Re-imagining the shopping experience around bundles
I led the initiatives for a re-design of the Spectrum.com Packages page, a guided-first experience shaped by funnel data, competitive research, and a clearer decision model for customers.
DURATION
2 months
ROLE
Senior UX Designer - Lead
TEAM
PM, 1 Designer, Engineering, UX Researchers (2)
TOOLS
Figma, Adobe Analytics, Prototyping, Kiro AI
Tested · moderated, pre-launch · n = 6
5 of 6
felt confident in the recommendation
5 of 6
would buy straight from the recommendation
6 of 6
read it as help, not a sales pitch
Live · A/B test, 3 months post-launch
+12.6 pts
package selection rate
+1.6 pts
conversion after checking availability
+0.16
services in the average order
Tl;dr
Problem space
Shoppers were interested. The page just never helped them choose.
The Packages page treated a multi-step decision like a single click. Customers had to settle on internet speed, mobile lines, and TV all at once, with no running total, no sense of how the services fit together, and no guidance if they didn't already know what they wanted.
Most visitors engaged with the options and then left without choosing — and the ones who explored hardest, looking for the right fit, were often the least likely to buy.
Tap below
Showing Internet + Mobile bundles.
Internet + Mobile
Internet + Mobile
Internet + Mobile
Interactive recreation of the spectrum.com packages page. Pricing is illustrative.
Solution
Added clarity around what users want for bundles with an informed decision making process.
Instead of showing every option at once, the experience asks a few plain questions about the household — how many people use the internet, how many phone lines, whether anyone still watches live TV — and builds the package in front of you as you answer.
Core Experience
Building the package, one question at a time
The flow opens with a question about the household rather than a grid of plans. Each answer narrows the options and updates a running total, so the package assembles itself as you go.
Process
Secondary Research
Customers engage broadly and choose rarely, indicating that the current experience displays options rather than supporting a decision.
My secondary research consisted of reviewing prior UX research on the Packages experience alongside nine months of production analytics. Together they made the failure point obvious — almost everyone engaged with the options, and nearly half left without choosing anything. But why?
Packages page funnel — Aug 2025 to May 2026
Pulled 9 months worth of KPIs of the existing packages page on Spectrum.com
Engaged with the service options
84.1%
Went on to select a package
53.6%
Completed a purchase after checking availability
12.5%
Interest was never the problem. Nearly half of engaged visitors left without choosing anything.
What prior research already told us
Two unmoderated studies from June 2025 — 10 participants on desktop, 13 on mobile. Four findings shaped the concepts.
Verbatim · Mobile study · n=13
"I want to see my prices first to help me choose which package is right for me or else I'll just be wasting my time."
↓
Accurate pricing has to arrive before the decision, not after it.
Verbatim · Desktop study · n=10
"When I hear the term packages I think of predetermined bundles and things that are already put together at a discount."
↓
Packages means curated, not configured.
Finding · Desktop study · n=10
The majority preferred the layout that let them compare preselected plans side by side.
↓
Comparison is the job, not a feature.
Finding · Mobile study · n=13
Half of participants never noticed that pricing changed after they entered their address.
↓
Competitive Analysis
After understanding the experience of packages across various competitors, I found that they only serve as temporary solutions.
Every provider I audited leaves customers to assemble the package themselves, one product at a time. That works for someone who already knows what they want. It is precisely what our research showed everyone else struggling with.
Analysis
Every customer was handed the same single moment of choice, whether or not they knew what they wanted.
Three sources pointed at the same conclusion from different directions: nine months of behaviour on the live page, two rounds of prior research, and the constraints the buyflow imposed on everything downstream of a selection.
01 — Behavioral data
Decisive shoppers convert. Undecided ones leak.
02 — Prior research
Customers rated the page highly, then behaved otherwise.
03 — Business constraint
The page can start the decision, but it cannot finish it.
What the three had in common
The paths that worked belonged to customers who arrived already knowing. The paths that failed belonged to customers still working it out — and the page offered both the same thing.
Opportunity
Does an undecided customer need the page to narrow the field for them — or does the page simply need to make comparing the field good enough that they can narrow it themselves?
Ideation
After designing multiple phases of two experiences, we went into usability testing.
I sketched each direction far enough to be usable rather than persuasive — a guided flow that sequences the decisions, and a single page that makes comparing them effortless.
Final Experience
A few questions, then a package that tells you why.
People don't want a wall of plans. They want to be asked. A study on a similar tool found people already expect this — and that they will start it if you tell them it is short. So the flow opens with one question about your home, and shows how many are left.
6 / 6
found the questions quick and easy
6 / 6
read it as help, not a sales pitch
Mobile Breakpoint
5 / 6
would buy straight from the recommendation
Ending Notes
What I would carry into the next one.
This was a page that had already been redesigned once, with research that already existed and a solution already half-decided. Most of the work was deciding which of those things to trust. Three things stuck with me:
→
Research tells you what people say. The funnel tells you what they did. The design that tested best in 2025 is the one losing half its traffic today.
→
Building both was cheaper than arguing either. Two prototypes cost less than two more weeks of opinion.
→
A page can only promise what the next screen can keep. Knowing where my work ended was as much a design decision as anything I drew.
There is more here than fits on one page — the concepts that did not survive testing, and the handoff constraints that shaped the final scope. :)
←
Return back home
©2026 David Park
| Made with love in NYC
Thanks for being here.