Manufacturing Execution System

For modular furniture factory workflows

Product:New product from scratch
Team:Lead Designer, PM, EM, CTO, 1 Dev
Role:Research, design & testing
Time:3 Months to first release
Infurnia MES Portal — Modular Unit list and 3D panel drawing view MES product preview

Simplified illustration of the cabinetry workflow (AI-generated)  ·  View product

Business impact

Comparing data from the 1st week to 6 months

Revenue Indicators chart — Revenue in Rs, Panel added for Manufacturing, Unit Enabled, Work Order Created
Revenue:
$0.162M to $3.82M
+2258%
Work orders created:
633 to 5650
+793%
SKUs processed:
5.9K to 152K
+2476%

Source: Infurnia Mixpanel . *Revenue shown only for process fee of panels

Adoption & Net promoter score

Ratio of panels (SKUs) added to total users (in 6 months)

All panels added / all user profile ratio chart

NPS test conducted with users

Production designers 60 NPS Score: 61
-10003070100
Operators / managers 57 NPS Score: 58
-10003070100

At a glance

Why

  • Bring the exact kitchen design into the customer's home
  • Faster and error-free manufacturing

What

  • Digital aid for furniture factory workflow
  • Similar to PLM products for cabinet units

Whom

  • Cabinet production designer
  • Different types of operators
  • Production managers / Supervisors / Owners

How

  • A whole new product for factories
  • Various touchpoints (tracking, scheduling, etc.)
  • Paperless process, portable solutions

Product brief

Modular cabinetry factories ran production on spreadsheets, WhatsApp, and paper, with no link between the interior CAD design and the factory floor. Disconnected workflows delayed deliveries, which gave Infurnia a big opening.

A 0→1 Manufacturing Execution System that covers the whole process, from interior CAD design to furniture ready for delivery. It handles digitisation, communication and tracking across the cabinetry factory.

Animation: interior CAD design flowing through the factory to delivery

Product glimpse

Why a design platform built a factory system

1. Business goals: quick revenue

Unlocking rapid revenue growth with minimal engineering overhead.

2. Easy to acquire new user base

Modular furniture factories are already our clients’ customers, so they're easy to reach.

3. Manufacturing outputs were already built

Our design app already generates the 3 main outputs manufacturing needs.

We already have a design portal for designing modular units.

DesignX portal screenshot — 3D cabinet view with Option used by production designer / Output needed for Factories callouts

4. Addressing different personas

Needs of Interior designer and production designer separately.

Interior designer: focus on look & feel of cabinet

Kitchen interior render photo
  • Interior design (appearance and size)
  • Cabinet uses (no. of drawers, pull-outs, etc.)
  • To sell cabinet design for people to buy
  • Unconcerned with factory manufacturing details

Production designer: focus on manufacturability

Technical cabinet construction drawing
  • Cabinet-oriented construction
  • Deals with carpenters & shop floor
  • Focuses on hardware specs, drilling patterns, & CNC machining
  • Materials procurement & hardware lists

Until MES, both used the same design portal.

5. Completing full suite in interiors

Easy for a business to buy a full suite that takes care of design to manufacturing

6. Faster time to market

Other architecture verticals take years to build. MES shipped in months.

Inside the factory floor

Who uses this system

Different types of users in the factory 5-circle spectrum diagram: Architect designer to Interior designer to Production designers to Production Managers to Operators, with Factory and In focus range labels

Current workflow in the factory

Overall workflow

Manufacturing execution system workflow illustration

Interior designer hands over the design file to production designers

Production designers verify and correct all units in the file, then give production sign-off

Hands over manufacturing files to supervisor who oversees production

Different Machine Operators get the files & perform their respective actions, tracking etc.

Ecosystem map

Full ecosystem flow diagram: Kitchen designer to Design File to Production designer's end (Cabinet editing/list/Panel listing) to Factory operator's end (Board layout/Cutlist/Manual+Automated cutting/Edge banding/CNC machining/Inspection/Packaging/Shipment) to Production manager/owner (Inventory management/Scheduling/Progress tracking/Reporting) to Assembly at client's end

Detailed flow

Detailed flow in factory diagram

Three solutions, one production floor

1 — Production Designer

A flat, searchable list of units. Designers can edit a unit quickly, open it in detail, or bulk-edit with templates; verify, delete and restore units in bulk; and pull in design changes before giving sign-off.

2 — Production Manager

Real-time status of every work order, down to the panel. Managers create stations, map operators to them, schedule work, and course-correct as issues come up.

3 — Board Cutting Operator

A daily work order with sheet layout files, assisted sawing for faster, guided cuts, custom board settings, and auto-printed labels per panel.

Keep scrolling for the solution. My process is further down, under "How I got there".

Sanjay, Production Designer

Solution 1: for The Production Designer

Name: Sanjay · Age: 35 · Qualification: Diploma in product design

Pain points

  • Units are scattered across the design file, so they're hard to check one by one.
  • Remembering verified units
  • Editing requires a lot of clicks & juggling in design app
  • Handling frequent interior design changes

Needs

  • Creating & managing multiple work orders
  • Verify, edit & give sign off to production
  • Easy coordination with both manufacturing & interior designer.

Qualities

  • Keen on details
  • Tech savvy
  • Understands manufacturing constraints
  • Collaborative

User flow

Creates work order (term used for a file in factories)
→
Verifies cabinets
Edits cabinet's properties
→
Verifies output files
→
Gives production sign off

Work order creation flow

Production designer

Needs solved

1. Converting design to factory work file

2. Creating & managing active work orders

Basic actions inside a work order

Production designer

Needs solved

A flat list of units means far less navigating than the design app.

See all essential details of units upfront.

Search, filter & sort units

Multiple iterations, incorporating user feedback

Iteration 1 — larger 3D viewport layout where panel details and 3D cannot be viewed together

Different ways of editing units

Production designer

Needs solved

Three ways of editing units and panels

Quick edit for short edits.

Detailed edit in design interface

Bulk edits using templates

Verifying and removing units

Production designer

Needs solved

Delete units which are not required

Restore deleted units

Verify units one by one or in bulk

Design changes and production sign-off

Production designer

Needs solved

Resetting the wrong changes

Import new changes made by interior designers

Giving production sign-off by publishing

Ramesh, Production Manager

Solution 2: for The Production Manager

Name: Ramesh · Age: 40 · Background: Business owner
See SolutionHide Solution

Pain points

  • No big picture / real-time feedback of factory operations
  • Giving frequent inputs to the operators
  • Errors going unnoticed

Needs

  • Track each order's progress through to completion
  • Track each operator's efficiency
  • Managing rejections & breakdowns in factory
  • Plan and schedule work to operators

Qualities

  • Leadership / team player
  • Business oriented
  • Handles sheer volume

User flow

Published work orders
→
Schedules / plans work order for execution
↑
Creates stations / manages operators
→
Tracks work order execution
→
Course correction in factory

Tracking work order executions

Production manager

Needs solved

Seeing the overall status of every work order

Seeing details down to each panel level.

Managing stations & operators

Production manager

Needs solved

Creating new stations and mapping operator to stations

Planning execution by scheduling work to the operators

Ajay, Board Cutting Operator

Solution 3: for The Board Cutting Operator

Name: Ajay · Age: 40 · Qualification: Skilled Technician
See SolutionHide Solution

Pain points

  • Hard to handle paper documents in factory
  • Time and effort to calculate dimensions & cutting order
  • Getting the right label to stick after cutting the panel

Needs

  • Knowing which work order to execute on a given day
  • Way to understand and cut panels faster
  • Option to customise the board settings to suit their needs.

Qualities

  • Not tech savvy
  • Years of muscle memory
  • Handles sheer volume
  • Own style of working

User flow

Open assigned work order of the day
→
Sheet layout files
Gets plywood sheets from the factory
→
Follow the sheet layout manually, or use assisted sawing
→
Cut the panels
→
Prints the panel label and sticks it on the cut panel

Accessing board layout for cutting

Machine Operators

Needs solved

Seeing boards and changing board setting if needed

Assisted sawing for easier cutting

Machine Operators

Needs solved

Assisted sawing: a mode that lets operators cut faster

Printing labels and sticking on panels

Designing for everyone

MES is a B2B tool used by sighted professionals, so we didn't aim for full screen-reader coverage. We did design for temporary and situational limits such as glare, one-handed use and low vision.

Adhering to WCAG 2.1 AA minimum requirements

Alt text for SKU images

All SKU image names are coded in alt text
alt="<Type>-<sku name>"

All images in the platform come from SKU display images, so the same naming is used in the alt text.

Code editor showing alt text implementation + Modular Units screen

How it looks for low vision & color blind users

Information uses icons and text as well as colour, so the app still works in greyscale.

Grayscale & color-blind simulation screenshots of Security Central / MES screens

Keyboard navigation

All options in the interface are accessible by keyboard, with an intuitive navigation order that addresses the important element first in each interface; focus state is clear and recognisable.

Work order list with numbered tab-focus orderBoard optimiser panel-map with numbered tab-focus order + "Focus state" callout

Text readability & contrast ratios

Body text: 14px #202126   Used 95%
Small text: 12px #323338   Used 5%

Default text & Background contrast. Used 95%

10.31:1

✅ AA (Normal)  ✅ AA (Large)
✅ AAA (Normal)  ✅ AAA (Large)

Least contrast of text & background in interface. Used 5%

5.07:1

✅ AA (Normal)  ✅ AA (Large)
❌ AAA (Normal)  ✅ AAA (Large)

Other accessibility elements

  • ✅ Predictable & consistent navigation, similar to Design App
  • ✅ Progressive disclosure unfolds elements on selection
  • ✅ Using plain language free from technical jargon
  • ✅ Link purpose is clear from the link text itself
  • ✅ Using icons along with color to communicate status in alerts
  • ✅ Tooltips accessible (dismissible, hover-able, persistent)
  • ✅ All buttons are large; the smallest is 32px
  • ❌ No captions or transcripts, since there are no video elements
  • ❌ No voice-only interface, since the user base is narrowly defined

How I got there

  • Getting into the factory — Factory visits, interviews & affinity mapping with real users.
  • Learning from what exists — Competitive teardown of Naadi, Ardis & 3 other MES tools.
  • Meet the users — Personas, journey maps & need prioritisation.
  • Shaping the solution — Wireframes, IA, and iteration with stakeholders.
  • Bringing it to life — Visual design system & prototyping.
  • Working with the team — Dev handoff & cross-functional collaboration.
  • The hardest problem — A tough tradeoff with engineering, and how I resolved it.

Putting it to the test

Task success rate and path analysis:

  • All events were logged in Mixpanel
  • Checking the task flow and drop-off rates using Mixpanel's user flow analysis feature
  • Some defined task flow funnels are:
    • Work order creation flow
    • Publish flow
    • Bulk edit SKU structure
    • Downloading CNC and board layout
Mixpanel user-flow Sankey diagram: Work Order Opened, Modular Units Tab, Cutlist/Board layout/Panel view events

Think-aloud & retrospective testing

Users were given a set of tasks while we observed their workflow. We probed frequently to understand their feedback.

Testing performed with:

5 customer success executives 5 users in the factory

Given JTBD:

  • A. Create work order from design file.
  • B. Make 5 units of a design ready for production (includes editing, verifying, etc.)
  • C. Production manager checking work order progress in detail
  • D. Saw operator using assisted sawing to cut boards

Some changes made based on user feedback

Flow: A. Create work order from design

User Feedback:

Entering the design link was hard and needed a lot of back-and-forth with the interior designer.

MES case study research photo 28b

Changes made:

A new flow where the interior designer publishes the design, and it then shows up in the production designer's selector.

MES case study research photo 28c
View more feedback & design changes

Flow: B. Make 5 units of a design ready for production, (editing / delete / reset / verify units)

User Feedback:

In the first iteration users could see only 3D or panel info, and could not relate the units and panel information without 3D.

MES case study research photo 29a

Changes made:

Changed the layout to access all required information upfront, and have access to the 3D viewport all the time; the user can access the 2D panel drawing and 3D at any time, with 3D as the default option. There was a tech limitation which restricted showing the panel 2D view & 3D viewport at the same time, so it was moved into tabs.

MES case study research photo 29c

User Feedback:

The home menu and work order menu looked similar and confusing, and users felt the work order contents were cramped.

MES case study research photo 29d

Changes made:

We changed the work order layout to match the design app, with tabs, giving the content more room.

MES case study research photo 29e

User Feedback:

All personas were accessing work orders at the same level. There was no access control, so everyone saw options they didn't need.

MES case study research photo 29f

Changes made:

We made persona-specific flows using role restrictions, so each user sees only the content they need.

MES case study research photo 29g

Flow: C. Production manager checking work order progress in detail

User Feedback:

After spotting a rejected panel, the user needed its details to take action, and the page didn't show them.

MES case study research photo 30a

Changes made:

Full panel details now open on the same order progress page, so the production manager can act right away.

MES case study research photo 30b

Flow: D. Use assisted sawing to cut boards

User Feedback:

Sheets were displayed too small to see in detail, and users also wanted to see cut progress and a sheet preview.

MES case study research photo 30c

Changes made:

Allocated more space for the sheets in the new interface and added zoom and pan in the sheet canvas. Sheet progress and previews were also added.

MES case study research photo 30d

User Feedback:

In assisted sawing, some operators are used to their own flow: they want to select the panel they're cutting rather than have the system dictate it strictly.

MES case study research photo 30e

Changes made:

A semi-assisted sawing mode was built, where the user can select a panel they want to cut, while the system also suggests which one to cut next if the user doesn't give any input. The progress is also shown in the interface.

MES case study research photo 30f

SUS evaluation

10
Avg SUS score: 75.25
SUS survey response table: 10 respondents x 10 questions, Raw SUS Score, Sus Final score

Net Promoter Score

For Production Designers

60 When:After the org’s 1st renewal

“How likely are you to recommend MES product to a friend or colleague?”

Method:Survey form
Task:Creation to publishing work order
Promoters: 42 Passives: 13 Detractors: 5
NPS Score61
-10003070100

For Operators & supervisors

57 When:After the org’s 1st renewal

“How likely are you to recommend MES tracking, pack-n-ship and outputs to a friend or colleague?”

Method:Via Renewal feedback calls
Task:For sawing / tracking / packaging & shipping, etc.
Promoters: 39 Passives: 12 Detractors: 6
NPS Score58
-10003070100

Did it work?

What success looks like

No. of work orders created from Design

Why: Proportional to adoption of the MES system

No. of panels / units added to production

Why: Proportional to revenue, since we charge per panel

No. of outputs downloaded

Why: Indicates usage of the system

No. of panels tracked / scanned

Why: Indicates that operators are working on the panels

Data events are logged in Mixpanel for tracking

Growth over time in the first 6 months

Increase in “panels added” in past 6 months

2251% increase from start From 40.6K to 954.4K

MES case study metrics chart 35a

Ratio of “panels added” to total no. of users.

0 to 20.33*All design clients were also recorded as users

MES case study metrics chart 35b

Data of all outputs downloaded

(Board download clicked and CNC files downloaded)

MES case study metrics chart 35c

Usage of tracking feature

(panel cutting, CNC, Edge band packaging)

MES case study metrics chart 35d

The business case (revenue metrics, first 6 months)

Data compared from the first week to 6 months after release.
Revenue shown below is only for processing fee of the panels; the MES add-on fee isn't included.

Revenue Indicators chart, Jul 22 to Dec 30 2024: Revenue in Rs, Panel added for Manufacturing, Unit Enabled, Work Order Created

Source: Data from Infurnia Mixpanel.com *Revenue shown only for process fee of panels

Revenue:
$0.162M to $3.82M
+2258%
Panels added:
40.6K to 954.4K
+2251%
Work orders created:
633 to 5650
+793%
SKUs processed
5.9K to 152K
+2476%

What I learned

Pilot study helped refine research direction

Running a pilot study before the real interviews tested my assumptions and made sure I asked the right questions.

Finding edge cases early minimises redesigns

Working through user scenarios before wireframing meant fewer rounds of iteration later.

Readymade design systems aren't always a perfect fit

A ready-made design system sped things up, but I still had to fix contrast ratios in some components and adapt it to our users.

Knowing Implementation feasibility accelerates execution

Knowing what was technically feasible made working with developers faster.

Versioning designs for smooth implementation

Since the designs were implemented in phases, maintaining clear versions of design files helped developers track changes & avoid misalignment.

Thank You

Present