What is the optimal IT spending strategy for a fast-growing startup? If you're Brian Morgan, CTO at Catalant, a Boston-based platform startup that delivers business expertise on demand, IT investments are dictated by what the company is trying to achieve, period.
"We have had to adapt to the stage the company is in," said Morgan, who shared his budget strategy for building an agile technology-fueled company.
Catalant, with a global network of more than 27,000 boutique firms, custom teams and independent experts, provides a platform that allows enterprises to access business expertise on demand.
The Boston-based platform startup, founded in 2013, now serves thousands of clients. Catalant caters to large enterprise organizations, while its HourlyNerd site targets small to midsize businesses.
From the start, Catalant sought to be an efficient way to connect businesses with expertise through a virtual storefront, Morgan said. But its goal was to be more than a job listing; Catalant uses algorithms to match the best consultants with companies based on project descriptions and requirements.
"As we go into the Fortune 100 companies, we want to solve their knowledge problems as well as their staffing problems," Morgan said.
But delivering on that business goal was not guaranteed. HourlyNerd's early technology -- a website built by an outside development firm -- was not up to such tasks, according to Morgan. Moreover, it wouldn't allow the fast-growing platform startup to capitalize on growth opportunities.
Shoring up and automating the foundation
Morgan said his first task as CTO was to nearly freeze innovation and rebuild the entire foundation. He wouldn't disclose actual costs, saying only that the company used a good portion of its "tens of thousands" in seed money.
Brian MorganCTO, Catalant
His biggest expenditure then, as it is now, is in engineering headcount. However, Morgan said it wasn't just about adding talent; he had to put in place the processes that would allow the company to respond and grow as quickly as the market did. Early investments thus were around automation and adding redundancy.
"Without those, your day-to-day risk is so high," he said.
The cloud figures heavily in Morgan's IT spending strategy. Early automation investments were made with respect to deploying software and provisioning machines, including servers hosted on AWS and servers running in a virtual machine on developers' computers. This allowed each new employee to hit the ground running and assured that all machines were built and deployed correctly, he explained.
Later he said he brought in tooling like Jenkins, an open source continuous integration software tool which allowed his team to automate more tasks and remove the human involvement almost completely. As for redundancy, he said his team leveraged Amazon Web Services heavily to give them machines in multiple isolated environments so that they could run multiple instances of each type of machine needed.
"Once we got what was there stable, we, to some extent, froze 90% of future investments to rewrite original platforms. It was on Ruby on Rails, which was fine, but it was hard to maintain," he explained, adding that he built a new stronger foundation on a Python stack.
Imperfection by design
The new system didn't always run perfectly -- by design, Morgan acknowledged.
"We were trying to be as agile as possible. Why should we spend six weeks polishing this product? The lifetime of the feature might only be three months. We just wanted the functionality to be out there," he said.
It's part of the trade-off when growing fast. But he said startups have an advantage when glitches happen, because early adopters are forgiving as long as the startup provides fast fixes.
"You can take these perceived bad moments and turn them into moments of delight because the early adopters see it as amazing customer service," said Morgan, who had worked at four startups prior to joining Catalant.
"This was a strategic decision. It was something I stumbled on about two companies prior. And there's no other way [to do it] in my mind," he said.
'Ambidextrous' platform startup
That's a strategy all businesses can learn from, said Jennifer Bailey, an assistant professor in the technology and operations management department at Babson College.
"Every business has to maintain that delicate balance between reinforcing the old and creating the new," she said. "They need to be ambidextrous: exploiting their existing infrastructure and the capabilities which they have already built while simultaneously exploring new opportunities and innovating for the future."
Morgan said he didn't lose sight of how the technology needed to support business objectives and strategic goals. His early re-engineering work, in addition to correcting problems, allowed for the addition of new functions that could drive company growth.
As part of his strategy for the growth, Morgan said he moved his team away from a monolithic design toward a client front-end model, exposing APIs which led the team into building out its enterprise portal.
Meanwhile, true to his approach of aligning IT investments to the business goals of the platform startup, Morgan adjusted IT processes as the company started to add more large enterprise clients. Catalant's larger enterprise clients had less tolerance for technical glitches even if he could deliver fast fixes. As a result, he said he invested more in testing and measuring quality of releases before they went live.
Cloud a critical resource for fast-growing startups
Babson College's Jennifer Bailey said the need to balance speed, resources and growth are typical for startups and platform startups, in particular.
"A platform-based startup faces two critical uses: limited resources and the need to scale quickly," she said, noting that "a cloud-based strategy allows startups to scale quickly and cost effectively because it limits capital expenditures, shortens implementation timelines and keeps operating costs low."
She added: "For both startups and established companies, a cloud-based strategy allows the firm to be more agile. Because the implementation sunk costs are lower, in terms of equipment, personnel, expertise, et cetera, the firm is less constrained and is more able to adapt to change and to take advantage of new opportunities."
Those companies that get the balance right are the ones who succeed.
Standards and keeping the startup vibe alive
Still, Morgan is careful not to move away from the startup mentality that allows agility. To that end, he restructured his technology teams earlier this year, creating more autonomous teams who own different parts. For example, the platform team can bring in new tooling when new feature requests come in and the product team focuses on delivering business value.
"Each team can work like a startup, so they don't have to worry about what else is going on in other areas of code," he explained.
Morgan cautioned, though: "This approach does not come with all good, but the benefits outweigh the disadvantages."
He said one of the biggest concerns is that the architecture and coding style differ so much from group to group that it could be hard for an individual to change teams. To counter that concern, two self-organized groups -- one that's focused on the front end and the other on the back end -- meet weekly to discuss best practices and patterns they would like to adopt or shed.
"These self-forming teams have the power to make decisions that are best for the way they want to work," he said.
There's also the challenge of ensuring that the teams work toward common goals, he said. To counter that, he holds regular meetings to discuss company priorities, short-term objectives and long-term visions.
Investing in people
Morgan said his IT spending strategy means that nearly all his budget (about 95%) goes to headcount, with infrastructure costs taking up the remainder.
He credits AWS for "driving our costs to run all our environments to almost nothing," and open source for keeping software costs down. Morgan's main costs, he said, are on "process and tooling," including code repositories and monitoring systems.
His IT spending strategy for analytics and tracking has been tweaked to accommodate the business's growth. "Early on we exclusively used free Google Analytics, but as our product and organization grew and we changed the type of metrics and insights we needed, we moved to a paid Mixpanel service. On the database side, we use MySQL, which is free. We only pay for hosting it."
Moreover, he said his tech philosophy is also reflected in how his team spends its time, with about 80% of his staff time building new product features with another 10% focused on skunkworks-type innovations.
"I do not have an FTE focused 100% of their time on infrastructure," he added. "The remaining 10% of the staff is focused on IT services and platform level tooling. This team is spending more of its time writing tools and libraries to allow the product teams to work more effectively."
Recent stories by Pratt:
Scalability of cloud boosts Ohio housing authority
CIO presence on boards growing, but still small