↑
Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing Case StudyManufacturing
Donaldson • United States

A New Revenue Stream, Built on Data

Donaldson put sensors in its products and started selling customers iCue, a subscription service that uses data to enable machine health monitoring. SunnyData rebuilt the system behind it on Databricks in 10 weeks, and cloud costs are expected to drop 55%.

55% Lower cloud costs Expected reduction in total Azure spend.
10 weeks From start to go-live Delivered in three phases.
3 to 1 Azure tools replaced by one platform One place for data, alerts and reports.

The Opportunity

A product sold once. Data sold every month.

Donaldson makes industrial air filters. Its plan was to put a sensor in every product and sell customers a subscription, iCue. The service watches their equipment in real time and sends alerts before it fails.

That turns a one-time sale into recurring revenue. It also gives customers fewer surprises and less downtime.

  1. Sale
  2. Month 2
  3. Month 3
  4. Month 4
  5. Month 5
  6. Month 6

One payment

Illustrative. Not client revenue data.

Client Challenge

A new revenue stream, running on an unreliable system

The system behind the subscription service for clients was built on Azure, and it wasn't ready. It used three separate tools that broke often, cost too much and copied the data at every step.

Data scientists couldn't easily test new ideas. For a service customers pay for, that wasn't good enough.

Two engineers review a Donaldson dashboard showing a map of connected equipment across the United States

The Solution

One platform, from sensor to screen

SunnyData moved everything onto Databricks,
so data now flows from the equipment to the customer in one place.

Collect

Live readings come in from every device.

Clean up

The data is organized and checked as it arrives.

Deliver

Customers see live dashboards and get alerts before something fails.

Keep it running

Updates are tested automatically, and SunnyData stabilized the system after launch.

What it runs on
Delta Lake logo
Delta Lake One trusted copy of the data.
Databricks logo
Databricks SQL The reports customers see.
Databricks Declarative Automation Bundles + Azure DevOps Safe, automated updates.

How It Works

From sensor to decision, with no manual steps

Follow one reading through

Five stages, no manual steps

Illustrative READING STATUS Normal Alert: service needed soon
01EquipmentA sensor reports its status.
02Live data feedReadings are collected as they happen.
03DatabricksData is cleaned and checked against alert limits.
04Customer portalLive dashboards and reports.
05Facility managerGets the alert and plans the fix before a breakdown.
01
Equipment
A sensor reports its status.
02
Live data feed
Readings are collected as they happen.
03
Databricks
Data is cleaned and checked against alert limits.
04
Customer portal
Live dashboards and reports.
05
Facility manager
Gets the alert and plans the fix before a breakdown.

Early warnings

Alerts go out before a small issue becomes a costly repair.

Prevent downtime

Better planning

The data helps forecast when to service or replace equipment.

Smarter planning

One view

Every connected machine appears on a single dashboard.

Full visibility

Key Benefits Achieved

Business results, before and after

MetricBeforeAfterChange
Cloud costs Baseline About 55% lower (expected) Cost down
Systems to manage 3 Azure tools, with data copied between them 1 platform Simpler
Reliability Broke often Stable and supported More reliable
Data science work Wait for data from another system Work directly on the platform Faster

What They Can Do Now

More machines, more users, more ideas

Since go-live, Donaldson has connected more machines and brought more users onto the platform.

On the horizon

A smarter support team

When a customer reports a machine problem, support reps can spend a long time finding the cause. Donaldson is exploring an AI assistant, built with Databricks Genie, that guides them step by step to the answer.

Being explored. Not yet built.
Next

Shared data

Sharing data with partners and dealers from the same platform.

Ready to turn your data into revenue?

SunnyData helps you turn live product data into a service customers pay for.

Portrait of Harrison Johnson
Harrison Johnson Account Executive, Retail & Manufacturing
Is your data platform holding back your next big idea?