Enterprise Search

The Problem

Every tool at Nike had its own search box, and none of them talked to each other. Time Off, The Source, Sedgwick, The Van, Passplay, Workvivo, Zero, dozens more, each an island. To find a person, a policy, or a location, an employee had to already know which of thousands of apps to open first, then search again once inside it. Often they just gave up.

I had been prototyping the fix since before this project had a name, in employee-search concepts that never shipped. This was the work that finally put it in front of the whole company: a 2,600-app sprawl narrowed to one search bar, on an intake system deliberate enough to keep growing after I handed it on.

Process

Nike's internal audit of its own digital footprint stopped counting at 2,600+ sites and apps, different shapes, technologies, ages, and sizes, built up over decades with no central inventory. Search could not wait for that inventory to finish.

Before I designed a single screen, I sat with employees across roles and watched how they actually hunted for things, and three habits stood out: the constant tab-switching, the ask-a-colleague fallback, and the quiet giving up. Each one pointed at a decision. Tab-switching called for one federated box in place of dozens. The ask-a-colleague habit set the core bet: when the fastest path to a policy is the person who owns it, search should lead with that person, not bury them under the document. And the quiet giving up called for putting results where people already worked, so one index could surface inside many Nike apps, not just the standalone search.nike.com. The people-first bet was not a fresh guess: I had been prototyping it since 2018, in an earlier Digital Employee Experience concept shown further down.

That last move, one index feeding many surfaces, only held up with a system behind it. I started with roughly a dozen systems, chosen for how essential the information was and how wide the audience reached, and built the machinery to keep adding the rest: an intake process, a data-cleanliness process, and a data-structure-mapping approach that could absorb a new source without a bespoke integration each time.

The architecture itself was federated: every source kept its own data structure, but each result surfaced through one consistent, Nike-branded interface. I designed each result type to behave like the thing it represented:

• People cards built around real athletes (Serena Williams, LeBron James, Michael Jordan) as stand-in profiles
• Location results with integrated maps and venue details
• Content cards with FAQ accordions, related content, and right-rail context panels
• Mobile-first responsive layouts maintaining full functionality on phone

Every round started from a real employee question, not a lorem prompt. One of them, "Where can I get my badge?", carried through three directions: a first pass buried under a green ServiceNow header, then a color-coded results layout tagged by source, then a refined version that moved the person and the answer to the top. I labeled its lead tab simply "Human." One word, and it carried the whole argument the project was making: lead with the person, not the paperwork.

The result types and the one box did not go straight from my screen to the whole company. They were pressure-tested first through Experience Insiders, a standing volunteer panel that ran a search user-flow A/B test on members' own time. That panel is where the people-first ranking and the result layout had to prove out, against how employees actually searched rather than how I assumed they would, before any of it went company-wide.

Solution

Type parental leave once, and the answer came back from wherever it lived: the policy itself, the person in HR who owns it, and the benefits office where you could go to ask, all in one ranked list. No more guessing which of thousands of apps to open first.

Under that one box sat a federated search that queried across the integrated systems and grouped what came back by what each result actually was, a person, a place, a policy, a piece of content, so the ranking could lead with the most useful kind instead of blending everything together. People and places carried their own card types: a profile you could act on, and a location with a map and directions, each built responsive for desktop and phone on Nike's NCSS design system.

The third bet, results where people already worked, is easiest to show in motion. This November 2021 interaction prototype, built in Bowerman, runs an Upcoming Events module of the internal experience: hover a card and the way deeper is right there, "Learn more on Workvivo," the app that owns the event. Content met employees where they already were, one hover from its source.

Impact

I designed and shipped the first wave, roughly a dozen systems, and built the intake-and-mapping machinery under them to scale past any one person. By the end of 2021 that machinery had carried the index to 72 systems, surfacing inside the apps employees already opened every day, so finding a colleague, a policy, or a location stopped being a separate errand. The clearest vote of confidence was organizational: the work outgrew its start and earned its own dedicated product and engineering team, at which point I moved into a strategic and service-design consulting role elsewhere at Nike, including redoing the global HR website and its brand.

I cannot point you to a clean before-and-after on time-to-find. These internal tools were never instrumented for it, and a wave of reorgs moved teams and priorities before that measurement loop ever closed. The honest outcome is reach and durability: the search spread across the company and became a capability worth staffing as its own team, not a project that ended when I moved on.

Key Deliverables

• Federated search UI with dedicated result templates for people, places, and content
• Athlete/person profile cards with stats, initiatives, and overview tabs
• Desktop results with left-rail filters and right-rail context panels
• Mobile-optimized search with touch-friendly cards
• FAQ accordion component with structured content
• BetaFrames comprehensive design reference (20+ screen states)

Meet the Team (a concept I explored)

I kept coming back to one idea that did not make the final product but still says something about how search should work: when you land on the content you were looking for, surface the people behind it too. Find a project, and see who works on it, with a face, a name, and a way to reach them. A document is rarely the thing you actually need; the person who can act on it is.

People in search, since 2018

The "people behind the results" concept above did not start with search.nike.com. I had been prototyping it since 2018, in an earlier proof of concept for Nike's Digital Employee Experience, before any company-wide employee search existed.

The DEE search POC tested a simple bet: one query should return people, policies, shopping, sports scores, and news together, and it should put the people first. Searching "Design" pulled a live NHL card, Nike.com results, a North America parental-leave policy, and a column of employee profiles into one view. Federated search, years before the production version shipped.

The boldest part was the profile card. Instead of returning a document and leaving you to track down whoever owns it, the POC surfaced the person behind it, with their role and a way to reach them right there in the result. (The cards ran on placeholder data, my own details and a teammate's, standing in until a real directory was wired up. The teammate's name and contact details are swapped for a fictional stand-in here.)

Two parallel prototypes pushed the same thesis from different angles. The HR Contact Hub reframed "contact HR" as a search-first experience: type what you need, and get answers back as cards instead of a phone tree.

The HRBP Directory was the most literal version of the idea. Enter an employee's ID and it returned the exact HR Business Partner who supported them, plus named backups, each as a contact card. Finding the policy was never really the goal. Finding the person who could apply it was.

None of these POCs shipped as they were. But the bet they tested, that finding information is really about finding the person who can help, is the same one that carried into search.nike.com and the Meet the Team concept above. That thread runs straight from a 2018 prototype to the one search bar Nike put in front of all 86,000.