Shotstack

Making a powerful product easier to enter

Video editing is familiar in principle and demanding in practice. Even experienced users have to find their way around footage, layers, timing, controls and output settings before they can make something useful. Put that experience in a browser, remove many of the conventions of an installed desktop application and ask a first-time customer to succeed quickly, and confidence becomes as important as capability.

Shotstack's editor brought a sophisticated, timeline-based video workflow to the web. The product needed to support serious creative control while remaining understandable to people encountering it for the first time. Every moment between opening the editor and seeing a finished result carried weight: if the next action was unclear, the depth of the tool could become a barrier rather than an advantage.

That tension sat at the centre of my role. As Senior Product Engineer, I worked across product design, engineering and growth, improving both the editor and the journeys that helped new users get something worthwhile from it.

Treating onboarding as part of the product

Onboarding is often added around a finished interface: a sequence of messages explaining what the product team has already built. That is rarely enough for a creative tool. Labelling every control does not give somebody confidence. Making progress does.

At Shotstack, onboarding was treated as an end-to-end product problem. The work considered the whole route through the experience: how a new user entered, what they understood, where uncertainty accumulated and what had to happen before the product felt useful rather than merely impressive.

Working across design and engineering meant I could move from observation to intervention without losing the context in between. Growth experimentation provided the feedback loop. The team could identify where users appeared to be struggling, make a focused change and see whether more people advanced.

We designed measurement into the journey from the outset. Funnel data showed where progress slowed, session observation revealed what that hesitation looked like, and support and usage signals added context. Together, those inputs helped us distinguish a confusing interaction from an unclear proposition or a poorly timed prompt.

Early platform choices were treated as provisional too. When the original onboarding-form experience stopped serving the journey, it was replaced within months rather than allowed to become a permanent constraint. The decision was guided by observed behaviour, not attachment to the first solution.

The question changed from “have we explained the interface?” to “are more customers succeeding?” Onboarding also stopped being a one-off launch task. The first-use journey became a measurable part of the product that could evolve alongside the editor.

Reducing the distance to the first win

Much of the editor's complexity was also the reason it was valuable, so removing it was not the answer. The useful design question was how to reveal that depth at the pace a new user needed it.

This meant paying attention to the sequence and clarity of actions rather than treating every available capability as equally important. The first experience had to establish a sense of cause and effect: bring material into the editor, make a meaningful change, understand the timeline and move towards an output. More advanced control could remain available without dominating the earliest decisions.

We had to balance guidance with agency. Too little structure leaves people facing an empty workspace; too much instruction turns a creative product into a rigid tutorial. I wanted to create momentum while preserving the feeling that the user was making the video, not following the software's script.

The clearest expression of that thinking was a three-stage “Getting Started” journey that replaced a wall of documentation with a sequence of achievable outcomes.

The first stage helped the user produce a real video from a working example and see the result inside the flow. The second made that output personal, introducing dynamic content through direct manipulation instead of explanation. The final stage carried the working setup into the full dashboard, so the transition into the product felt like a continuation rather than another beginning.

Progress was made visible throughout, with each completed stage reinforcing momentum. The important interactions were measured from launch, giving the team a clear view of where people advanced and where the journey still needed work. By the time a user reached the dashboard, they had already made something, changed it and understood the product's central promise through experience.

The change in behaviour was substantial. Activation increased from 15% to 40%, lifting the proportion of new users reaching the product's activation milestone by 25 percentage points. For a technically demanding creation tool, that could not be explained by better copy or a new walkthrough alone. Product design, engineering and measurement had all been pointed at the same outcome.

Building the editor around creative flow

Beyond onboarding, the work covered the interaction model at the heart of the product: arranging material over time, working across layers, handling assets, navigating efficiently with the keyboard, reversing changes and moving from an edit into an export.

These behaviours have to feel immediate because creative attention is fragile. A delay, an unexpected selection or an action that cannot be confidently undone interrupts the user's train of thought. That made the engineering inseparable from the experience design. Responsiveness, predictability and feedback were part of the product experience, even when the work behind them was deeply technical.

The browser added another constraint. A conventional page interface was not enough for the density and fluidity of a video timeline, so the editor used a specialised rendering approach for its most interactive surface. That choice created more freedom over performance and behaviour, while also requiring many familiar interaction patterns to be considered deliberately rather than inherited from standard web controls.

The product needed to carry that complexity so the user's attention could stay on the creative work.

Creating a shared product language

As the editor grew, consistency became a delivery problem as well as a design problem. Panels, inspectors, timeline controls, menus and dialogs all needed to behave like parts of the same tool. Without a shared foundation, each new feature risked introducing another pattern for users to learn and another one-off decision for the team to maintain.

I built a modular component system around the interface patterns that kept recurring. This was a practical investment, not abstraction for its own sake. It gave designers and engineers a common language, made established behaviour easier to reuse and created more room to focus on what was genuinely new in each feature.

That investment improved two forms of speed at once: the team could deliver more confidently, and users could move through the editor with less relearning. In a product with this much surface area, coherence was a feature.

From editing one video to orchestrating many

The editor was designed around the craft of making an individual video. By 2023, a different customer need was becoming visible: producing the same kind of output hundreds of times, responding automatically to events elsewhere in a business and combining several generative-media capabilities without manually moving work between them.

Workflows extended Shotstack from a creative surface into a visual automation product. Customers could assemble generation, transcription, audio and rendering steps into a reusable sequence, then run it with a single input, across a larger dataset or in response to activity in another system. Video production could become part of a wider business process rather than a task that had to be initiated and managed by hand each time.

The central experience decision was to make that complexity visible and manageable. A visual canvas showed the sequence of a workflow and how its parts related. A focused test action let customers validate the result with one real input before committing to a much larger run. That safeguard mattered because a mistake becomes far more expensive when it is about to be repeated hundreds of times.

As with the timeline, we designed the interactive surface specifically for the product instead of inheriting the assumptions of a generic diagramming tool. It required more investment up front, but gave the team control over how workflows were created, tested and understood. It also kept the interaction language consistent with the rest of Shotstack.

Workflows was also positioned as a step up in the commercial offer. Automation delivered a different kind of value from editing alone: repeatability, scale and the ability to connect video production to an existing operation. Packaging it accordingly made the product decision and the business model reinforce one another.

The timing mattered. Generative-media services were evolving quickly in 2023. Many were individually capable, but the ecosystem was fragmented. Workflows made an early bet that customers would need a dependable layer across those services instead of a growing collection of disconnected tools. It expanded Shotstack's role from a place where video was assembled to a system through which video production could be orchestrated.

The outcome

Across my 2022–2024 tenure, Shotstack gave me the opportunity to work on a production creative tool used by hundreds of people and to contribute at the point where product craft and commercial performance meet. By the end of that period, I had become the principal contributor across much of the dashboard's evolution, providing continuity between onboarding, editor behaviour, the shared foundations beneath them, and the later push into Workflows.

The clearest result was the activation increase from 15% to 40%. The broader contribution spanned the product's full arc: stronger first-use journeys, considered interactions across a demanding timeline, a shared interface system and a new automation layer that extended Shotstack into repeatable production at scale.

The experience still shapes how I approach growth. In a genuinely complex product, some of the strongest growth improvements come from making value easier to reach. The promise does not need to become louder when the product becomes easier to succeed with.