Cadencea case study
SharePoint Pilot · UX consulting

Cadence

Rework a team’s ignored portal around a new way of working — and leave behind a blueprint the company could repeat.

Cadence’s QA team had an internal portal that was cumbersome and mostly ignored. As UX consultant I led the pilot that reworked it around community and file sharing — and, just as importantly, produced a repeatable design and migration blueprint the rest of the company could follow.

community portal — production skin
Cadence|SharePoint
Search this site
BU
CQIQuality Community PortalQA Organization☆ FollowingShare
HomeCommunitiesDocumentsSupportAbout
+ New↗ Send by emailⓘ Page details✎ Edit
Community Portal
Find the teams, discussions, and shared files that move QA’s work forward — and jump into the communities that matter to you.
FEATURED
Test Automation
Frameworks, CI, and shared harnesses.
👥 214 members · 💬 1,120
FEATURED
Release Quality
Sign-off criteria and gate checklists.
👥 168 members · 💬 842
FEATURED
Defect Triage
Reproduction, severity, and ownership.
👥 203 members · 💬 977
What’s HotRecent Communities
Performance & Load
Benchmarks, soak tests, and dashboards.
142 members · 💬 610
Test Data & Envs
Fixtures, seeding, and environment parity.
96 members · 💬 388
Accessibility QA
Audits, tooling, and remediation playbooks.
118 members · 💬 502
UX designInformation architectureCard sortingContent & functionality auditMigration strategyCadence Design Systems
01The gap

The team had a portal. Almost no one used it.

When I came on, the QA team already had an internal portal — the problem was that almost nobody used it. It was cumbersome and mostly outdated, so the broader team had never really adopted it: too complex to be worth the effort. A place meant for collaboration had quietly become a place people worked around.

Cumbersome

Hard to navigate, harder to contribute.

Outdated

Content no one trusted or maintained.

The real failure wasn’t the software — it was that the team had no shared way of working in it.
02The problem

A version upgrade wouldn’t fix adoption.

IT wanted to move the whole company onto a better setup, and the temptation was to treat this as a version upgrade — lift the content, drop it somewhere new. But moving a cumbersome site somewhere new just relocates the problem. Adoption is about how people work — not where the files live.

Migrate the content

Lift the files, call it done.

Drop them on a newer setup — and inherit the same low adoption.

Change the way of working

Rebuild around how the team collaborates.

So people want to live in it — and it works for QA and scales to every other team.

And the company was far too big to prove that everywhere at once.

03The goal

Two outcomes, one pilot.

So the team ran a pilot: pick one content-heavy, well-defined group — QA — rework their portal end to end, and come out with two things at once.

Outcome 01

A portal they live in

QA’s site reworked around community and file sharing, so finding and sharing work is easier — not just relocated.

Outcome 02

A repeatable blueprint

A design and migration approach that hands cleanly to the next team, and the next after that.

Two principles held it together → every choice had to be scalable, and the new site had to be genuinely easier than the old one.

04Hypothesis

Two beliefs to make true.

We framed the work around two beliefs — and the design’s whole job was to make both of them true.

Belief 01

Adoption rises if it’s built around how people work.

Communities, discussions, shared files — not a faithful replica of the old tangle.

Belief 02

Prove it with one team, and it justifies the whole company.

One content-heavy team is enough to earn the company-wide rollout.

05The assumption

“We understand this well enough to build it.”

The biggest assumption baked in at the start was that we understood this new way of working well enough to build it. I was skeptical — I was living the gap firsthand. So instead of papering over it, we closed it: the team found room in the budget to bring on two Microsoft consultants, to work alongside us.

Why bring them on

Truly understand the platform’s capabilities.

And just as much

Make sure everything we built would scale across the entire company.

Name the belief you’re least sure of — then spend to make it true.
06The audit

Content and current-functionality audit.

My hands-on work was the UX spine, and it started with an audit — three moves to find out what the site really was before deciding what it should become.

Spidered the site
Crawled the entire QA portion of the existing installation.
Interviewed QA
Learned what the current intranet was missing, from the people using it.
Walked leadership
Sorted what was outdated from what genuinely needed to move.
The finding that reframed the assignment: the site’s complexity was the cause of its low adoption — so the pilot was a chance to fix that, not just to move it.
07Card sorts

Two card sorts rebuilt the structure.

With the audit done, I moved into information architecture — deciding how QA’s content should be grouped and named, instead of replicating the tangle that had failed to catch on. I ran both an open and a closed card sort with the team.

Their words · open sort

Let the team group and name content themselves.

Surfacing the real mental model, not the one the old site had imposed.

The test · closed sort

Checked content against the proposed groupings.

Validating the categories held up under real content.

The result → a grouping-and-naming backbone the site — and the rollout — could be built on.

08Building blocks

Two screens carried the whole argument.

Working with the platform’s existing modules plus some customizations — and handing the build to the Microsoft consultants — the design converged on two core screens. The ones that carried the argument for the whole migration. Everything else is a refinement of these two.

Screen 01

Community portal homepage

Featured communities up top; a What’s Hot / Recent toggle; each tile carrying its own member and discussion counts, so the space feels alive rather than archival.

Screen 02

Discussion overview page

A highlighted best reply, up/down voting on responses, a Top Contributors rail, and sort controls — so a busy thread stays navigable.

The two screens

Wireframe first. Then a production SharePoint skin.

Each screen shown twice: the low-fidelity wireframe that proved the structure, and the same screen taken to a production skin built in the SharePoint Web UI Kit — the platform’s real visual language, with the site header, hub nav, command bar, and web parts.

09 → 12Community portal homepageBefore · wireframeAfter · SharePoint
before · wireframe
CQI Portal - Windows Internet Explorer
◄ ► ✕ ⌂
http://www.cadence.com/
Secondary NavigationBob User⚙ ?
BrowsePage
Cadence Logo
Home
Link
Link
Home › 123 › 456 › 789
Community Portal
Search Communities  ⌕

Introduction Text goes here. Something about the purpose of this specific community and what you’ll find here

Featured Community
Description
<<???>> members
<<???>> discussions
Featured Community
Description
<<???>> members
<<???>> discussions
Featured Community
Description
<<???>> members
<<???>> discussions
What’s Hot  |  Recent Communities
Community Name
Description
<<???>> members
<<???>> discussions
Community Name
Description
<<???>> members
<<???>> discussions
Community Name
Description
<<???>> members
<<???>> discussions
Community Name
Description
<<???>> members
<<???>> discussions
Community Name
Description
<<???>> members
<<???>> discussions
Community Name
Description
<<???>> members
<<???>> discussions
after · sharepoint web ui kit
Cadence|SharePoint
Search this site
BU
CQIQuality Community PortalQA Organization☆ FollowingShare
HomeCommunitiesDocumentsSupportAbout
+ New↗ Send by emailⓘ Page details✎ Edit
Community Portal
Find the teams, discussions, and shared files that move QA’s work forward — and jump into the communities that matter to you.
FEATURED
Test Automation
Frameworks, CI, and shared harnesses.
👥 214 members · 💬 1,120
FEATURED
Release Quality
Sign-off criteria and gate checklists.
👥 168 members · 💬 842
FEATURED
Defect Triage
Reproduction, severity, and ownership.
👥 203 members · 💬 977
What’s HotRecent Communities
Performance & Load
Benchmarks, soak tests, and dashboards.
142 members · 💬 610
Test Data & Envs
Fixtures, seeding, and environment parity.
96 members · 💬 388
Accessibility QA
Audits, tooling, and remediation playbooks.
118 members · 💬 502

Structure first: a row of featured communities up top, a What’s Hot / Recent Communities toggle over a grid of community tiles, each showing member and discussion counts — the whole layout signalling activity so the portal reads as a living place. The production skin dresses that same structure in the platform’s real visual language: site header and hub navigation, a command bar, featured community web parts, and a What’s Hot pivot over live cards with a face-pile of members.

10 → 13Discussion overview pageBefore · wireframeAfter · SharePoint
before · wireframe
CQI Portal - Windows Internet Explorer
◄ ► ✕ ⌂
http://www.cadence.com/
Secondary NavigationBob User⚙ ?
BrowsePage
Cadence Logo
Home
Link
Link
Home › Products › Xyz › Features
Community Name
Question Goes here
Person's name

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec justo arcu, vehicula suscipit laoreet vitae, congue in elit. Nulla gravida tempor adipiscing. Sed quis lorem nunc, id dapibus orci.

<<post time>>   like   reply
Best Reply <<who it was chosen by>>
Person's name

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec justo arcu, vehicula suscipit laoreet vitae, congue in elit. Nulla gravida tempor adipiscing. Sed quis lorem nunc, id dapibus orci.

<<post time>>   like   reply
All Replies
A-ZZ-ARecentPopular
Person's name

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec justo arcu, vehicula suscipit laoreet vitae, congue in elit. Nulla gravida tempor adipiscing. Sed quis lorem nunc, id dapibus orci.

<<post time>>   like   reply
Search this site… ⌕
This post
# Views
# Replies
# likes
Best Reply
Top Contributors
Person’s name
Person’s name
Person’s name
More…
after · sharepoint web ui kit
CQIQuality Community PortalBU
HomeCommunitiesDocumentsSupport
Home  ›  Communities  ›  Test Automation  ›  Flaky UI tests
24
Test Automation
How do we stabilize flaky UI tests in CI?
RNRiya N.asked 2 days ago★★★★☆
Our Selenium suite fails intermittently on the same tests — usually timing. What patterns have actually worked for your team to make CI runs deterministic?
✓ BEST REPLY
41
TMTheo M.
Isolate state per test and wait on explicit conditions, never fixed sleeps. We cut flakiness ~80% by adding retry-on-idle and pinning the test data seed.
All replies
A–ZZ–ARecentPopular
12
AKAmir K.
Quarantine the flakiest specs into a separate lane so they stop blocking the main pipeline while you fix them.
About this discussion
1.2kviews8replies34likes
Top contributors
TMTheo M.2,140
JPJess P.1,880
AKAmir K.1,502

Modelled loosely on the Stack Overflow pattern: a question at the top, a highlighted best reply, up/down voting on every response, sort controls, and a right rail carrying the post’s stats and its top contributors — so a long thread stays navigable. In production, that pattern renders as a real SharePoint page, the accepted best reply called out in green. This is the screen that made community feel worth showing up for.

11The challenge

Every choice had to scale.

The hardest constraint wasn’t any single screen — it was that every decision had to be scalable. A design that worked beautifully for QA but couldn’t be carried across every other team would have failed the actual goal. That constraint sat behind every layout and content choice.

Great for QA. Impossible to repeat. — the failure mode we designed against. The KPI wasn’t a vanity metric; it was that the pilot be delivered in a way that was scalable across the entire company.

14Implementation

Split cleanly along expertise.

Implementation split cleanly along expertise — then converged.

Microsoft consultants

Offloaded the pilot content onto the new server.

All the content tagged for the pilot, moved off the current installation.

Me

Built the site infrastructure off the card sorts.

The open and closed sorts became the structure directly.

Then, as a team, we populated the content and converged on one consistent design concept that could carry across the entire rollout.

15Results

It launched as my engagement wrapped.

I’ll be honest about the ending. The QA quality portal went live on the new setup — right before I rolled off the engagement. So I saw it ship and the migration approach take shape, but I left before the longer-term adoption and rollout numbers came in, and I won’t claim results I didn’t see. What I can say is that the pilot did the two jobs it was scoped to do.

Delivered 01

A live portal

A community-driven quality portal QA could actually work in.

Delivered 02

A repeatable blueprint

A design + migration approach ready for the broader rollout.

16Learnings

Structure and process, over any one screen.

IA at real scale

The card sorts and content audit did the heavy lifting.

They were the difference between a site QA would adopt and one they’d quietly ignore.

A repeatable pilot

The deliverable wasn’t one good screen.

It was a process solid enough for other teams to repeat without me in the room.

The lesson underneath both: when you introduce a new way of working, name the assumption you’re least sure of — and spend to close it.
Close

An ignored intranet, reworked into a portal a team lives in — and a blueprint the whole company can follow.

The lesson that outlasts the project: adoption isn’t about where the files live — it’s about designing the way people work together, and making that repeatable.

Honest framing: I rolled off as the pilot launched, so no adoption or rollout figures were captured and I don’t quote any — the outcome was a live site and a repeatable blueprint.