Fluctuation
Analysis
Catch the spike fast — then dig straight into why.
Crisis management is one of the biggest selling points of an enterprise social platform. Analysts needed to catch large fluctuations — good or bad — the moment they happened, and drill immediately into what was driving them.
Make the big moments impossible to miss.
In a sea of data, a large fluctuation has to announce itself — and then let the analyst go straight from “something happened” to “here’s exactly what.” A spike of positive posts might be a campaign landing; a spike of negative ones might be a crisis breaking. Either way, speed to the root cause is the product.
Large fluctuations visibly obvious in large amounts of data.
Drill straight into the root cause of the spike.
What even counts as a fluctuation?
This was the hardest thing to pin down, because the honest answer is it depends on who’s looking.A threshold that reads as a crisis for one brand is a quiet Tuesday for another. The conversation about which spike was “large enough” to trigger the analysis could have stalled the whole feature.
Simple, sufficient, complete.
I scoped the threshold using a framework our CTO used that I still reach for. It splits a feature into three honest phases — and it told me exactly where to draw the MVP line so an unanswerable question couldn’t hold the release hostage.
Simple — the MVP we shipped
Thresholds set programmatically by our data analysts. A spike of positive or negative posts past that line counts as a fluctuation. It ships, and it works.
Sufficient — the next version
Most functionality, launch-ready — moving toward giving users a say in the threshold.
Complete — the full feature
Users set their ownthresholds — granular control over what “large enough” means for them. The roadmap, made explicit.
Letting users set their own thresholds was clearly “sufficient/complete” territory — so the MVP set them for you. That framing kept us shipping instead of stalling.
The wireframe I designed and tested: summary, then the terms driving the spike.
A spotlight, and a way in behind it.
Surfacing the spike is only half the job. From the flagged fluctuation, the analyst drills into a summary — mentions, posts, sentiment — and then into the specific terms and soundbites driving it. The wireframe here is the artifact that shipped as the polished view up top.
Like Annotations, this was as much a visualization-engineering problem as a design one — and it pushed my working knowledge of Highcharts to surface and drill into spikes cleanly.
You can ship on a question that has no perfect answer.
“Simple, sufficient, complete” turns an unanswerable debate into a phased plan. Scope the honest MVP, name the roadmap, keep moving.