How do you create options when the original thesis weakens?

Odeo was pursuing podcasting when Twitter began inside it in 2006. Apple's move into podcast distribution was part of the competitive context. It is tempting to compress the story into a giant killing a startup and a visionary rescuing it. The public record is more crowded. Jack Dorsey, Biz Stone, Evan Williams, Noah Glass, and other contributors appear in accounts of the transition, with different roles and recollections. A useful case should preserve that complexity instead of awarding a single person an inevitable insight.

Dorsey's Stanford account describes staff breaking into groups to explore alternatives. In a 2007 MediaShift interview, Dorsey and Stone described a short prototype effort and subsequent internal use. The proposed interaction was limited: people could send brief status messages that others chose to receive. Rather than solve every communication problem, the prototype let the team observe a particular one. Could small updates create enough awareness and connection for people to keep contributing?

The constraint was part of the design. Messages had to fit the environment of text messaging, and participation could happen without writing a long post. That made the experience different from a full publishing system or a private conversation between two people. For a team whose earlier work involved audio, the new direction changed both the medium and the behavior being supported. Some expertise in communication and web services transferred; assumptions about the original product did not have to transfer with it.

The early experiment also brought costs into view. MediaShift's contemporaneous report noted unanswered questions about user numbers and revenue, alongside the expense of SMS traffic. This is valuable evidence precisely because it precedes the eventual global platform. People could find the interaction engaging before the company had demonstrated a durable business model. A leader examining an internal prototype needs to distinguish permission to learn from permission to scale indefinitely. Usage can justify further testing without settling the economics.

The interests around the experiment were not identical. Participants wanted reliable delivery and an audience they cared about. Engineers needed a service that worked as activity expanded. Existing investors had claims on the company and its assets. The people who built the product had interests in credit, authority, and ownership. Historical accounts describe Williams buying back Odeo's investor interests and the creation of Obvious, followed by Twitter becoming a separate company. Those organizational decisions belong in the story alongside the product decision.

The team made something small enough to be used, then exposed it to broader participation. It was not necessary to predict every later use of public short messages to learn from the first users. Nor does later adoption prove that the venture's governance was handled without conflict. The origins remain an example of experimentation within an existing organization, not a template for dismissing colleagues or treating prior stakeholders as obstacles to be bypassed.

A weakening thesis creates a need for options, not simply a new conviction. A short experiment can test a behavior that a strategy discussion would only speculate about. But an option needs boundaries: what resources it can consume, what evidence warrants continuation, and who has authority over the result. Twitter's emergence shows how a constrained interaction can open a much larger field of use. Product discovery and organizational legitimacy still need separate evaluation. The question is not whether every challenged company should hold a hack week. It is whether the company can learn something decisive before committing to another large, insufficiently tested direction.

For a discussion tailored to your team, explore a workshop with Julien.