Editor’s note: Steve Sokol is the product lead for Respoke, a WebRTC platform as a service that allows web and mobile developers to add live voice, video, text and data features to websites or apps.
I have been part of several of startups, and I love the energy, excitement and sense of risk that comes with launching into uncharted waters. My last startup was acquired in 2007, and I’ve been working at the acquirer ever since. A bit over two years ago, I got the itch to do something entrepreneurial again.
When I talked with my boss about quitting to build a new business in the nascent WebRTC/web communications space, he made me an offer I couldn’t refuse. And that’s how I landed in my current position: For two years, I’ve had the unique opportunity to work as a co-founder at a virtual startup, housed under the corporate umbrella. The ride is not quite the same as your typical startup, but it’s a long way from business as usual.
If you’ve read The Lean Startup, you know a startup is intrinsically different from an established company. The mission of a startup is to find a business model, and the mission of a company is to capitalize on that model. If you’ve read The Innovator’s Dilemma, you also know existing companies have to keep innovating – sometimes disrupting their own primary business model – if they want to remain relevant as times and technologies change.
As it turns out, it can be really challenging to innovate inside of an existing, successful company. Everyone is busy doing their part to move the existing products and services forward. Everyone is intellectually and emotionally invested in the current model. The sales organization has established channels. The marketing team has a set of customer personas they understand. The engineering team has a mix of skills tailored for the current products. And nobody has the time to disconnect and try to think about a completely new approach.
Large companies solve this problem by acquiring innovative, small companies. Smaller companies have to take a more creative approach.
I work for Digium, a business telecommunications company that launched as a startup in 1999. Digium led a wave of disruption early in the millennium with Asterisk, an open source software project that replaced over-priced proprietary PBX systems with commodity computers.
Digium started out selling add-ons and support, and evolved to offer a full line of business communications products and services. I came aboard in 2007 when Digium acquired my Asterisk-focused training business. After five years working as the product manager and later marketing manager for Asterisk, I was ready for a change and saw WebRTC as the next big wave in communications.
It turns out that Digium had a similar thought. The leadership team recognized both the opportunity and the threat posed by the new technology and wanted to explore its options. I stayed, and within a few weeks had assembled a small team tasked with researching the technology and looking for the opportunities it presented. We spent several months generating plans and MVPs (minimally viable products), which we pitched to Digium’s board of directors. After a couple of iterations we landed on the concept that became Respoke – our web communications PaaS and API.
The idea that we should operate Respoke as a virtual startup gets credited to our VP of Engineering, David Deaton. Having worked at a combination of startups and established tech companies, he recognized that it is hard to successfully mix the two. He proposed the startup structure as a way of giving the new project the flexibility to experiment, and as a means to avoid the distractions that pervade any established business.
Our group moved into an unused section of the company’s headquarters and started building our product. At the time, none of us knew of any other virtual startups, but that changed when we met HubSpot’s Christopher O’Donnell during our early customer discovery phase.
When HubSpot, the Boston-based inbound marketing startup, decided to enter the competitive sales automation space, it chose to launch what VP and virtual founder Christopher O’Donnell terms a “startup-within-a-startup.” HubSpot chose the startup approach for several reasons. The new business was diving into a different market and was intentionally adopting a different go-to-market strategy. Launching Sidekick, the new product, as a startup allowed O’Donnell and his team to test and pivot.
The HubSpot experiment came with very few strings attached. O’Donnell was given three mandates: it had to be for the sales automation market, it had to involve email, and it had to be self serve and no-touch.
The process started out as lean as you can get: O’Donnell worked by himself for the first several months creating and testing hypotheses. Next, he brought in a few engineers to build an MVP. The result were initially validated by letting HubSpot’s sales team give it a try. When the product caught on organically (i.e. without a “you must use this tool” directive from sale leadership) they knew they were onto something.
The story O’Donnell tells is not particularly different from other startup tales. The initial concept required several pivots to find a solid fit. Growth is just as critical. It calls for the iterative approach that O’Donnell describes as “sketching in pencil, then re-drawing in pen.” As an experiment, he views the Sidekick startup as a complete success. In less than 18 months the team was able to create both a new product and a new channel.
Matthew Hodgson of Matrix tells a similar story. Matrix was launched as a “curated startup” by Amdocs, a thirty-year-old multinational company that specializes in billing and directory platforms for the telecom market.
Recognizing a potential existential threat in the ongoing decentralization (and devaluation) of traditional communications, Amdocs created Matrix as something of a hedge. Hodgeson’s team was charged with building an independent ecosystem around an open source standard for federated communications.
Matrix was not the first curated startup from Amdocs, but it may be the most audacious. Creating an ecosystem around an open source product is no easy task (I know – I spent six years working as product manager and marketing director for Asterisk).
Amdocs views Matrix as a “moonshot” – an opportunity (and a gamble) to own the story of their own disruption. It’s the kind of experiment that you can’t do directly with an established company – especially one with 20,000 employees and offices in nearly 70 countries.
“Moonshots” seem much more at home at Google or Facebook than someplace like Amdocs. Hodgeson admits that getting the necessary support and approval for a startup inside of legacy enterprise business is unusual. Matrix had the good fortune of finding a champion at a very high level in the Amdocs org chart. As he explains it, “Without buy-in from senior leadership – preferably at the C-level- it is very hard to succeed.”
Pros and Cons
In theory, a virtual startup is an autonomous group within an established company that acts as a business model generator / validator. The idea is to create enough separation from the parent company that the startup group can hypothesize, experiment and pivot when necessary.
The virtual startup does most of the same things that a “real” startup would: hiring a team, developing a MVP, testing for product / market fit, etc. Once a viable model has been found and validated, the startup transforms into a business unit of the parent company.
In practice, the lines aren’t quite as clearly drawn. There are some huge advantages and a few disadvantages to being a virtual startup.
In a real startup you spend a lot of time looking for money. That’s less of an issue in a virtual startup. Digium understood and acknowledged the opportunity in front of them, so we were able to skip out on much of the fundraising process.
The Sidekick team at HubSpot had the same advantage, though O’Donnell was expected to show actual revenue much more quickly than a traditional startup. Even with that challenge, most would kill to be able to “one-and-done” the fundraising.
A down side to the virtual startup is one of the strengths for a “real” startup – equity. Established companies with existing value have a stock option program geared towards rewarding a combination of high performance and loyalty. Traditional startups, on the other hand, have a huge pool of (initially) worthless stock. It’s far easier for a traditional startup to offer a mix of salary and equity to attract and retain talent.
Virtual startups (including Respoke) have to live within the option program of their parent company, which means the compensation mix is much more heavily tilted towards salary. This isn’t entirely a negative. A salary-heavy blend is often attractive to experienced developers who tend to be able to work “smarter” rather than pulling an infinite series of all-nighters to crank out the product.
Another difference that’s a bit of a mixed bag is the existing infrastructure the parent company can provide. Not having to set up payroll, insurance and other business basics is a plus. But having to make use of existing systems that were selected or built for a different business model can be a challenge. We had to spend a good deal of time integrating with existing accounting and billing systems.
A traditional startup would have built a solution on Stripe or Chargify in half the time. This isn’t a huge deal, but it proves there are limits to how nimble or lean you can run when you’re attached to a larger organization.
Virtual startups generally don’t have to worry about finding office space. The local members of our mostly distributed team simply moved into a section of the company’s existing headquarters. We knew from the beginning we would be hiring remote employees so we also invested some time to build the kinds of collaboration tools that make remote work painless. I’d call both instant office space and remote workers a win for the virtual startup.
Other aspects of virtual startupdom are pretty much the same as the classic model. In the early days, virtual startups spend a lot of time on HR stuff because assembling the right team is critical. They argue (amiably) over the details of the MVP – how “minimal” can it be? (Hint: a bit less than it would be for a classic startup because you don’t want to risk damaging the parent company’s reputation with a spectacular flop.) How do you price it? When do you hire the next engineer? What’s the burn rate? How can we get useful feedback from early adopters? These types of questions are shared concerns no matter what type of startup you are operating.
If you’re working in an existing, established business but long for the excitement of a startup, consider pitching a virtual startup idea to yourboss. It’s an evolve-or-die world, and finding a way to innovate should be on everyone’s roadmap.