All work
Professional

2023 · Product journey mapping in fashion e-commerce · 4MIN READ

E2E Return Optimization

Product journey board, 2023

TeamSystemCustomer facingBreaking point

1. Create

Season briefing to supplier handover

Design

Product approval

Tech / fit

Master and block fits

1

Merchandise

Orders and pricing

PDM / PLM

Specs to supplier

2. Store

Where product data lives

SAP

Retail and wholesale master data

Data lake

Pricing, orders, status

2

PIM

Central product info

Built end-to-end

3. Shoot

Eight-step studio and retouch chain

Samples

Stockroom, express delivery

Shooting

Stills and model

Approval 1

Selection

External retouch

12h turnaround

3

Approval 2

3h on declines

Colour correction

Style on hand

4

Built end-to-end

4. Publish

Assembling the page

DAM

Approved imagery

Copy & translation

48h to publish

5

Storefront feed

Data to the shop

5. Sell & learn

What the shopper sees, and says back

Product page

Imagery, copy, size chart

Fit guidance

Third-party tool

6

Returns & feedback

Weekly returns list

7

Breaking points

  1. 1Fit decided here, never written back into the page.
  2. 2Data sits in four places, so nobody trusts one version.
  3. 3Retouching corrects colour away from the real garment.
  4. 4Colour check happens once, with no reference locked earlier.
  5. 5Copy is written from mood, not from garment behaviour.
  6. 6Fit tool has no access to the updated size tables.
  7. 7Return reasons too coarse to send back to any owner.
Redrawn from the original FigJam board, left to right as it was built. Named owners and internal identifiers are replaced by roles and system types - the point of the exercise was that nobody could see the whole chain at once.

Efficiency

  • Missing product info and incomplete fit and size description lists
  • Fitting check (true to size, normal fit) missing on some pages, no automation behind it
  • Outdated size charts
  • Inaccurate colours and imagery against the real garment
  • Inconsistent or missing copy highlights, prints in particular

Technical

  • Model size missing because customer feedback never came back into the record
  • Data not released into the downstream systems
  • Model information missing or broken

Strategy & visibility

  • Owners and stakeholders unknown across the chain
  • Loss of manpower turning directly into loss of knowledge
  • Undocumented processes and no shared visibility
  • Decentralised data across PDM, SAP, PIM and Salesforce
  • Inconsistent size labelling per category
  • Purchasing strategy: not every size is bought for e-commerce

Development status at handover

25% Done50% In progress25% Not started

High priority

  • Copy guideline for product descriptions
  • Mapping the end-to-end product journey
  • Size table update, inch to cm
  • Fit Analytics visibility updates
  • Adding model information
  • Measurement info bugs

Medium priority

  • Repositioning the product info box
  • New visuals for “true to size”
  • Transfer of the new range plan
  • Product tagging app surfaced in Power BI via the data lake
  • New product strategy as a chance to unify size strategy
  • Fixing translations

Low priority

  • Evaluating a new fit tool
  • 3D images for flats
  • Video introduction for selected categories
  • Process input from product and styling

Process summary

How the work was made

Click a step to jump to the artefact it produced.

01

Mapping the chain

I contacted the teams along the path - design, tech, merchandise, photostudio, copy and translation, PIM and storefront - and drew the product journey end to end in FigJam: every hand-off, every system, every service-level agreement, and the owner of each node.

02

Finding where the promise breaks

With the map in place, the defects grouped themselves into three families: efficiency, technical and strategic. It became visible that most customer-facing errors were downstream symptoms of upstream data gaps, not mistakes made on the page.

03

Turning findings into a backlog

Each problem was paired with a solution and sorted by effort and impact - low-hanging fruit against high-hanging - then re-cut into a high, medium and low priority list the team could actually work through.

04

Shipping and testing

The first fixes went live: a copy guideline, the journey map itself as a shared reference, and the size table converted from inches to centimetres. One change went into an A/B test before the company entered insolvency and the programme stopped.

Participants

Eight teams and six systems mapped; stakeholders interviewed across design, tech, merchandise, photostudio, copy, translation, PIM and storefront.

Result

Three fixes shipped and one A/B test started against a one percent return-rate target worth thousands of euros a season; the programme ended with the insolvency.

What I would test next

Whether pairing fit guidance with model measurements on the product page moves the return rate on the categories with the worst fit complaints.

Outcome

  • An end-to-end product journey map covering eight teams and six systems, adopted as the shared reference
  • Every known defect traced back to its cause and sorted into a prioritised backlog
  • Three items shipped: the copy guideline, the size table conversion in centimetres, and the journey map itself
  • One further change taken into an A/B test; the programme ended when the company entered insolvency

What I took from it

The target was a one percent reduction in returns - small on paper, thousands of euros in practice. I learned to argue for design work in the language of that number, and that the hardest part of a systems problem is not solving it but getting everyone to agree on what the system actually is.

FigJamFigmaPower BIMicrosoft 365
Service DesignE-commerceJourney MappingStakeholder ResearchPrioritisationA/B Testing

Next case study

Personal Finances