John Roberts Asked for Six Months. Then He Opened the Code.

John Roberts asked his wife for six months. He wanted to see whether business software could be built differently, with the code available for strangers to inspect and improve. It was 2003, and he had spent enough time inside enterprise software companies to know the usual procedure. Build in private. Sell with conviction. Keep the machinery out of reach.

His proposed experiment disturbed that sequence. People would get to examine the machinery before anyone had persuaded them to buy it. Some would use it without paying. Some might help make it better. A business would have to emerge from those encounters. There are easier proposals to bring home.

Roberts described that six-month request years later, after the idea had become SugarCRM. The detail gives his story a useful starting point: a person with industry experience, a family and a limited stretch of time, trying to find out whether his irritation could become something other people wanted.

Six months to try it

He came from Washington, D.C., and earned a business degree at Virginia Commonwealth University. He graduated in 1990, worked at IBM in Georgia and felt drawn toward California. In his account, the move west came in 1995. Enterprise applications supplied his working education.

His career included Aurum Software, BroadVision and E.piphany. The jobs involved sales engineering, product management and product strategy. He learned how customers encountered a product and how companies decided what to build. At BroadVision, he directed the e-commerce product line. At E.piphany, he returned to CRM, the software used to organize relationships with customers.

What began to bother him was the production process around it. Companies assembled teams behind closed doors, developed applications and employed expensive sales organizations to carry them into the market. Roberts questioned whether that arrangement consistently produced the most useful software. He also questioned its geography. Why should a good idea need a Silicon Valley address?

Travel had made the available talent elsewhere harder for him to ignore. Opening development offered a way to include it. His complaint had become a plan.

“There are smart people all over the world.”

John Roberts, 2008

Three people, then the world

In April 2004, Roberts, Clint Oram and Jacob Taylor left E.piphany and committed themselves to SugarCRM. They initially financed the development themselves. Their combination mattered: these were colleagues who had already spent time building CRM, choosing to reconsider how such a product reached its users.

Their first months had a rather less ceremonial setting than the phrase “founding team” suggests. They worked remotely, coordinating through Yahoo! IM and internet voice calls. An office came later. SourceForge provided places to distribute files, report bugs and discuss requests. Much of the practical infrastructure was already waiting.

SourceForge selected SugarCRM as its October 2004 Project of the Month. Recognition arrived while the project was still young. The founders were already hearing from people putting early versions to work. Their audience could answer a question that a polished presentation could only pose: would anyone actually use this thing?

The name helped carry the intended mood. Roberts wanted business applications to feel more approachable. Salespeople, marketers and support staff had to spend time with them; making the experience less forbidding had a business purpose. A CRM system could look perfectly respectable while being quietly avoided by everyone expected to enter information into it.

Free to inspect. A business to run.

Opening a product does not settle the question of paying salaries. Roberts was explicit about the word “commercial.” He and his colleagues could work without pay for a time. A growing company could not treat that as a permanent operating plan.

By 2008, he described three editions emerging from a common source-code tree. Community Edition was free. Professional was sold by subscription. Enterprise supplied additional functionality, support and training. Customers could see source code, while commercial editions carried their own redistribution terms. The arrangement asked a practical question: which capabilities and obligations would organizations pay for?

It also let different kinds of user enter the same conversation. A small organization could begin with the community product. A larger buyer could want a supported commercial relationship from the outset. Roberts resisted describing the free edition as merely a trial. People could keep using it.

The same concern for choice extended to deployment. Hosted software suited some organizations. Others wanted it on their own infrastructure. Roberts supported both. The customer’s circumstances were supposed to influence the decision, even when the industry conversation preferred a single fashionable answer.

The crowd gets a vote

The international response gave the approach substance. In December 2007, Roberts reported 2,000 paying customers in more than 30 countries and translations into 75 languages. These were company-reported figures from that moment, rather than a census of everyone who had downloaded a copy.

That distinction matters. A download is an invitation accepted. A paying customer has made another commitment. Roberts needed a business capable of serving the second group while preserving reasons for the first group to participate.

He also saw public engineering as a means of attracting interest. In 2008, he described engineering salaries as contributing to marketing: publish useful work and people could discover it, share it and call the company. It was an argument about how spending on the product might help bring in the next customer.

Public participation brought disagreement along with distribution. SugarCRM’s earlier attribution-based license attracted criticism within the open-source community. People could admire the application and still dispute the conditions attached to it. A founder offering openness would be asked to defend the fine print.

In July 2007, SugarCRM announced GPLv3 for the forthcoming Community Edition 5.0. Roberts supported the change as being in the community’s interest. The episode complicates any tidy account of a founder simply giving something away. The terms needed work, and the people outside the company had opinions about that work.

“Let’s write software in public.”

John Roberts, 2006

Back to the customer, in Santa Cruz

Roberts’s first stint as a founding CEO ended in May 2009. He stepped down at SugarCRM, and board member Larry Augustin took over on an interim basis. The company said Roberts was pursuing other opportunities. The departure belongs in the chronology without requiring an invented private explanation.

Two years later, he founded X2Engine. He returned to the same category with a complaint about how complicated sales applications had become. The September 2011 public beta, X2Contacts, organized contacts and sales activity in a compact, blog-style interface. Familiar social features, including feeds and shared information, were part of its approach.

Santa Cruz supplied the setting. The team worked in the Cruzio building downtown and spent months developing the product before its public appearance. Roberts’s next attempt began with a small group working on a specific application, much as the first had.

The early release announcement included San Jose BMW Motorcycles as a customer design partner. That is a useful change of scale from arguments about the entire software industry. Somewhere, a sales team needed to make the next contact, remember the last conversation and avoid losing track of a prospective buyer.

By May 2013, X2CRM 3.0 included marketing automation, quotations and invoicing. A website visit could trigger a sequence involving a sales representative, a lead score and a follow-up message. The company reported more than 4,500 installations across 135 countries. Again, installations were the measure being reported, rather than paying accounts.

More capabilities created the familiar product problem: keeping the application manageable as it learned to do more. X2Studio let administrators alter fields and layouts or create modules. Customers could adapt the system to their own work. The effort to keep things compact had to coexist with the demand for flexibility.

Another connection from Roberts’s earlier career reappeared. David Buchanan, who founded Aurum Software, said Roberts invited him to advise X2Engine. Buchanan acquired the company in April 2017. A July release announcement identified Roberts as founder and vice president of engineering. The roles changed; the relationship had crossed more than one company.

In July 2017, X2CRM version 6.9 added deployment options for Amazon AWS and Google Cloud. The announcement described dedicated and shared cloud arrangements alongside downloadable software. That was another expression of a choice Roberts had discussed at SugarCRM years before. The hosting decision could follow the customer’s requirements. An application might live on someone else’s infrastructure while its buyer still wanted a say in how it was configured and run.

The name changes. The question remains.

In April 2026, SugarCRM announced a new name: SugarAI. Its CEO, David Roberts, described a direction in which software would help account managers identify where to focus and what to do next. John Roberts’s founding tenure had ended seventeen years earlier. The two Roberts belong to different chapters of the company.

For John, the through-line is visible across the work he actually undertook: applications people could customize, choices about where they ran, and repeated attempts to make CRM less cumbersome. X2Engine’s own later history continued to credit his 2011 founding. His first experiment had become the beginning of a second.

That is the appeal of the original six-month request. It gave an industry argument a deadline and exposed it to customers. Roberts could explain why the old process frustrated him. Then he and his colleagues had to ship something useful enough for strangers to keep.