Case study · 02 / 06 · 20254 min read
  • Conversion
  • Monetization
  • Design

Fullbay Connect

How do you turn a half-decade-old Integrations page into a scalable, subscription-tiered service?

Where
Fullbay. Shop management software for heavy-duty repair.
Role
Product designer. UI, the tiering model and its screens, cross-team delivery. The executives set the price points.
The shipped Fullbay Connect page: connections grouped by job to be done, with plan badges and the Upgrade Plan button in the header.
$1.1M revenue growthNot design-only

At a glance

  1. The problem

    A five-year-old Integrations page with 2,000+ shops on it and a pile of new third-party connections coming that it could not absorb.

  2. What I did

    Rebuilt it on the components the rest of the app already used and collapsed several branded buckets into one tiering model, with plan state and the upgrade path on the page.

  3. What happened

    $1.1M in one quarter's revenue growth attributed to the launch by the company. No holdout group. The size is the company's number and the direction is what the design can point to.

01Scope

  • One quarter's revenue growth from the Connect page, attributed to the launch

    Up, the quarter after launch

    The company's figure, reported against the launch. A redesign and a new tiering model shipped at the same time, so the $1.1M is not the design's alone.

  • 2,000+

    shops the page had to serve

  • 3

    problems reframed into design goals

  • 1

    legacy page rebuilt from the ground up

02The challenge

The Integrations page was five years old and visually inconsistent, and it was straining under the 2,000+ shops it served. It also had to absorb far more third-party connections in the near future without collapsing again.

This was a UI and monetization project. The research was competitive analysis and the usage numbers already in the product, because there was no budget for more. The usage numbers are where the underperformers came from. A few served very specific enterprise workflows, so those accounts kept their integrations as legacy and the bloat came off the page for everyone else.

The page as it was. Every connection under its own branded heading and settings scattered per card, with nothing to say which plan you were on.

03The goal

How might we make integrations simple and modern?

Two visual languagesOne

Rebuild on the Bootstrap 5.3 components already in use across the non-legacy parts of the app, rather than maintaining a second visual language.

Everything to every shopOnly what applies

Make the key features findable and retire the underperformers. Show each shop only what applies to it.

Branded bucketsOne tiering model

The page had marketing language and several Fullbay-branded buckets doing duty as price brackets. Collapsing those into one coherent tiering model was the design problem.

04The solution

Fullbay Connect

One page, grouped by what a shop is actually trying to do: payments, marketing and communication, parts and labor. Plan state, entitlement and upgrade path are in the same view.

Collapsing the branded buckets into four plans left some shops entitled to less on paper, and the plans were arranged so the loss was compensated. What a shop actually used moved into the plan it landed on. QuickBooks was the most used connection by a distance and stayed reachable everywhere. The enterprise shops did not mind paying more for the feature sets in the higher plans, and new integrations went into Starter, Pro and Elite so the step up bought something visible.

Fullbay Connect integrations page, showing payments and marketing connections grouped into cards with per-integration toggles and an upgrade path.
Connections grouped by job to be done, with plan state and upgrade path in the same view.
The rebuilt page, in motion. Payments, marketing and communication, parts and labor, with the toggle and the upgrade path on each card.
The shipped Manage Your Plan screen: Starter, Plus, Pro and Elite side by side, every integration marked per plan, Downgrade, Current Plan and Upgrade under the prices, and an annual pricing toggle above.
The tiering model as it shipped. Four plans, every integration marked per plan, and the move up or down on the same screen as the entitlement. The price points were set by the executives. The grouping, the model and the screens are mine.

05What I dropped

  • A second visual language, a fresh system just for this pagethe non-legacy parts of the app already ran on Bootstrap 5.3. Rebuild on that.
  • The Fullbay-branded buckets doing duty as price bracketsnone of them fit the tiering model. They were marketing copy.

06The impact

The release video Fullbay shipped it with. About a minute, with sound.
ImpactEvery figure with what it is not
One quarter's revenue growth from the page, the quarter after launch$1.1Mup
Churn across the affected base, the same quarter+1 ptupUp by about a point across the shops the tiering touched, in the same quarter as the revenue. The level is not printed, because I no longer have it to a confidence worth printing. A point of churn on a base whose pricing had just changed shape is what I was watching for.
Overclaims$0.00

Corrections & caveats

Variables
A redesign and a new subscription tiering model shipped at the same time, so the offer changed alongside the interface.
Cohort
The quarter after launch. The revenue figure is the company's, reported against the launch rather than against a held-back control group.
Confidence
I can point to the direction and not the size. This shipped as a redesign and a new subscription tiering model at the same time, so the offer changed alongside the interface. The $1.1M is not the design's alone.
Not measured
  • No holdout group, so there is no counterfactual to compare against
  • How much of the lift is new tier adoption versus existing shops upgrading
  • Longer-term retention on the new tiers

The page had been a five-year-old liability nobody wanted to touch. We put a couple more connections on it before I left, and the tiering worked as one model instead of several competing ones.

07What I would measure next

  • Holdoutvs. controlThis launch had no counterfactual. Hold back a slice of shops so the lift has one.
  • New-tier adoption% of upgradesSplit from existing shops moving up. The number I could not separate.