Research and analysis, behind the service blueprint

01 · CONTEXT

The operation we walked into

Scale, sprawl, and a single flat map.

180m+item records tracked in the product master

100k+retail and reseller partnerships feeding data

60+countries under active coverage

300+product categories maintained

Weeklydelivery cadence to the client platform

PROVENANCE Public published references Archive project records and survey data Derived reconstructed from evidence

02

The flat map, excerpt

A slice of the legacy process map teams navigated the platform with. Function focused, running versus supporting, no ownership, no dependencies. The full map is in the viewer above.

Archive

Data coding

Auto matching

Manual matching

Object creation

Data signoff

Matching QC

Item QC

Codeplan management

Codeplan review

Ad hoc clarification

Brand management

WFM set up, Data in, Data enrichment, Data out …

Now covered by blueprint phases Item matching, Item creation and Coding sign off

03

Documentation audit

Every artefact teams navigated the platform with, inventoried at kickoff. One was in company-wide active use.

Derived

Artefact	Owner	Last updated	Coverage	Status
Flat process map, WFM set up to Data out	Ops enablement	4 yrs prior	End to end, shallow	Active
Separation rule notes	Global operations	2 yrs prior	Per market, uneven	Conflicting
Codeplan wiki	Service centres	1 yr prior	Fragmentary	Orphaned
Coding instruction library	Enablement	3 yrs prior	Two phases	Stale
Feature dictionary	Reference data	Current	Attributes only	Active
Brand dictionary	Reference data	Current	Brands only	Active
Partner register	Commercial ops	2 yrs prior	Partial	Stale
Projection methodology memo	Data science	3 yrs prior	Handover only	Orphaned
Capacity plan spreadsheet	Workflow management	1 yr prior	Non canonical	Duplicated
Delivery calendar	Delivery	Current	Timeline only	Active
Escalation contact list	Support	Current	People only	Active
Store check programme notes	Enrichment teams	2 yrs prior	One market	Stale

1 of 12 artefacts in company-wide active use, the flat map. No artefact carried ownership, dependencies or rules.

04

Topology audit

Systems, files and references behind each lifecycle phase. No single system saw the whole flow. This inventory became the blueprint's topology layer.

PublicArchive

Separation runPartner feeds gatewaySeparation rulesCoverage monitorRaw archive

Workflow managementWorkflow systemOrder registerDay planInstruction library

Item matchingMatching engineProduct masterMatch workbench

Item creationSpecification captureEvidence archiveStore check log

Coding sign offCodeplanFeature dictionaryBrand dictionarySign off pack

HandoverProjection engineData warehouseRelease fact tablesDelivery rendererClient format libraryValidation report

05

Process baseline

Where time went before the blueprint. Reconstructed from delivery logs and workshop evidence.

Derived

Phase	Time in phase	Handovers	Rework rate
Separation run	1–2 d	2	4%
Workflow management	1–2 d	2	3%
Item matching	3–6 d	4	18%
Item creation	4–7 d	3	15%
Coding sign off	3–5 d	3	12%
Handover	2–3 d	2	5%
Roadmap sign off	21 d	6	—

06

Research brief

Why research came before blueprinting.

Archive

Goal: Replace opinion with evidence before drawing a single blueprint line. Quantify where the platform actually stalls and who is affected.
Scope: All six lifecycle phases and seventeen steps, all four involvement tiers, seven regions.
Questions: Where does work stall and why. Who owns each decision today, and who believes they do. What would a trusted reference need to carry to displace the flat map.
Observed symptoms: Unclear ownership. Fragmented workflows. Planning slow because nobody could point at where work stalled.

02 · FIELD EVIDENCE

Listen at scale

Nine months of structured fieldwork across seven regions.

Stakeholder interviews

42

Semi structured, all four tiers, seven regions. Waves 1–3.

Consumer voice survey

118

Responses analysed and recoded. Fielded February 2022, 63% response rate.

Focus groups

9

Sessions in service centre and local market hubs, mixed tier rooms.

Importance analytics

1,412

Statements coded and weighted by severity and reach.

08

Fieldwork timeline

The research operation across the nine month engagement. Waves overlap the blueprint build so findings fed each layer as it formed.

ArchiveDerived

nM1M2M3M4M5M6M7M8M9

Consumer voice survey, analysis118

Interviews, service centres and cross domain26

Shadowing in service centres5

Focus groups, mixed tier9

Interviews, SME local experts10

Pulse survey, service centres32

Interviews, value chain6

Mirror group validation rounds4

Regional sign off reviews7

09

Field observation log

Shadowing in the service centres, watching the work instead of asking about it.

Derived

Centre	Role shadowed	Focus	Findings
Pune	Service centre team member	Matching exceptions	11
Vadodara	Service centre team lead	Routing and approvals	8
Chennai	Service centre domain lead	Codeplan clarifications	9
Warsaw	Service centre team member	Separation variance	7
Manila	Service centre team member	Coverage checks	6

10

Coding codebook

The machinery between 1,412 raw statements and three themes.

Derived

Code	Freq	Theme
approval-wait	142	Ownership & Accountability
unclear-owner	128	Ownership & Accountability
dependency-block	111	Ownership & Accountability
missing-docs	124	Tools & Documentation
policy-ambiguity	118	Process & Standards
rework-loop	102	Process & Standards
tool-switching	98	Tools & Documentation
training-gap	96	Long tail
standard-drift	91	Process & Standards
data-trace	89	Tools & Documentation
comms-lag	84	Long tail
other, 4 codes	229	Long tail

“We need clear visibility into bottlenecks, where data gets stuck, who needs to approve, and how to unblock.”

Engineering · biggest driver of delays

Ownership & Accountability

“We don’t understand how dependencies and governance affect accuracy, costs, and timelines.”

Global Ops · policy ambiguity

Process & Standards

“Give us clarity on data flow, what’s essential versus what isn’t.”

Data Science · downstream impact

Tools & Documentation

03 · PRIORITISATION

Quantify the noise

From complaint volume to a defensible priority order.

12

Theme heatmap

1,412 coded statements clustered by keyword frequency, weighted by severity and reach. Three themes carried 71% of the signal and became the spine of the blueprint.

Archive

27%

Ownership & Accountability

381 statements · biggest driver of delays

dependenciesdecision rightsapproval waitsSLAs

22%

Process & Standards

311 statements · governance and rework

governancepolicy ambiguityrework

22%

Tools & Documentation

311 statements · accuracy and visibility

accuracydata flowdownstream impact

7%

Training & onboarding

6%

Comms cadence

6%

Loading times

5%

Access & permissions

5%

Other signals

13

Keyword frequency

Top terms across the coded corpus, mapped to the theme they cluster under.

Derived

Term	Count	Theme
approval	142	O&A
owner	128	O&A
documentation	124	T&D
policy	118	P&S
dependency	111	O&A
blocked	104	O&A
rework	102	P&S
tooling	98	T&D
training	96	Long tail
standard	91	P&S
trace	89	T&D
handover	83	O&A
escalations	78	O&A
accuracy	74	T&D
volume	62	P&S
coverage	57	P&S

14

Severity × reach backlog

Issues scored by severity and reach, with research confidence. Score set the order the blueprint addressed them in.

Derived

ID	Issue	Score
R-01	Approval bottleneck at stage gates	25
R-02	No owner for matching exceptions	20
R-03	Duplicate handling interpreted locally	16
R-04	Rework loop, sign off to creation	16
R-05	Partner feed specs undocumented	12
R-06	Tool switching in coding sign off	12
R-07	No trace from evidence to decision	12
R-08	Coverage threshold invisible	9
R-09	Training gap in service centres	9
R-10	Access delays at onboarding	6
R-11	Comms lag across time zones	6
R-12	Loading times at peak volume	4

04 · PEOPLE & SERVICE

Map the people and the service

Who decides, who delivers, where the rules bind.

15

People matrix

150+ stakeholders in four involvement tiers, mapped by location and decision group. The four groups became the blueprint's decision hierarchy.

Archive

Tier	People	Where	Group
Specialists	60	Service centres and global ops · Pune, Warsaw, Manila, Chicago	Delivery
Cross domain	50	Centre and global leads · Vadodara, Chennai, Nuremberg	Mirror
SME	30	Local experts and data science · Madrid, Milan, Berlin, Amsterdam, Geneva	Contributors
Value chain	15	Local management and sales · London, Paris	Informed

16

Stakeholder field

Every mapped stakeholder, coloured by tier.

Archive

Specialists 60Cross domain 50SME 30Value chain 15

Service centre team lead

Country group lead · Vadodara

Cross domainMirror

Service centre team member

Client focus · Pune

SpecialistDelivery

Local expert, coding

Market knowledge · Madrid

SMEContributor

Global operations manager

End to end view · Nuremberg

Cross domainMirror

Data science

Methodology · Geneva

SMEContributor

18

RACI matrix

Decision rights for the activities research flagged as fractured. Agreed in workshop, carried into the blueprint's decision hierarchy and its Actors and personas layer.

Derived

Activity	Delivery	Mirror	Contributors	Informed
Define stage requirements	R	A	C	I
Approve lifecycle objectives	C	A	C	I
Own constraint register	R	C	A	I
Resolve matching exceptions	R	A	C	—
Approve codeplan changes	C	A	R	I
Validate segmentation	C	R	A	I
Sign off blueprint layers	C	A	R	I
Escalate risks	R	A	C	I
Set SLA and SLO targets	C	A	R	I
Publish reference updates	R	C	A	I

19

Segmentation criteria

Needs based criteria behind the user landscape, each validated before use.

Archive

Criterion	Validated by	Drives
Decision authority	Interview follow up	Group assignment
Data touchpoints per phase	Workshop card sort	Layer entry point
Phase coverage	Survey cross check	Tier placement
Escalation role	Shadowing evidence	Risk layer routing
Downstream application need	Internal data user interviews	Internal data users layer

20

Service mapping workshops

Pipeline structure and connected lifecycle objectives, built with the teams who run each phase.

Archive

Frontstage

Retailer feed lands

→

Data order raised

→

Release published

Backstage

Separation

→

Matching

→

Creation

→

Sign off

→

Handover

Supporting

Codeplan management

→

Reference data

→

Capacity and SLAs

21

Connected lifecycle objectives

One objective per phase, agreed in workshop. These became the blueprint's phase headers.

Archive

Separation runReceive retailer data and split it into processable units by market and category

Workflow managementRegister, route and own each run through the production workflow

Item matchingMatch incoming items to the product master, resolving what the algorithm cannot

Item creationCreate new product records with validated specifications

Coding sign offCode every item to the taxonomy and approve it for release

HandoverProject, package and release the data to internal users and the client platform

22

Constraint tokens

Every binding rule sized and scored, so a stall could be scheduled, not just described. Rating is impact weighted by sentiment. Tokens feed the blueprint's constraints layer directly.

Archive

Efficiency constraint

Large variable data volume

Weekly partner volumes vary by market and season, so runs queue behind capacity instead of value.

volume

Consumer voice survey, February 2022

Partner confidentiality

Oversight of large retailers

Inter-country teams share responsibility for large retailers. Missing information sources for novel data types drive escalations between centres and countries.

escalations

Handling speed30

Data relevancy13

Compliance13

Negative sentiment13

Variable retailer data10

Loading times7

Consumer voice survey, February 2022

Coverage threshold

Coverage below release threshold

Runs are held when partner coverage sits under the 80 percent release threshold, but the threshold is invisible to delivery teams until the hold lands.

blocked

Representative weighting pending workshop data

Token	Category	Keyword	Rating
Match confidence threshold	Match threshold	accuracy	80
Unverified specifications	Verification	accuracy	79
Evidence for manual matches	Evidence	stuck	76
Duplicate handling interpreted locally	Policy	ambiguous	72
Set global process requirements	Efficiency constraint	onboarding	68

23

Risk register & SLA baseline

Hazards with mitigation and escalation, and service levels set against the process baseline. Fed the blueprint's risks layer and its metrics and targets.

ArchiveDerived

Hazard	Mitigation	Escalation
Late partner feeds	Automated feed monitoring	Partner manager
Coverage below threshold	Run held, threshold shown in context	Operations lead
Capacity shortfall	Work rebalanced across centres	Delivery manager
Unowned stage	Owner assigned same day	Centre team lead
Codeplan drift by market	Single codeplan, reviewed	Global operations

Service level	Baseline	Target
Feed acknowledgement	2 d	Same day
Coverage before release	Unstated	80%
Exception resolution	8 d	3 d
New item lead time	10 d	5 d
Roadmap sign off	21 d	≤7 d

05 · SYNTHESIS

From evidence to strategy

Three responses, one blueprint.

24

Traceability

Every strategic response traces back to evidence and forward to a blueprint layer. The funnel, auditable at a glance.

Archive

Evidence

R-01 · approval bottleneck27% theme weight, fieldwork interviews

R-02 · unowned exceptionsShadowing in Pune, RACI workshop

Constraint tokensConsumer voice survey, ratings 68–96

Process baseline + SLAsDelivery logs, workshop reconstruction

Topology + documentation auditSystems and artefact inventories

→

Strategic response

Full data lifecycle clarityService mapping, SLA and SLO design

Clear standards for faster deliveryConstraint analysis, constraint tokens

Trusted source for decisionsAcceptance criteria, segmentation

→

Blueprint layer

Phased lifecycle · Timeline

Actors and personas · Frontstage · Backstage

Constraints · Supporting processes

Topology · Evidence and artefacts

Risks · Opportunities · Metrics and targets · Internal data users

Response 01

Full data lifecycle clarity

Unclear roles, stages and SLAs caused delays. Mapped the end to end lifecycle to remove silos, improving planning and accountability.

Service mappingSLA/SLO design

Single source of truth adopted by 150+ stakeholders

Response 02

Clear standards for faster delivery

Bottlenecks and inconsistent processes slowed delivery. Set accessible standards to improve compliance, cut errors, and speed execution.

Constraint analysisConstraint tokens

Reduced rework, clarified ownership of every rule

Response 03

Trusted source for decisions

Poor documentation and low risk visibility led to defects. Created a company wide single source of truth to boost alignment.

Acceptance criteriaSegmentation

Faster, clearer, more confident decisions

26

Playback & validation log

The blueprint earned adoption through review rounds, not a launch email.

Derived

Round	Audience	Changes	Layers touched
R1 walkthrough	Mirror group	14	Phased lifecycle · Timeline
R2 centre review	Service centres	9	Frontstage · Backstage
R3 leadership	Value chain	5	Opportunities · Metrics and targets
R4 regional sign off	All seven regions	0 open	All layers

27

Measurement plan

Baseline, target and result for every claim the case study makes.

Archive

Metric	Baseline	Target	Result
Roadmap sign off	21 d	≤7 d	3 d
Adoption	1 flat map	100	150+ · 7 regions
Post adoption NPS	—	≥60	92
Rework rate, item matching	18%	<10%	Reduced
Coverage before release	Unstated	80%	Visible per run

Roadmap sign off21 d

3 d

Delivered over nine months. These three responses became the blueprint. Six phases, seventeen steps and a full stack of layers, from timeline and constraints to risks, evidence and metrics, each one traceable back to this canvas.

→ Explore the service blueprint in the next viewer