DevCycle
Winter 2014AcquiredFeature Flag Management Built for Developers
Pressure-test this opportunity
Explore the risks and possibilities with a prompt for ChatGPT, Claude, or your agent.
Feature Flag Management Built for Developers
Explore the risks and possibilities with a prompt for ChatGPT, Claude, or your agent.
DevCycle was the second act of Taplytics, a Toronto company that began with mobile A/B testing and spent more than a decade controlling what software users saw. In 2022, the team turned that experience into a developer-first feature-flag platform built around the OpenFeature standard.[1]
Dynatrace acquired DevCycle in January 2026 because feature flags were becoming part of observability's control loop. DevCycle could change production behavior instantly; Dynatrace could measure the result. The deal validated DevCycle's pivot while showing why an independent flag dashboard was strategically incomplete.[2]
The company that became DevCycle started as Taplytics. Aaron Glazer led the original business, while Andrew Norris, Jonathan Norris, and Cobi Druxerman are the founders YC now associates with DevCycle. The team joined Y Combinator's Winter 2014 batch and initially sold A/B testing for mobile apps.[3]
Taplytics let a product team change an app experience, split users into cohorts, and measure behavior without waiting for a complete release cycle. By May 2015, it had expanded into analytics and push notifications. Glazer described the ambition as “a full-blown mobile optimization solution” and said the company wanted to “drive higher mobile engagement and higher monetization through the entire user lifecycle.”[4]
The company raised a $2.4 million seed round from YC and a long list of operator-investors, including Alexis Ohanian, Paul Buchheit, Mark Pincus, and Kevin Hale. Glazer said Taplytics had delayed announcing the already-closed round because “we’ve just been putting all of our resources towards product development.”[4]
Over time, the mechanism beneath experimentation became the larger opportunity. Feature flags decoupled deployment from release: teams could ship dormant code, expose it gradually, test variants, and turn it off without another deployment. In March 2022, Taplytics raised $5 million to launch DevCycle as a separate feature-management suite for engineering teams.[1]
DevCycle gave developers a control plane for features already present in deployed code. An engineer wrapped a behavior in a flag, created targeting rules, and exposed the feature to selected environments or cohorts. The team could increase exposure gradually, compare variants, schedule a release, or disable a problem without rebuilding the application.
The platform included SDKs for common languages, centralized flag configuration, real-time updates, permissions, audit logs, experiments, reusable targeting, and integrations with GitHub, Jira, Terraform, Slack, webhooks, and Vercel Edge Config.[6] The product also emphasized flag cleanup, a necessary defense against old branches becoming permanent code debt.
Its decisive design choice was OpenFeature. The CNCF project standardizes the application-facing API for feature flags, separating instrumentation from the vendor that evaluates a flag. CNCF accepted OpenFeature in June 2022 and promoted it to Incubating in November 2023.[5] DevCycle publicly committed to the standard across its supported languages and onboarding in 2024.[7]
That openness reduced customer lock-in, but it also made DevCycle easier to combine with another control system. Dynatrace could consume OpenFeature data and join a flag change to latency, error, or conversion signals without forcing customers to replace every SDK.
DevCycle sold to engineering, platform, and site-reliability teams that released frequently enough to need runtime control. Taplytics had originally addressed product and marketing teams; DevCycle moved the buyer and daily workflow closer to code, CI, and incident response.
The company did not publish a defensible category-size estimate. Its scale claims were operational: the 2022 launch said Taplytics had delivered billions of feature flags and supported customers such as Grubhub, Warby Parker, and RBC Royal Bank.[1] YC later described millions of end users and billions of events, but those are vendor-supplied figures rather than audited financial measures.[3]
DevCycle competed with LaunchDarkly, Split, Flagsmith, Unleash, cloud-vendor configuration products, and internal systems. Proprietary vendors could win on mature enterprise workflows. Open-source tools could win teams that valued control and low cost. A homegrown boolean table looked cheap until it needed low-latency evaluation, targeting, auditability, SDK coverage, and safe failure modes.
OpenFeature changed the axis of competition. Vendors could no longer rely as heavily on SDK lock-in, so value moved toward workflow, governance, experimentation, and the data used to judge a rollout. DevCycle embraced that shift. Dynatrace owned far more telemetry, which made it a natural buyer.
DevCycle sold hosted feature management with free and paid tiers, though no reliable historical price schedule or revenue breakdown was located. Taplytics raised $2.4 million in seed financing and later announced $5 million specifically for DevCycle's launch and growth.[4][1]
The economic challenge was structural. Flag evaluation is high-volume infrastructure, while developers expect low latency and generous usage. Enterprise revenue therefore depends on governance, support, audit controls, experimentation, and integrations rather than the toggle itself. Public sources do not disclose revenue, burn, profitability, or acquisition price.
Taplytics named Kobo, Shyp, Frank & Oak, and Wave as early customers in 2015.[4] The 2022 DevCycle launch cited Grubhub, Warby Parker, and RBC Royal Bank and claimed billions of delivered feature flags.[1]
Dynatrace kept the product operating after acquisition and began integrating its capabilities. DevCycle co-founder Cobi Druxerman wrote, “the DevCycle you know and love today will be here tomorrow,” while telling customers that functionality would move into the Dynatrace platform over time.[8]
Taplytics began as an experimentation interface for product teams. DevCycle isolated the underlying runtime-control system and sold it to engineers, who felt the cost of risky releases directly. The pivot narrowed the buyer but deepened the workflow: flags became part of deployment, incident response, and code cleanup rather than a marketing experiment screen.
OpenFeature weakened a common SaaS defense because customers could change providers without rewriting every application call. DevCycle accepted that cost and competed on execution. The decision widened its integration surface and aligned it with Dynatrace, which helped establish the standard in 2022.[2]
This was not charity. A standard can commoditize the interface while increasing the value of the best control and data layer. DevCycle became easier to plug into enterprise environments; Dynatrace could buy that control layer without asking customers to abandon vendor-neutral instrumentation.
A feature flag can change who sees code, but it cannot judge whether the change is healthy. Dynatrace can correlate the rollout with errors, latency, and business outcomes. Its acquisition note described flag systems as standalone controls that often lacked direct visibility into consequences.[2]
The non-obvious mechanism was the split between action and evidence. DevCycle owned fast action; observability vendors owned the evidence needed to automate that action safely. TechTarget placed the deal within a broader convergence of observability and feature management.[9]
The countercase is straightforward: DevCycle could have remained an independent, open-standard vendor and connected to many telemetry systems. Public sources do not reveal revenue, growth, or acquisition terms, so the sale cannot be scored financially. The strategic logic is clearer. Druxerman wrote that the team had built feature-flagging and experimentation software for 13 years, and Dynatrace offered a way to join that control with production intelligence.[8]