out on their lives," he says.

But by 2000, he was back in the game again.

This time, Roller started a software firm with a slightly different take on the same industry. Based in Bolivar, Missouri, Duck Creek Technologies ranked #63 on last year's Inc. 500 list — with astounding growth of 1,131 percent over three years.

In 2005, the company had 150 people and $11 million in revenues. This year, Roller says, Duck Creek is on target for 200 employees and $25 million.

That's more than doubling sales in one year, again.

How do they do it?

Well, it helps that Duck Creek's vertical is ripe for the picking. The insurance industry, especially the intersection between agents and carriers, cries out for automation.

Need to manage endless tons of information? Need to access that data quickly? Need to customize how it's presented? Software can do all that.

But it takes business savvy to make good on the most promising opportunity. And Roller seems to have savvy to spare.

We asked him to give us an inside look at what it takes to push a company this far, this fast. And here are his 12 proven strategies for growing a successful vertical-market software firm.

Vertical-market strategy #1: Deliver both a platform and the software to use it.
To understand the Duck Creek business model, you need to understand something about insurance. Don't worry, this won't take long.

Agents sell insurance. To do this, they need information from insurance carriers: What coverage is available? What does it cost? What rules govern it?

Enter Duck Creek with an industry-specific platform, called EXAMPLE. This platform enables carriers to quickly transform products into web-enabled insurance services such as rating, underwriting, and policy processing.

In effect, it's a delivery mechanism for getting information into agents' hands quickly and easily.

A range of software tools then come into play for this platform. Duck Creek's EXAMPLE Author helps a carrier define rates, rules, algorithms, documentation, screen flow, and policy transactions. EXAMPLE Server transforms data into web-enabled services. Other products customize industry data and support policy administration.

The common themes in all this? Speed and simplicity.

"Our vision is to continually reduce the time it takes to imagine, implement, and maintain insurance products, so that the carrier is able to get to market faster and less expensively with higher quality," Roller says.

Vertical-market strategy #2: Rise like the phoenix.
Okay, it's not exactly like a phoenix, since Roller's previous venture never went up in flames. Quite the opposite, in fact.

Founded in 1981, Micro-Magic Systems grew steadily until he sold it in 1986 to AMS Rating Services, the largest insurance rating firm in the United States. Roller stayed on to help run it for another 10 years.

Still, Duck Creek does represent a rebirth of sorts. The company drew together a number of former colleagues from the old venture looking to try something new.

"It was kind of like a western, where all the old gang comes back together," Roller chuckles.

Nor does the new business merely replicate the old. Instead, Duck Creek builds on the experience gained from the first time around, while taking that expertise in an entirely new direction.

"In the old business, we were providing rate updates to agents, but we had a real hard time getting the information from the carriers in a form that made it easy for agents to use," Roller says.

That got them thinking: If they were having trouble in that area, that meant there was a problem in need of solving.

"We asked ourselves: If we had the opportunity, how would we change this process?"

The answer was to create software that backed up the process by one step. They would start not with the agents, but with the carriers. They'd build software that could help carriers easily generate all sorts of rating and underwriting information, then give them a platform on which to deploy it.

Vertical-market strategy #3: Don't underestimate what it takes to create vertical-market software.
Take the pieces from half a dozen jigsaw puzzles, and mix them all together on a table.

That's the insurance business. Roller calls it, "one of the messiest businesses in the financial-services world."

"If you're selling commercial insurance, there are literally thousands of coverages, from business interruption to flood and fire to theft to valuable papers to employee dishonesty," he says.

And that's just one corner of the vast puzzle.

"By the time you tally them up, insurance companies become extremely complex, with hundreds and even thousands of possible coverages," Roller says.

And these products vary from state to state, depending on numerous regulations.

You have to work pretty hard to make anything out of this puzzle. Developing software for the insurance industry is a major undertaking. It's got to be led by developers with a stomach for extreme complexity. Nothing less will do.

That means each new Duck Creek product represents "an enormous challenge. It is a massive development project where you have teams of people building and testing and validating," Roller says.

For instance, take the company's commercial coverage product, which is delivered as web content.

To build it, Duck Creek used its in-house toolset on data from national insurance bureaus, including the National Council on Compensation Insurance (NCCI) and Insurance Service Organization (ISO).

Wait, stop here. Spend $3.5 million over two years to get that content into usable shape.

Okay, now you can start developing again. Make sure you have a team of 40 programmers to move that content to your system.

Now do the same for every new product. That's what it's like doing software for the insurance industry.

Every vertical market presents its own puzzles. And no ISV can afford to underestimate how tough it will be to solve them.

Vertical-market strategy #4: Tailor your pitch to the size of your prospect.
Even with such huge challenges, growth was in the air by 2004. And Roller wanted to be ready for more.

"It's necessary to build infrastructure ahead of growth. In our case, training, service capacity, and the management team were all important for us to reach the next level," he said.

Help came in a round of venture funding (value undisclosed) led by Pequot Ventures based in Westport, Conn.

That money went right to work.

Duck Creek's sales force has grown from three to five in roughly the past year. The sales cycle can range from two to 18 months, with four to six months as an average.

In ramping up sales, Roller has learned one basic lesson. Selling software is not a one-size-fits-all affair. Customers come in a range of classes, and a successful sales strategy must vary to match the different needs of different-sized customers.

Duck Creek's larger clients with mature IT departments typically buy the company's core platform, and then invest in Duck Creek's toolset to speed up their application development. With these clients, sales staff go in talking about the technical merits of the product, development issues, and the virtues of the company's diverse software.

In these cases, the challenge to closing the sale comes from the in-house IT staff, who may see Duck Creek as competition against their in-house applications.

So the company will let its own app take a back seat, "and instead sell ourselves as an enabler for the customer's development, not as competition against their application," Roller says.

Then there are the mid-sized carriers, with less IT support, who may be looking for out-of-the-box solutions. In these cases, Roller will sell them a turnkey solution, and capitalize on the skills gap by expanding the sale to encompass a professional-services team who can come in to customize the software.

Often that service generates more revenue than the software itself.

But sometimes even this won't be enough to close the deal with a smaller prospect.

"There are times when we have to walk away from the sale," says Roller.

"That typically happens when we look at the maturity of the organization, and we come to the conclusion that they're really not large enough to support a service implementation from us."

Vertical-market strategy #5: Make it clear what you mean by "quality."
If we tell you Duck Creek has a "quality team" you'll probably roll your eyes — with good reason.

The concept of "quality" has been abused to the point where it's almost meaningless. The obvious definitions have been reduced to platitudes: Be the best you can be. Do the right thing. Yada yada yada.

Yet Roller has some interesting thoughts on what quality means to his organization, and why it takes a special team to make it happen. It has to do, first, with the changing nature of his company.

"When we started, we were seven or eight, and we were all entrepreneurial. We were working elbow-to-elbow, and communication was no issue, because everybody knew everything. We got an incredible amount done during that first year," he says.

But things change over time.

"By the time you're at 30 or 40 people, you can't get everybody into a room as easily. You can't sit down for a Monday morning meeting with the whole team."

So how do you do it? How do you keep that sense of cohesion, that immediacy even as the company grows?

Enter the "quality team."

What is it? Roller defines the team in part by what it isn't, by what it doesn't do.

"Their goal has nothing to do with dollars," he says. "We ask them to look at what they do, almost independent of the monetary factor. In that way, the quality team serves as our conscience."

Ken Toler, a company founder, is technical lead to the team, which includes members of customer service, HR, recruiting, and several other disciplines from the company.

The team directs the Duck Creek University training program, defines interdepartmental processes, and stays on top of customer service functions.

Right now, they're working to define the handoffs between the account management team that does customer service and sales support, and the product support team that does technical support. A process analyst on this team is driving these definitions, and in the end the documented processes will be published on the company's intranet.

Roller says the team arose out of necessity, as it became clear that some of the most crucial functions in the organization did not fit neatly under any existing heading.

"We needed a place for some of these things to land, where these things would not be an afterthought. Things like training and customer service, for instance, which often go together."

The quality team at Duck Creek, in effect, gives a home and a structure to any processes that cross the lines between departments.

Vertical-market strategy #6: Learn to balance sales and service.
Every juggler has a breaking point. Put one too many balls in the air, and the whole thing comes down on your head.

Software companies face a similar challenge, especially those that deliver a mix of software and services. Sell too much software, and you may not be able to service all your customers.

Roller maintains equilibrium in a number of ways.

"We try to balance it by being careful who we sell to, making sure we advertise and sell to carriers within our specific price point and our ability to service," he says.

Sell to a customer too big, and you'll be drained by the sheer volume of support needed. Take on the very smallest businesses as clients, and you'll get stretched too thin by all the hand-holding you'll need to do.

"We do a good job of filtering out those who are too small or too big, or who need the wrong kind of solution," Roller says.

Vertical-market strategy #7: Look for partners to help build more capacity.
Next, he put together a major-league support organization, with a full-time VP of service driving a 55-person professional-services team.

Finally, he's ready and willing to employ outside firms to get the job done.

"We are aggressively using partnerships to provide additional service capacity," he says.

That could mean using partners to integrate the Duck Creek platform into a carrier's legacy environment. Other partners may come in to help his staff build and customize software for a customer. In all, about 20 percent of the service capacity is being outsourced.

This strikes Roller as an ideal way to maintain the balance between how much he sells and how much he can service.

"It gives us the ability to meet that unanticipated account that comes along, to be able to contract without it affecting our internal organization as much," he says.

The caveat? When you use outside help, make sure they're trained properly.

Just like all new employees at Duck Creek, outside contractors must undergo a three-week boot camp in which they do a real-life build-out of a product.

A part of Duck Creek University, the boot camp is relatively new, so the rules are not yet set in stone. Roller said he expects to charge partners "a reasonable fee to cover the cost of our efforts" and that most will be happy to pay, "since they know they have the potential to gain significant service revenue as a Duck Creek partner."

Vertical-market strategy #8: Define every role clearly.
Of course every person has a job. But does every job have a person?

Roller learned early on how important it is to define every function in an organization, and to attach that function to a specific department or person.

He recalls, for example, the time a customer called to report a feature they needed to see in the software. The project manager took the call, and confidently assured the customer that the feature would appear in the next release.

"The release occurred — and guess what? It wasn't there," Roller says. "So the customer who'd been counting on this feature to complete their project suddenly finds it isn't there."

What went wrong?

As Roller describes it, no one was really in charge of the process, in this case, the process of recording, reporting, and addressing customer issues. The buck stopped nowhere.

"At the time this discussion occurred, there really was no owner," he says.

Since then Roller has tightened things up. Now he has a product manager responsible for defining all the roles and functions for a given project.

In addition, implementation team members no longer wear two hats as customer service reps. That job now belongs to a full-fledged customer service department.

"Now we have a customer service rep who lives with that customer for the lifetime of that customer," Roller says.

In other words, a single individual is now responsible for keeping each client happy on an ongoing basis. They know it, the customer knows it, and the rest of Duck Creek knows it.

Vertical-market strategy #9: Help newbies grow, without squelching them.
What's Roller's biggest challenge? Finding people. But not the same way you might think.

Most software CEOs say they struggle to find and keep talented personnel. Roller's problem is more specific: He needs to find talented people who understand the arcane nature of insurance-industry software.

Not too many people walk in the door with that knowledge.

"In many instances, you have very bright people doing something for the first time," he says. "It's like when you teach a child to ride a bike. He may be perfectly capable of doing it, but someone has to be there to walk him through the steps."

At Duck Creek, the ability to function comfortably with the software is important, but even that isn't enough. Newcomers need to get up to speed on the bigger picture, too.

But how do you make that happen?

Roller uses another special task force. In this case, it's the Solution Assurance Team, whose job it is to support decision-making by coming in at critical points and working closely on the project to ensure the right decisions get made.

This team includes three of the founders, along with three veteran service people. But let's be clear: They don't actually make any decisions.

"They do not take responsibility for the delivery, but they are there to teach," Roller says. "This team never directly interfaces with the customer. They don't get directly involved in product support. They are always involved behind the scenes."

Beyond overseeing the process, they are there to teach people how to make the right call, the steps for making good decisions.

"There is always someone there in a learning role who's going to learn from their presence," Roller says. "The goal is ultimately that the team won't be needed anymore."

Once its wisdom has been shared throughout the organization, he foresees a day when the Solution Assurance Team can be disbanded. But for now, it's serving a valuable role in knowledge transfer and training.

Vertical-market strategy #10: Pay your dues in your vertical.
You can't get much more vertical than Duck Creek.

Insurance is a beast unto itself, which is a good thing — if you know what you are doing.

The trouble is, most people don't know what they're doing. Roller says a common flaw among software companies is trying a vertical play without having enough intimacy with the industry.

In insurance, this usually means trying to get in the game by grabbing onto a single piece in that big crazy puzzle.

"I'll get a call from somebody who says, 'We are looking to get into the policy-administration business, and looking for a good partner.' When they come with that approach you almost know they are going to fail, because they don't have depth," he says.

"The way to be successful in a vertical is to solve the hard problems in that vertical. You can't come in with a horizontal solution and say, 'Here, now make this fit.'

"To be successful in a vertical requires two things. It requires a great knowledge of the vertical that you're working in, and it requires a great vision for a solution."

So how does any software company looking to grow pick the right vertical? That's easy, says Roller.

"Find a messy one. Find one where the problems are really hard." The greatest challenge brings with it the greatest reward.

Vertical-market strategy #11: Dig deep for customer feedback.
Duck Creek's software never sits still for long.

The complexities of the industry call for constant mutation in response to regulatory changes, demand for new insurance products, or the need to provide agents with better, faster information.

Duck Creek responds with two major releases a year, one in the spring and one in the fall. In addition, the development teams put out maintenance releases about every six weeks.

All these changes are customer-driven, yet the input from clients is strictly regimented in a way that's somewhat unusual in the software industry. While other firms rely on user feedback and casual surveys, Duck Creek runs a highly formalized system for seeking customer guidance.

A customer advisory board convenes four times a year, twice in person, and twice by conference call.

In addition, Duck Creek executives put together customer focus groups in advance of each new release.

It sounds like overkill: focus groups on top of a formal advisory board. But there is a big difference. While the advisory board is mostly senior-level people, the focus groups are typically end users.

Roller says he needs to hear both sets of voices.

"We want to watch the people who use the software, and have them tell us their problems," he says. Then they can roll both corporate-level and end-user comments together to create a set of requirements for the next upgrade or new product.

Vertical-market strategy #12: E pluribus unum. Out of many, one.
That's how development works at Duck Creek, and it's a sound model for the software world.

With multiple product lines in the mix, the company has several development efforts underway at any point. All the wheels have to keep turning.

But how can a savvy software executive ensure that these efforts don't all go spinning off in their own directions? How can you make sure that the wheels all stay on the same axle?

At Duck Creek, the engineering team is subdivided into eight feature teams: one for the Author product, one for Server, and another for Express. Another team handles documentation. And four more teams build content for the different lines of business.

The effort to keep all this running smoothly begins with reporting. Each team must present a periodic plan detailing how it plans to meet its development goals.

But are they on the same track?

That's the job of an internal board made up of executives who get together to set big-picture goals, guided by the marketplace and the needs for internal support.

This group lays out the vision from the top. Those big-picture goals get wide play within the company, which helps to make sure everyone is moving in the same direction.

And every six months, before beginning work on the next release, leaders from the development side of the company set aside a week to review the road map, evaluating the company's long-term goals along with its short-term market needs.

They meet in a rented conference room half a mile down the road. While not exactly glamorous, it is this "product planning retreat" that ultimately brings it all together.

In the first year, Roller says, "We had various people in attendance for each product line, with a core group of about 10 people for the whole meeting. We had about 25 people who participated in some role."

By the time it's over, "all these end results are unified, and communicated to everyone," he says. "Everyone walks out of there knowing what their part is."

And when everyone knows their part, it gets a lot easier to face that great hulking vertical called insurance. In a thorough, methodical way, this is just what Duck Creek has managed to do.