Fetch: people-first editorial intelligence
Editorial writers know their stories better than anyone. Yet, standard analytics tools force them to build multi-level queries using unfamiliar jargon just to get basic answers. This product eschews the typical ‘wall of charts’ and puts writers and their content at the center, delivering answers in plain English.
This project comes from proximity rather than a brief. I spent years building and running content platforms alongside McKinsey's editorial teams, and the same exchange kept repeating: an author or stakeholder would ask a straightforward content question and I would go pull the answer for them. The analytics tool was built for analysts, which meant that the people closest to the work had the least direct access to how it was doing. This was my answer to quickly getting insights to both the authors and the stakeholders.
Same route, opposite defaults
The product has two audiences with opposite needs. An author wants their own work. An editor or leader wants the whole publication. Fetch does not split those into separate apps. It keeps one home route and changes the default by who you are.
In the demo, a persistent Viewing as: Author | Editor toggle sits in the header. Most portfolio demos describe a design decision. This one lets you flip it. Author view opens in second person with anomaly cards highlighting trends first. Editor view opens on the publication and puts filters up front. If something starts to look like a query builder or a chart dashboard, it is wrong.
What the demo shows
- Anomaly cards dynamically highlight emerging trends such as resurfacing content, underperformers, and new audiences, all written in plain language and linked to the story.
- Sparklines and horizontal reader breakdowns instead of pie charts. Every trend and bar has a text label, and color never carries meaning alone.
- AI-powered “Ask Fetch” absorbs search, meaning one input for both lookup and questions.
(Everything here runs on hardcoded synthetic data. No backend, no API calls, no browser storage. Names, headlines, and numbers are invented for this portfolio piece.)
What I cut
Three things came out between the planning sketches and the build. Each was removed for the same reason: it put back the complexity the product exists to remove.
Search. The first nav had Home, Search, and Ask. Search and Ask were answering the same question at different levels of specificity, so Search came out and Ask took the whole job. Typing “Dana” and typing “how did Dana do last month” now land in the same input.
The filter row under the author view. Planning put Author, Topic, and Date range directly below an author's own stories. That is the query builder again, only smaller, and it competes with the default the page just personalized. It became a single “Explore all content” link. The filters still exist, they are one step further away, which is where an escape hatch belongs.
A permanent Reset. The editor filters were drawn with a Reset control always visible. With three dropdowns resting at “All topics,” “All authors,” and “All time,” there is usually nothing to reset. “Clear filters” now appears only once a filter is set.
Demonstrable states
Real product states, with copy in the product's voice, are linked here for easy review:
- First run — nothing published yet
- Empty anomalies — a quiet, normal week
- Loading — skeletons at final dimensions
- Error — degrade gracefully, list still readable