NetBase
Enterprise
A high-precision engine for trained analysts — my job was making its power legible and shareable.
The platform had power to spare. It didn’t have reach.
Enterprise could process billions of posts and slice sentiment a hundred ways — but all of that lived with the trained analysts who ran it. A spike on a chart means nothing without context; a curated view means nothing if only the analyst can build it; a crisis means nothing if no one catches it in time. My feature work kept answering the same question in different shapes.
Giving the data context — right on top of the data.
Analysts sat on valuable data that was useless to anyone without the story behind it. Annotations let them mark a launch, a campaign, an event — so a spike could be read correctly by anyone with access. My first concept was a side panel of notes; it didn’t hold up. The stronger, riskier idea was annotating in-chart— something we weren’t sure Highcharts could even support.
Design for the best idea to pull engineering forward — keep fallbacks, don’t pre-compromise.
The full story — the rejected side panel, the in-chart bet, and pushing Highcharts past its defaults:
Read more about AnnotationsCarrying the data to people who’ll never open the platform.
A major global retailer needed social data in front of managers and C-level execs without giving them the raw Enterprise platform. It started as a simple emailed snapshot and grew — through ongoing work with the client — into a live dashboard of analyst-curated data. I made the call for live over a static report because it fit the real use case.
A one-customer feature that had to serve many — de-risked by pulling other clients’ feedback in early.
The full story — snapshot-to-dashboard, the live-vs-static call, and designing one-for-many:
Read more about the Executive DashboardSpotlighting the spikes that matter — good or bad.
Crisis management is a headline selling point for an enterprise social platform. Analysts needed to catch large fluctuations fast and drill straight into the cause. The hard question was the whole feature: what even counts as a fluctuation? — because the honest answer is it depends on who’s looking.
“Simple, sufficient, complete” — ship a threshold-based MVP now, hand users control later.
The full story — defining the undefinable, and scoping an MVP that could actually ship:
Read more about Fluctuation AnalysisHow I kept shipping on questions with no perfect answer.
Our CTO’s scoping frame: simple is the MVP, sufficient is enough to launch with most functionality, complete is fully realized. It let me scope an honest MVP when a question — “what counts as a fluctuation?” — had no perfect answer, instead of stalling.
Rather than pre-compromise a design to the platform’s current limits, I designed for the best idea and let it pull engineering forward — fallbacks in my back pocket. In-chart annotation is exactly that: a bet that stretched the team past building only to the defaults.
On a dense platform, the design work is making the power usable and shareable.
These features shipped with KPIs framed as goals — more data shared between analysts and non-analysts, faster understanding of fluctuations — rather than measured figures I can stand behind, so I don’t quote numbers here. The dedicated feature screens are illustrative reconstructions of the design decisions, not the original production UI.