How did LaunchDarkly learn what developers would pay for?

Warm relationships could hide weak demand

Harbaugh described a user named Grant refusing even a very low payment because his existing system worked better. UrbanSitter co-founder Andrea told her that basic releases mattered before experimentation. Harbaugh initially avoided charging to preserve pricing flexibility. The accounts differ on time to the first ten customers, so the evidence supports a slow, labor-intensive start without a single exact milestone. [1][2]

[Inference] These conversations changed the level of the problem being sold. Experimentation can sound attractive while a buyer is still struggling with release basics. A founder should test the customer’s immediate priority before asking them to adopt a more advanced practice. Warm introductions help secure attention, but a realistic payment decision provides evidence that goodwill and installations cannot supply.

The developer trial and the team purchase were different

In the WorkOS interview, Harbaugh said developers could always start a trial, while most sales required assistance because a manager or director needed to bring a team along. She recalled nervously asking a large prospect for $20 for the beta. That request was an early learning moment, not a verified standard enterprise price. The first paid account remains unidentified here. [2]

[Inference] A developer can establish technical fit, while release practices require shared responsibility. Product-led evaluation and human-led conversion can therefore serve different stages of the same account. Qualification should locate the organizational owner early and identify what must change for the whole team. Otherwise, a technically enthusiastic trial may remain disconnected from authority to adopt and fund the workflow.

The first seller relieved an observed constraint

Harbaugh said that at roughly eight employees she had become a bottleneck for demonstrations, quotes and legal redlines. The first salesperson came from New Relic, although the inspected interview did not name him. She transferred her field knowledge and described salespeople as advisers helping customers understand an unfamiliar category. Compensation and formal training details were not established. [2]

[Inference] This is a more concrete hiring trigger than reaching an arbitrary headcount. There was visible commercial work waiting for capacity. The hire needed to explain a practice, understand objections and complete the purchase process, not simply accept an order. Founder knowledge transfer mattered because early learning about the buyer otherwise risked remaining trapped in individual conversations.

Enablement makes release control repeatable

The November 2026 plan review showed free Developer, usage-based Foundation and custom Enterprise packages with unlimited seats. Service connections, client-side monthly active users and AI runs appeared as distinct usage units. The offer included Academy and certifications, guided onboarding, advisory sessions and enterprise customer-success managers. These are current boundaries, not the original beta terms. [3]

[Inference] Expansion could follow more applications, environments and workflows, but adoption requires ownership and governance. Unlimited seats can ease collaboration while usage billing shifts the commercial question toward deployment scale. Success work should establish who manages flags and how the team uses them consistently, rather than count SDK installation as completion. The account needs a sustainable practice before broader scope becomes useful.

Charge early enough to learn

[Inference] The strongest lesson is the correction from a compelling advanced idea to a more basic paid job. CAC, margin and cohort retention remain unverified. A better early test would combine a realistic commitment with observation of actual release work. As the offer broadens, the same discipline applies: prove a specific operating improvement, then earn the next expansion conversation. A friendly user’s enthusiasm is valuable feedback, but it is not a commercial outcome.