Ideas are important, but execution is everything. Here’s one example of just how true that is.
You have an idea. It’s a good idea. In fact, you think it’s a great idea.
But is it a business?
That’s the question every would-be startup founder asks–and needs to answer.
Here’s another in my series in which I pick a topic and connect with someone a lot smarter than me.
This time, I discussed how to know if a business idea could be a viable company with Stuart Frankel, the co-founder and CEO of Narrative Science, the makers of Quill, an advanced natural language generation platform. (Basically, Quill transforms massive amounts of data into a story that sounds like someone wrote it, highlighting the most important insights.)
Prior to launching Narrative Science, Stuart was the president of the Performics division of DoubleClick; before that he was a practicing attorney and CPA. (I told you he’s a lot smarter than me.)
Unlike many tech entrepreneurs, you didn’t start out as a coder who developed software and tried to turn it into a product.
I went to DoubleClick in 2004 through the acquisition of Performics and worked on the senior team. When that business was sold to Google, I stayed there for a time and then looked for my next opportunity.
I had told my wife I would never do another startup. I felt like I had checked that box in my life. But in 2008, the financial crisis meant there wasn’t a lot of activity, so I connected with two computer science professors at Northwestern University who ran an innovation lab and were working on some interesting artificial intelligence concepts and technologies.
Most of the projects that were running in the innovation lab didn’t have much commercial opportunity, but I thought one did: a piece of software that automatically generated natural language baseball game stories from data, which ended up becoming the precursor to Quill. I thought it was game changing in terms of the impact it could have on a wide variety of industries. With all the data being collected, we knew the current analytics and visualization applications were insufficient simply because of the sheer volume of data. We felt there was a macro-trend taking place that would force new technologies in the area of big data.
I couldn’t articulate it as well then, but it was clear something was changing.
The idea almost seems too simple: turning data into readable, engaging text.
Many companies monitor and measure and track every activity undertaken by their employees and customers. That means they’re sitting on a massive amount of data. But much of it is unintelligible or simply goes unused.
Most knowledge workers–and most consumers–are forced to stare at walls of numbers and make decisions on the basis of what they see. Reading is a much better way to digest and absorb information. Machines are better with numbers; people are better at reading.
Data is useless unless people can understand it, digest it, and use it to make the right decisions. Ultimately, that’s what our business provides.
Still, cool technology alone does not make a successful company.
Absolutely not. So I did some research and looked in particular at the media space–newspapers, magazines, online publications–and we decided technology that could take data and turn it into natural language was something they would be interested in. That was my first indication there could be a market; I felt there would be an insatiable demand for content if we could do it at scale.
Technology is great, but there has to be a market: either one that is well defined or one you can clearly see emerging over a reasonable period of time. We went after media because it was an existing market. We really started the business in that market, but in the back of our minds we knew there could be even bigger opportunities in other industries.
That makes sense, but that’s also a tough decision to make–do you try to service a silo or think more broadly?–because that decision drives everything you do in terms of development.
Over the first couple of years, we received interest from financial services organizations, health care, and government agencies, and that was a pretty big proof point. And it helped us decide whether we should be a long-tail media company selling services to publishers, or a software company that builds and licenses software to enterprises.
We decided on the latter, especially because of all the investments companies were making in big data. We knew there would be a market for what we do.
Ultimately, that’s what gave us the final boost of confidence that we could build a thriving business.
I keep bringing up challenges, so here’s another one: Once you think you have a good idea, you still have to implement it.
Implementation is the hardest stage. Execution is really, really difficult. One lesson I learned early is there are no easy days.
In our case, we had to figure out who our customer base would be, but on the other hand, our salespeople needed something to sell. We had to figure out who and where to sell because we couldn’t sell to everyone; we couldn’t be successful in all industries simultaneously.
For a while, we engaged a number of organizations in conversations and did business with some as we worked to determine what places within the broader market would lend themselves well to using our technology.
So even though you chose to be an enterprise solutions company, how did you decide where to focus?
In time, we began to focus on financial services. Quill relies on lots of high-quality data, and that’s a given with financial services companies, which absolutely must have their data in order.
Plus, they had the resources to invest in technology and a real desire to improve their communications with customers.
If you think about it, the information you get as an investor can be a little obtuse. You get a table of your holdings, a pie chart that shows your allocation of funds, and then you get the performance of your portfolio compared with some selected indices, and it’s really difficult to tell how you’re doing. You get pretty pictures, but you don’t get much in the way of useful information.
We knew wealth management firms were fighting for customers and needed to improve the customer experience. Quill can write an individual portfolio review targeted specifically to you–how you’re doing against your goals, against prior years, in comparison with other choices you could make–and it’s easy to read.
We have another product, Quill Portfolio Commentary, that helps mutual funds generate their quarterly performance summaries. For many funds, that process is incredibly cumbersome, surprisingly manual, and as such prone to errors, and we configured Quill to do that automatically: It’s more accurate, more efficient, and essentially error-free.
We can do that because we made the decision to be a company that creates and licenses software to enterprises. That allows us to develop specific solutions that address specific problems. In time, that’s given our salespeople a portfolio of products to sell–and made it possible for us to seize new opportunities, too.
So while financial services has been good to us, it’s not the only industry we’re working in or focusing on for the future.
I get the automation and accuracy side of it, but I assume the customer relationship benefits are nearly as important.
While it sounds obvious, smart organizations know there’s a real benefit to building customer loyalty. Even if sometimes you’re delivering bad news, when you build trust, that goes a long way toward solidifying relationships.
Take a company like USAA, one of our customers. It’s one of the largest financial services firms in the world with 10 million members, and it consistently shows up on lists of most admired companies in America. It does a phenomenal job of building customer loyalty, and a lot of that has to do with providing its members with the right information so they can make the best decisions about their financial and insurance needs.
So we do provide a data and information generation service, but what we do for USAA is also consistent with its mission of developing customer loyalty.
Once you gained a foothold, how did you weather the tough transition from startup to growth?
The pressure only intensifies when you move from the startup to the growth stage.
In the early days, you have high expectations of yourself, but there are very few people that have expectations for you; relatively few people are looking at you or depending on you. Plus, the statistics say you’ll likely fail. So there’s a weird lack of external pressure.
As you grow and take on investors, get a board, start to have functional pressure that didn’t exist, get customers you need to service and keep happy, etc., there’s pressure–good pressure, but pressure nonetheless–associated with those things.
At least for me personally, the more successful we are the more pressure there is to keep succeeding. You do breathe easier that you won’t go out of business or miss payroll, but you also have a new set of challenges that emerge.
With that said, though, I’m wildly optimistic about our company. We have a much broader set of choices in front of us. We’ve shown success in financial services; we have customers in other industries. Now we have the luxury of deciding whether to move into other verticals or dive deeper into the areas we’re already in.
And the best part is that what we believed early on is coming true. The ability to generate narratives from data will someday be built into every dashboard. People now believe that technology like ours can improve operations, help customers, reduce risk. There’s no reason why technology like ours won’t be widely available if not ubiquitous.
So, over time, we’ve gone from deciding whether our idea could be a business to deciding how to take advantage of the tremendous opportunity we have.