Skip to content
Unit 4 of 18
RFPLumen DiagnosticsOpen anytime - every module Task uses this brief

Module 2 — Business Architecture

Contents · 26

What you'll be able to do

  • Sort what a customer calls "drivers" into business drivers, requirements and quality attributes, and tag each driver with the business-case category it proves
  • Write qualitative goals and SMART objectives that trace to drivers, including a baseline-measurement objective when the customer has no data
  • Build a stakeholder register that goes beyond the roles an RFP names, place each stakeholder on Mendelow's power/interest grid and choose how to engage them
  • Draw up a RACI matrix for a whole engagement with exactly one Accountable and at least one Responsible per activity
  • Tell a business capability map and a value stream from a functional decomposition of the system, and trace each capability to an objective
  • Classify requirements by BABOK type, and seed a RAID log with constraints, assumptions and clarification questions from an RFP

Business architecture is "a blueprint of the enterprise that provides a common understanding of the organization and is used to align strategic objectives and tactical demands." (OMG Business Architecture Special Interest Group; adopted in the Business Architecture Guild's BIZBOK Guide.) In practice it connects what a business intends — its model and its strategy — to what it actually does day to day, so that a change on either side can be traced to the other.

Each section below defines its terms and illustrates them with a short example. The Task asks you to build the Lumen artefacts yourself; a complete worked answer follows the Task — attempt the Task before you read it.


Drivers, goals, objectives

Business driver — anything a business depends on to keep succeeding and growing: a resource it uses, a process it runs, or a condition it operates in. Some drivers are within the company's control, and it should push those as far as it can; others, such as the state of the economy or trade policy between countries, it can only watch and plan around. Name a driver with a noun.

Four categories — everything in a business case must relate directly to proving one or more:

  1. Increase or protect revenue
  2. Reduce cost or increase efficiency
  3. Reduce risk
  4. Compliance

Examples of drivers: speed of revenue recognition; cost of goods sold (COGS) and operating cost; business productivity (more goods from the same inputs); business effectiveness (product-to-market penetration); market competitiveness (share); new-business opportunity; risk exposure; accuracy of planning and forecasting.

How to find drivers — root-cause analysis over the financial statements. Start from the statements and ask "what drives this line item?" repeatedly:

What drives revenue? → volume of products sold × average price. What drives volume? → the number of products and the number of salespeople. What drives salespeople? → the number and sizes of stores. What drives the number of stores? → this is a core business driver — an operational and capital decision, with nothing preceding it.

Repeat for each line item on each of the three financial statements, then determine which drivers are most important to focus on: those that impact the most areas of the business and have the largest effect on results.

Drivers in the customer's words

RFPs rarely separate drivers from requirements. A list headed "drivers" usually mixes three things:

  • Business drivers — conditions of the business itself. They would matter to the board even if no system were ever built: the cost of a process, a growth plan, a regulation.
  • Requirements and constraints — what the solution must do, or must connect to. They answer a driver; they are not one.
  • Quality attributes — how well the solution must do it: performance, availability, usability, modifiability. Module 3 makes these measurable.

Test each item with one question: would this matter if no software existed? If yes, it is a driver — name it with a noun. If no, find the driver it serves and record the item as a requirement or quality attribute traced to that driver. Two of Lumen's six aspects show the split:

Case aspectWhat it isWhy
Cost reductionDriver: cost of referral handlingA manual referral costs ~$9 of desk time whether or not a platform is bought
IntegrationConstraintIt names systems the solution must reuse and says nothing about the business; it serves the cost driver (no re-keying)

Then tag every driver with one of the four business-case categories. A driver you cannot tag is usually a requirement in disguise. Finally, look for the drivers the list leaves out: an RFP records what its author was thinking about, and regulation and continuity are often assumed rather than written down.

Goals vs objectives: goals are general, directional statements of desired achievement; objectives are the specific steps or actions taken to reach a goal. Goals can involve profitability, growth, customer service, and are expressed with qualitative words — "increase", "improve", "easier". Objectives are where the precision goes: each must be SMART — specific, measurable, achievable, relevant and time-bound. The goal "make reordering easier for repeat customers" is directional; the objective "raise the share of repeat orders placed without contacting support from 40% to 70% by the end of Q4, measured from the order log" is SMART.

Terminology varies: some companies use goals for the company as a whole, objectives for departments, and targets for individual employees.

Do not start out with what you have and then figure out what it can be used for. A far more successful approach is to start with what you want to accomplish, then determine what you have — and what you need — in order to accomplish it.

As an architect you must do the background research necessary to understand the goals; they are not always clearly explained in the case.

Rendering diagram…

The aim is a chain you can trace end to end — Drivers → Goals → Objectives → Capabilities — with every objective covered by some part of the solution, and every part of the solution traced to an objective. Two things make objectives hard on a real bid. First, an objective's metric needs a baseline, and customers often have none; when they don't, measuring it becomes the first objective. Second, the targets in a proposal are yours, not the customer's: label them as assumptions until the customer agrees them.


Business value, capabilities, value streams

  • Business value — the benefits a firm generates for its stakeholders, including long-term ability to create revenue, products, services, employment, quality of life and investment returns.

  • Business capability — encapsulates what a business is doing right now and what it needs to be doing to meet current and future challenges. Capabilities define what a business does rather than how (which processes describe), and are named with nouns. "Talent acquisition" is a capability necessary to a goal of having a competitive workforce; it says what is needed but leaves open whether there is an HR process from recruiting website to interviews to hiring administration, or whether everything is outsourced. Every capability decomposes into three overlapping components — People (human resources), Processes (business processes) and Technology (IT applications and technical resources) — so assessing or redesigning any one capability means examining all three layers, not just the process layer. Capabilities are the basic building blocks of a business, and the capability map in its entirety delivers a concise, non-redundant, business-centric view at its most basic level.

    Rendering diagram…
  • Value stream — an artefact allowing a business to specify the value proposition derived by an external (customer) or internal stakeholder. It depicts the stakeholder who triggers it, a sequence of value stream stages each creating a specific value item, and the value proposition delivered at the end — an end-to-end collection of value-adding activities creating an overall result for a customer, stakeholder or end-user. Name the stream from the stakeholder's side, verb plus noun ("Acquire product"), and give each stage a value item the stakeholder would recognise. Stages are not capabilities: each stage is enabled by one or more capabilities, which are mapped onto it. The value map shows how the organisation creates the value exchanged between itself and its stakeholders.

    Rendering diagram…

    A manufacturer's value stream: stages in sequence, each producing a value item, with capabilities (square boxes) mapped to the stages they enable. One capability can enable several stages — customer communication enables both the first and the last. The stages and capabilities vary by business; the pattern doesn't.

  • Value proposition — the promise that makes an offer worth choosing: which of the customer's problems it removes or which situation it makes better, what measurable benefit the customer gets, and why that beats every alternative on the table.

  • Strategy — the pattern or plan integrating an organisation's major goals, policies and action sequences into a cohesive whole (after J. B. Quinn, Strategies for Change, 1980). The balanced scorecard (Kaplan and Norton) is a strategy performance management tool — a semi-standard structured report supported by design methods and automation tools.

  • Information — the typical approach: extract information concepts from capabilities; identify capability-based relationships for information concepts; extract relationships from matching capabilities; define information concepts; identify their types; identify their use; validate the information map.

  • Organisation — the organisation map provides visibility on the business, describing business units and communication approach. An organisation typically has one or more leaders making strategic decisions. One way to describe it is via capabilities and their virtual relationships; another is an organisational chart describing structure, departments, groups and teams hierarchically.

The value of a capability map lies mostly in analysing current vs desired levels of capability, and in uncovering capabilities the organisation already possesses but does not recognise or manage explicitly. Capabilities and capability levels in a target business architecture give high-level direction for change — this is the core of capability-based planning.

Building a capability map (conventions after the BIZBOK Guide):

  • Name every capability with a noun, and let each appear once. Level 1 is the top tier — a few dozen capabilities for a whole enterprise, fewer for the scope of one solution; level 2 breaks each one down.
  • Keep it stable. A capability survives a reorganisation, an outsourcing deal or a new system; processes, org charts and applications don't.
  • Trace each capability in scope to an objective. Then rate each one today and at target: drawn as a colour-coded heat map, the biggest gaps show where the investment should go.

Capability map versus functional decomposition. A capability map describes the business; a functional decomposition describes the system — its functions, channels and adapters. The two are easy to confuse because a solution's functions often borrow capability names. The test: if the business went back to phone and paper tomorrow, a capability would still exist and a system function would not. "Order fulfilment" is a capability; "warehouse scanner app" is a function that supports it. Trace each function to the capability it supports, and each capability to an objective. A function that traces to nothing is scope you invented — or a sign you missed a driver.


Stakeholders

Formally (per the PMBOK Guide, 7th edition, PMI 2021), a stakeholder is any individual, group or organisation that may affect, be affected by, or perceive itself to be affected by a decision, activity or outcome of the project — a deliberately wide net that goes beyond just the people paying for or using the solution.

There are always several stakeholders and all are important, as they define official and often personal requirements and interests that must be taken into consideration. Understanding the organisation and business context helps identify vendors, partners, HR, clients and everyone else who needs to be on the architect's radar.

Identification techniques:

TechniqueUse
BrainstormingProduce the stakeholder list and identify roles and responsibilities
InterviewsInteract with specific stakeholders for more information about stakeholder groups
WorkshopsInteract with groups of stakeholders
Mind mappingIdentify potential stakeholders and understand the relationships between them
Organisational modellingDetermine whether listed units or people have unique needs and interests; models describe roles and functions and the ways stakeholders interact, helping identify who will be affected by a change
Process modellingCategorise stakeholders by the systems supporting their business processes
Stakeholder list, map or personasDepict the relationship of stakeholders to the solution and to one another
Survey or questionnaireIdentify shared characteristics of a stakeholder group

Definitions:

  • Stakeholder groups — depending on classification, a stakeholder may belong to different groups: organisation/organisation unit, or power/influence/interest.
  • Concerns — the issues, risks and constraints stakeholders have with the solution. This may include the use of the solution, perceptions of its value, and the impact it has on a stakeholder's ability to perform necessary functions.
  • Stakeholder requirements — describe the needs of stakeholders that must be met to achieve the business requirements; they may serve as a bridge between business and solution requirements.

The ISO 42010 distinction between a view and a viewpoint is worth carrying over from Module 1 here: a view is what you actually see, a viewpoint is the vantage point that determines what you see. An architecture viewpoint is a set of conventions — model kinds, notations and analytic techniques — for constructing and interpreting one type of view so that it addresses a specific set of stakeholder concerns (examples: operational, systems, technical, logical, deployment, process, information), while an architecture view expresses the system from the perspective of one or more stakeholders using those conventions.

Stakeholder register

A register has one row per stakeholder or stakeholder group, and five columns: Name, Role, Group, Concerns, and Artefacts that address their concerns. The last column is a promise: for each concern, the deliverable that will answer it — an architecture view, a prototype, a report specification, a plan. Architecture views appear in it, but it is not a list of 42010 views. A concern with no artefact is a gap in your plan.

NameRoleGroupConcernsArtefacts that address their concerns
Maya ChenPatientUserUsability on a phone, accurate appointment detailPatient-portal prototype, notification copy
Elena VossSponsor / COOCustomerCost, timeline, utilisation, regional expansionBusiness case, conceptual diagram, financial reports

Work through four groups: Users of the system, concerned with usability and the job they do with it; Customers — the sponsor who pays and is concerned with timeline and cost, the owners of the business rules and of the systems you integrate with (subject-matter experts, SMEs), and very often the CIO, enterprise architect, security, data protection and finance; the Project Team who design, implement and support the solution; and Regulator/External — people outside both organisations who constrain or depend on the solution: regulators, third parties whose APIs you call, and anyone who feeds the process without using the system.

Some stakeholders can be found in the case description, but most of them are hidden, and it is a real art to find all project stakeholders.

To find the hidden ones systematically, walk five trails through the case:

  • Follow the work — who performs, or is affected by, each step of the process, including the steps the RFP skips?
  • Follow the money — who pays, who gets paid, who reports and reconciles?
  • Follow the data — who is accountable for personal data, and who regulates it, in every jurisdiction?
  • Follow the rules — who decides what is correct (clinically, legally, commercially) when the system ranks, merges or combines?
  • Follow the interfaces — who owns each system and API you connect to, on both sides of the boundary?

Concrete tactics for surfacing the hidden ones: have an informal coffee with people around the project, build both formal and informal relationships, do background research on the company, and mine artefacts such as email threads, commit history, the author metadata on shared documents, and the client's own market and sales materials.

Power / interest prioritisation

Use Mendelow's power/interest grid (after A. L. Mendelow, 1981, in the form popularised by Johnson and Scholes in Exploring Corporate Strategy). The vertical axis is power — the ability to change the project's direction: budget, sign-off, veto, or control of a system or rule you depend on. The horizontal axis is interest — how much the outcome affects them and how closely they will follow it. How strongly a change affects someone belongs on the interest axis; it is not a separate dimension.

Low interestHigh interest
High powerKeep satisfied — meet their needs and consult them on decisions in their area, without burying them in detail; watch for their interest risingManage closely — the key players. Engage regularly, involve them in decisions and never let them be surprised. A sponsor almost always belongs here
Low powerMonitor — minimal effort: general communications, and a periodic check that their position hasn't changedKeep informed — tell them regularly what is happening and ask for their input; they are often the system's users and its best advocates

The grid is a snapshot, and positions move. A group with little power individually can hold real collective power — a few hundred users who won't adopt a system can sink its business case — and a stakeholder indifferent today becomes very interested once a change lands in their area. Revisit the grid at each phase. Each quadrant's approach then becomes a line in the communication plan: who, what they need to hear, how often, through which channel and from whom. The supplier's own project team is usually left off the grid: you work with them every day, so the grid adds nothing.

Beyond the power/interest matrix, two further techniques map stakeholder relationships from a different angle: the personal interest diagram and the social diagram. Both help avoid a common trap when identifying the powerful stakeholders — the most vocal person in the room is not necessarily the most powerful one, and real influence often has little to do with someone's official title, so both formal and informal relationships need to be observed.


RACI

  • Responsible — the person assigned to do the work
  • Accountable — the person who makes the final decision and has ultimate ownership
  • Consulted — must be consulted before a decision is made
  • Informed — must be informed that a decision or action has been taken

The most complicated part is understanding the difference between Accountable and Responsible. The accountable person signs off the work, has the authority to make a decision, and is the one whose head rolls if it goes wrong. The responsible person is responsible for the execution of the activity. Accountability comes before responsibility, and one activity must be associated with a single accountable person only.

A small example from a generic system replacement shows the rules at work:

ActivitySponsorBusiness leadPMArchitectDeveloper
Prepare and approve the business caseACRC
Choose the integration approachICCA/RC
Build and unit-test a featureIACR

Read across each row: exactly one A and at least one R. The sponsor is accountable for the decision only the sponsor can make; the PM writes the business case (R) without owning it; the architect both decides and does the work on the integration approach (A/R). Then read down the columns: a column full of A's marks a bottleneck, and an empty column is a stakeholder nobody involves.

A RACI covers the whole engagement, not only construction: initiation, discovery, design, build, test, transition and support each have different accountable people. A template can exist, but the project has its own RACI matrix tuned to the current project context.


Requirements (BABOK classification)

TypeDefinition
Business requirementsHigher-level needs of the enterprise
Stakeholder requirementsThe needs of stakeholders that must be met to achieve the business requirements
Solution requirementsSplit into functional and non-functional
Transition requirementsWhat the solution must do, or what must be true, only to get from the current state to the future one — converting data, training people, keeping the business running while old and new overlap. They are temporary: once the change is complete they are no longer needed, and that is what sets them apart from every other type

Functional requirements — what the system must do: the behaviour that lets a user finish a task, usually stated as what happens when a given condition holds. Write each one so that the stakeholder who asked for it and the developer who builds it read the same thing. On Lumen:

  • When a referral reaches Triaged and the matching rules find a suitable slot, the system sends the patient an offer by email and SMS.
  • The referral desk can search availability across partner slots and uploaded slots, filtered by modality, urgency, location and price.
  • When the referral system cancels a referral, the system cancels any offered or confirmed appointment for it and alerts the centre or partner concerned.

Non-functional requirements — not related to system functionality but defining how the system should perform:

  • Lumen: routine screens respond within 2 seconds at the 95th percentile with ≈1,650 concurrent sessions (today's derived peak).
  • Lumen: time to offer — from a referral being triaged to the offer being sent — is under 5 minutes at the 95th percentile for referrals the matching rules can offer automatically, sustained at the busiest-hour load of 1,680 referrals (14,000 a day × 12%). The 5 minutes is an assumed target to agree with Lumen.

Each example names what a testable NFR needs: a percentile, what is measured (a screen response, or a start and an end event), the load it must hold at, and the cases it covers. "Handle millions of users without deterioration" names none of them.

The other three requirement types are just as concrete on Lumen. A business requirement: reduce the desk cost of handling a referral from ~$9 towards the ~$2 automated target. A stakeholder requirement: the referral desk can see the current stage of every booking without phoning the centre or partner. A transition requirement: bookings already in progress at cutover are migrated with their current stage, and the desk runs the manual phone-and-email process in parallel for each region until that region's migrated bookings are reconciled. The transition requirement is the telling one: it costs real effort — migration scripts, reconciliation reports, double running — and none of it is needed once the last region is live, which is why it has its own category and its own line in the estimate.

Note the terminology debate: some people dislike "non-functional", because some such requirements — performance, for instance — might be a very important "function". They split the category instead into quality attribute (QA) requirements and constraints — a constraint being a requirement that removes design freedom (see the glossary), not just another name for a QA requirement.


Risks, assumptions, issues, dependencies

Together these make up the RAID log, one register with an owner and a next step for every entry.

TermMeaning
RiskAny specific uncertain event that might occur and affect the project or programme — usually negatively (a threat), although the PMBOK Guide and ISO 31000 also count positive risks (opportunities). Each has an associated probability of occurrence and an impact if it materialises. Example: a change in tax law could force rework, impacting the schedule by X and cost by Y. A risk management process must be undertaken, managing and mitigating risks and communicating them routinely and effectively to stakeholders
AssumptionSomething set as true to enable progress, typically during planning and estimation. Example: assuming access to 10 skilled specialists for the entire project duration — that assumption enables the plan. If it turns out false the project is negatively impacted, so all assumptions must be monitored and managed so minimal impact occurs
IssueAnything arising that must be dealt with to keep the project running smoothly. Issues differ from risks in that they exist as a problem today, unlike risks which might turn into issues in the future. Example: a key project resource has called in ill and is unlikely to attend for the rest of the week. Issues are managed through the issue management process
DependencyExists when an output from one piece of work is needed as mandatory input for another. Example: in a building project the architectural diagrams must be complete before the foundations can be laid. Managing inter-dependencies is critical regardless of project size

Two more lists sit next to the RAID log on any bid. Constraints are decisions already made for you — a fixed date, a mandated system, a regulation; you design within them rather than manage them. Clarification questions are gaps in the RFP. Ask them; until they are answered, log the answer you are planning on as an assumption with an owner and a confirm-by date, and its consequence if wrong as a risk. Never guess silently: a silent guess becomes unestimated scope.


Conway's Law

Conway's Law (Melvin Conway, "How Do Committees Invent?", Datamation, 1968) is an observation, not a rule: an organisation that designs a system tends to produce a design whose structure mirrors that organisation's own communication structure. Team boundaries — the supplier's teams, and the customer teams that own the systems you integrate with — tend to reappear as module and interface boundaries, so understanding the communication structure helps us design the modules and components and assign the right responsibilities to them; the inverse Conway manoeuvre deliberately shapes teams to fit the architecture you want. In the other direction, an optimised architecture might have an impact on the business by showing how work can be done more efficiently, or even just by exposing a gap.

Beyond formal, well-defined communication channels there are informal ones — a daily coffee chat. Identifying both is crucial to understanding the real need of the solution.

Note also that multiple solutions may be identified, not just one — and the solution is not necessarily limited by the boundaries of the customer organisation, and can cover or support multiple capabilities at the same time.

As a solution architect, put more focus on stakeholders, business processes and all the relevant requirements. Identifying the right stakeholders and maintaining a good relationship with them is crucial. Understanding the solution-related business processes helps identify more stakeholders and clarifies what the solution must support. Requirements define everything, including the constraints that impact the solution.


Analysis techniques (BABOK)

The techniques below are a selection from the 50 described in the BABOK Guide v3 (IIBA, 2015): those most useful for business architecture, including every technique this module applies. They are summarised here rather than quoted. Design-level modelling techniques (data, state and sequence models, use cases) are left to Module 5; read the guide for the full set and their detail.

TechniqueUse
Balanced ScorecardTrack how well strategy is being executed using measures beyond the purely financial, across four perspectives: Financial, Customer, Internal Business Process, Learning and Growth (Kaplan and Norton)
Benchmarking & Market AnalysisCompare the organisation's practices and results with best-in-class peers — competitors, public bodies, industry associations — and study the market, to find where operations, customer satisfaction or value can improve
BrainstormingGenerate many ideas quickly in a group, holding back judgement until the list is complete
Business Capability AnalysisMap what the business must be able to do, and how well it does it today, so that change can be aligned with strategy, scoped and prioritised
Business CasesArgue for a course of action by weighing the benefits a solution should deliver against the cost and effort of acquiring and running it
Business Model CanvasA one-page view of how an enterprise creates value for its customers, delivers it and earns from it (Osterwalder and Pigneur)
Business Rules AnalysisFind, state, check and organise the rules that govern everyday business behaviour and operational decisions
Collaborative GamesStructured, game-like exercises that get elicitation participants to build a shared picture of a problem or solution
Concept ModellingPin down the business terms of a domain and how they relate, so everyone uses the same vocabulary
Data DictionaryKeep one agreed definition per data element, so everyone reads the data the same way
Data Flow DiagramsTrace data from its sources, through the activities that transform it, to where it is stored or consumed
Data MiningSearch large data sets for patterns that improve decisions
Decision AnalysisCompare options systematically by valuing their possible outcomes when the future is uncertain
Document AnalysisMine existing material — policies, manuals, contracts, system documentation — for context and requirements
EstimationForecast the cost and effort of a course of action; methods include top-down, bottom-up, parametric, rough order of magnitude, rolling wave, Delphi and PERT
Financial AnalysisWeigh the costs, benefits and financial risk of an investment or solution option
Functional DecompositionBreak a process, system, functional area or deliverable into smaller parts that can each be analysed on its own, to tame complexity
Interface AnalysisFind every point where information crosses between solution components or the solution boundary — what is exchanged, when, how, why and for whom
InterviewsPlanned conversations with an individual or small group, with prepared questions and recorded answers
Item TrackingLog the issues and stakeholder concerns that affect the solution, each with a named owner
Lessons LearnedRecord what worked, what failed and what to change, so the next project or phase does better
Metrics and KPIsDefine indicators that show how well a solution or component performs against what stakeholders care about
Mind MappingCapture ideas branching out from a central topic, to explore them and connect them
Organisational ModellingShow roles, responsibilities and reporting lines, and check that they fit the organisation's goals
PrioritisationHelp stakeholders agree what matters most, using an explicit, shared method
Process AnalysisJudge how efficient and effective a process is, and where it could change for the better
Process ModellingDraw how work flows — who does what, in what order, with which decisions — to expose hand-offs, the stakeholders involved and the steps a system must support
Risk Analysis and ManagementFind the uncertainties that could affect a change, rate their likelihood and impact, decide how to treat each, and keep watching them
Root Cause AnalysisAsk "why" repeatedly, or draw a cause-and-effect diagram, to get from a symptom to the cause behind it
Scope ModellingDraw the boundary of the change — what is in, what is out, and what sits at the edge — often as a context diagram of the solution and the people and systems around it
Stakeholder List, Map, or PersonasList who is affected, map their power and interest, and describe representative users as personas
SWOT AnalysisSort internal strengths and weaknesses and external opportunities and threats, to frame a strategy or a bid
WorkshopsBring stakeholders together in a facilitated session with a set goal, to agree objectives, rules or priorities faster than a series of interviews could

Where this module uses them: brainstorming, interviews, workshops, mind mapping, organisational and process modelling and stakeholder lists, maps and personas to find stakeholders; root cause analysis to find drivers; business capability analysis and functional decomposition for the capability work; risk analysis for the RAID log; and scope modelling for the boundary the capability map and the constraints draw. SWOT frames the bid itself: a weakness such as Lumen's missing utilisation data, or the regulatory threat in a new jurisdiction, shapes the proposal as much as the opportunity to grow.

Benefits Dependency Network (BDN) (Ward and Daniel, Benefits Management). Useful because it keeps focus on benefits realisation during programme execution, and allows variations of the project or programme to be assessed for their impact on benefits realisation. A well-constructed BDN tells the story of the project visually: in one diagram it shows why the programme is needed, what objectives it aims to achieve, how the organisation needs to change, and what projects must be undertaken. It can be read left-to-right or right-to-left, helps identify critical paths, can serve as the basis for the programme plan, and enables discussion of the relative contributions of the different projects — which in turn enables the right resource-allocation decisions.

Business processes are a huge topic that this module does not cover directly — process-modelling notations such as BPMN and swimlanes are in Module 5 — but value streams and capability mapping, above, are a great baseline for further discussion.

Check yourself

  1. The case lists "Integration" among six aspects that are "effectively the business drivers". Why is it not a driver, and what is it?
  2. Is "increase partner utilisation by 10%" a goal or an objective for Lumen? What would make it SMART?
  3. The referral desk has about 350 clerks, each with little individual power. Why might the group need more attention than its quadrant suggests?
  4. A RACI row shows both the PM and the Referral SME as A. What is wrong, and what are two ways to fix it?
  5. Is "Partner portal" a business capability? Explain with the phone-and-paper test.
  6. In a value stream, what is the difference between a stage and a capability, and can one capability enable several stages?
  7. The RFP does not say how the Billed stage reaches finance. What do you do with that gap on a bid, and why not simply pick an answer?
  8. Give one transition requirement for Lumen, and say why it is not just another functional requirement.

Suggested answers

  1. It would not matter if no software existed: it names systems the solution must reuse and says nothing about the business. It is a constraint, recorded as interface requirements traced to the drivers it serves (cost of referral handling, compliance).
  2. As written it is too specific for a goal and too vague for an objective. A SMART version names the measure, the baseline, the target and the date — "raise partner-slot utilisation by 10 percentage points against the discovery baseline within 12 months of go-live" — and needs a prior objective to measure the baseline, because Lumen has no reliable utilisation data.
  3. Collective power: the cost objective depends on the desk using the platform instead of the phone, so 350 clerks who resist adoption can sink the business case. The grid is a snapshot; groups move quadrants when their collective behaviour starts to decide the outcome.
  4. Two Accountables break the one-A rule, so nobody truly owns the decision. Either choose one A and make the other C, or split the row into two activities with one A each — for example, the referral-system integration (customer-owned) apart from the SSO and partner-API integration (supplier-owned).
  5. No. If Lumen went back to phone and paper, the portal would vanish but the business would still need to manage partners and their capacity; those are the capabilities, and the portal is a system function that supports them.
  6. A stage is a step in sequence that delivers a value item the stakeholder recognises; a capability is a stable ability of the business that enables stages. Yes — patient communication, for example, enables several Lumen stages.
  7. Raise it as a clarification question, log the answer you are planning on as an assumption with an owner and a confirm-by date, and log the consequence of being wrong (a finance integration) as a risk. Guessing silently hides scope and cost from both sides.
  8. Migrating in-flight bookings with their current stage at cutover, or running the manual process in parallel per region until migrated bookings reconcile. Both are needed only during the change and disappear afterwards, yet they cost real effort and must be estimated.

Exercises — Module 2

Task. Produce a business-architecture pack for Lumen Diagnostics. Output: one document of about 6–10 pages, mostly tables: a driver → goal → objective table; a stakeholder register; a power/interest grid; a RACI for the whole engagement; a level-1 capability map, a value stream and a functional decomposition (not a WBS); a constraints and assumptions list; and a RAID log of at least five entries. Time: about 5 hours. Inputs: the case — its six aspects, the Volumes and usage table and the two clarification questions at the end — and the stakeholder, concern and view vocabulary from Module 1.

  • Sort the six aspects into drivers, requirements and quality attributes; add any drivers the case leaves out, and tag each driver with a business-case category
  • Write goals and SMART objectives for each driver, with a baseline-measurement objective where Lumen has no data
  • Document stakeholders
    • Fields: Name, Role, Group, Concerns, Artefacts that address their concerns
  • Place stakeholders on Mendelow's power/interest grid, with a one-line reason for each placement
  • Create a RACI for the whole engagement — exactly one A and at least one R per row
  • Draw a level-1 capability map, trace each capability to an objective, and map the capabilities onto a value stream for one referral
  • List the main system functions (functional decomposition) and trace each to a capability
    • Do not produce a detailed WBS; that is out of scope here
  • List constraints and assumptions, and seed a RAID log that includes the case's two clarification questions

Self-assessment

CriterionWhat good looks like
DriversEvery driver is a noun phrase that would matter to Lumen with no system, tagged with one of the four categories. Requirements and quality attributes are traced to drivers, not listed as drivers. Compliance and continuity are not missing
ObjectivesEach has a measure, a baseline (or an objective to measure one), a target and a date. No goal carries a number. Targets you invented are labelled as assumptions
StakeholdersThe register includes people the case never names — at least a clinical owner, a compliance owner, finance and one external party — and every concern points to an artefact that will address it
Grid and RACIEvery grid placement has a reason, and the sponsor is in Manage closely. Every RACI row has exactly one A and at least one R, and Lumen people are accountable for the decisions only Lumen can make
CapabilitiesCapabilities are nouns that would survive a return to phone and paper. Each traces to an objective, and each system function traces to a capability
GapsThe two clarification questions appear as assumptions or risks with an owner and a date, not as silent guesses in the design

Worked answer — Lumen business architecture

Attempt the Task before reading this. It is one defensible answer, not the only one. Every target in it is an illustrative assumption to agree with Lumen; every figure not in the case is labelled as derived or assumed.

Drivers, goals and objectives

The six aspects, put through the driver test:

Case aspectWhat it isTraces to
Cost reductionDriver — cost of referral handlingD1
AutomationSolution strategy — the means by which cost falls. It becomes functional requirements: automated offers, the partner-API bookingD1, D2
IntegrationConstraint — the referral system, the patient identity service and corporate SSO must be reusedD1, D4
ConsistencyQuality attribute — usability: one look, feel and process across every portalD1
FlexibilityBoth: the growth behind it is a driver; "highly configurable" is a quality attribute (modifiability, Module 3 QA10)D3
PerformanceQuality attribute — performance and availability. The business condition behind it, continuity through outages and surges, is a driverD5
(not an aspect)Driver — health-data regulation in every region, and a new jurisdiction within three yearsD4

Only one aspect, Cost reduction, is a driver as written. Flexibility and Performance are quality attributes with a driver behind each (D3, D5); Automation, Integration and Consistency are a strategy, a constraint and a quality attribute serving drivers. D2 comes from the case's problem statement — unused partner capacity, no utilisation data — and D4 from the data-security NFR and the growth plan.

DriverBusiness-case categoryEvidence in the case
D1 Cost of referral handlingReduce cost / increase efficiency~$9 of desk time per manual referral against a ~$2 automated target; ~350 desk staff
D2 Utilisation of imaging capacityIncrease or protect revenuePartner capacity goes unused because nobody can see it; no reliable utilisation data
D3 Capacity for growthIncrease or protect revenueReferrals +12% a year (×1.405 over three years); partner practices to double; two new regions
D4 Regulatory compliance in each jurisdictionCompliancePatient data must meet health-data regulation in each region; one new region is outside the current jurisdiction
D5 Continuity of the booking serviceReduce riskCentre outages and regional surges; outage-related cost accepted only up to ~$100,000 a year
Goal (qualitative)Objectives (SMART)
G1 Reduce the cost of turning a referral into an appointment (D1)O1.1 Measure the baseline — desk minutes per referral, status-enquiry calls and slot utilisation, by region — within 3 months of contract signature. O1.2 Cut the average desk cost per referral from ~$9 to $5.50 or less within 12 months of each region's go-live; that needs half of referrals booked with no desk handling (9 − 7 × 0.5 = 5.50). O1.3 Serve utilisation and financial reporting from the platform alone, with no side spreadsheets, from the last region's go-live
G2 Make every booking and every free slot visible to the desk, the partners and the patient (D2)O2.1 100% of confirmed bookings held in the platform as the single source of truth from each region's go-live. O2.2 Status-enquiry calls to the desk down 50% against the O1.1 baseline within 6 months of go-live. O2.3 Partner-slot utilisation up 10 percentage points against the baseline within 12 months of go-live
G3 Grow into new regions and twice as many partners without rebuilding the platform (D3)O3.1 Clinical-protocol, policy and prioritisation-rule changes applied by the configuration team within one working day, without a code release, from go-live (Module 3, QA10). O3.2 Year-3 volumes — ≈2,360 referrals in the busiest hour (1,680 × 1.405) and ≈1,200 partner practices — proven by load test before the first region goes live. O3.3 A new partner practice onboarded (slots, pricing, capability) within 2 working days without supplier involvement, from go-live
G4 Process patient data lawfully in every region, including the new jurisdiction (D4)O4.1 A data-protection impact assessment (DPIA) approved by Lumen as controller, with the Data Protection Officer's (DPO's) advice recorded, for each region before go-live. O4.2 The new-jurisdiction region launches with its residency and cross-border-transfer position confirmed by the DPO and built into the deployment
G5 Keep referrals flowing through outages and surges (D5)O5.1 Outage-related cost within $100,000 a year from go-live: at most ≈8.5 hours of booking-path outage in operating hours a year (100,000 ÷ 11,760, the cost of a peak hour: 1,680 referrals × $7 extra manual cost) — the method is in Module 3. O5.2 When a centre closes unexpectedly, every affected patient receives a new offer within 4 operating hours

G2 exists because G1 has a consequence: automation that nobody can see is still a desk full of phone calls. Goals are strategic, set over months or years; the objectives carry the metrics that prove achievement. O1.1 comes first because the case says Lumen has no reliable utilisation data — O2.2 and O2.3 cannot be measured without it.

Stakeholder register for Lumen

The case names some roles (Scheduling Manager, referral desk, partner practices, patients). The rest come from walking the five trails: the work (centre staff and radiographers), the money (finance), the data (the DPO, the regulators), the rules (the clinical lead) and the interfaces (the integration SMEs and the partner group's API owner).

NameRoleGroupConcernsArtefacts that address their concerns
Maya ChenPatientUserUsability on a phone (~45% of sessions are mobile), accurate appointment detail, confirming or declining without phoningPatient-portal prototype, notification copy
Omar HaddadReferral-desk clerk (one of ~350)UserTime to offer, conflict handling, manual booking formsDesk UI prototype, workflow stage model
Priya NairScheduling ManagerUserRule configuration, visibility of status, reportingAdministration-portal prototype, rule-configuration model, report specifications
Luca RossiPartner-practice coordinator (one of ~1,500 partner users)UserEffort of uploading slots, pricing, which bookings they receivedPartner-portal prototype, partner onboarding guide
Kwame MensahCentre manager, speaking for ~900 centre staff including radiographersUserHow the centre's own availability is published (clarification Q1), an accurate day list, combined visits that fit the equipmentCentre schedule view, availability data flow, consolidation rules
Elena VossSponsor / COOCustomerCost per referral, first release within 12 months, utilisation, regional expansionBusiness case, benefits dependency network, conceptual diagram
Dr Samuel AdeyemiClinical leadCustomerWhat counts as a "clinical match"; which studies are "clinically permitted" to combine; safe urgency rulesPrioritisation-rule catalogue, consolidation rules, clinical-safety review
Ingrid HolmData Protection Officer (compliance owner)CustomerHealth-data regulation in each region, residency for the new jurisdiction, retention (18 months online, then archive), access auditDPIA, data-flow and deployment views, retention policy
Marcus FeldFinance directorCustomerHow Billed reaches finance (clarification Q2), financial reporting, outage cost within $100,000 a yearFinance interface specification, financial report specification, cost model
Jonas BergReferral-system SMECustomerIntegrability, the 15-minute change extract, merge policyIntegration data flow, context diagram
Amira SoléPatient-identity and SSO SMECustomerCorporate SSO for staff, the patient identity service, session handlingSequence diagrams, security view
Tom ReyesChief Information Security OfficerCustomerAccess control, encryption (TLS 1.3 or better), assurance over a supplier-hosted serviceSecurity view, threat model, penetration-test report
George WalterProject ManagerProject TeamQuality, risk, cost, staffingPlan, RAID log
Andy DonaldSolution ArchitectProject TeamArchitecture, constraints, requirementsContext, component and deployment views
Nadia PetrovaBusiness AnalystProject TeamComplete, testable requirements; baseline measurement; business rulesRequirements catalogue, process models, baseline report
Lena FischerUX designerProject TeamOne coherent look, feel and process across portals; mobile useDesign system, prototypes, usability-test results
Ravi KumarDeveloperProject TeamScope, interfaces, environmentsSequence diagrams, component view, interface specifications
Sachin DaveTesterProject TeamTestability of workflow and search, test dataUse cases, performance scenarios
Satty KumarDevOpsProject TeamPromotion to live, multi-region deployment, CI/CDDeployment view, pipeline design
Rada KandolaIT Support (supplier service desk)Project TeamOperability, failing notifications, support hand-overOperations runbook, monitoring dashboards
Health-data regulators (one per jurisdiction)Supervisory authorityRegulator/ExternalLawful processing, breach notification, cross-border transferReached through Lumen's DPO: DPIA, records of processing
Dr Laura BrandtReferring clinician (representing referring clinics)Regulator/ExternalUrgent referrals offered quickly; knowing the patient has been bookedTime-to-offer report, booking status fed back through the referral system
Kenji WatanabeAPI owner, automated partner groupRegulator/ExternalCall volumes and rate limits, versioning, the data shared with LumenInterface contract, integration test plan

Put each concern with the person who can decide it: health-data regulation belongs to the DPO, not to the identity SME, who owns sign-in and sessions.

Power/interest placement

Low interestHigh interest
High powerKeep satisfied: finance director (budget, outage tolerance); CISO (can block go-live on security); patient-identity and SSO SME (controls access to SSO); partner group's API owner (can refuse changes to its API); health-data regulators (engaged only through the DPO)Manage closely: Sponsor/COO; Scheduling Manager (owns configuration and UAT sign-off); clinical lead (veto over clinical rules); DPO (gives advice a region's launch will not proceed against); referral-system SME (owns the feed every referral arrives through, and an integration on the critical path)
Low powerMonitor: referring clinicians — their referrals still arrive through the referral system, so their work barely changes; move them up if the platform changes what a referral must containKeep informed: referral-desk clerks, centre managers and radiographers, partner-practice coordinators, patients (through launch communications, usability research and portal feedback)

Two placements to argue over. The desk clerks are individually low-power, but O1.2 depends on all ~350 of them using the platform — treat them as a group that moves to Manage closely if adoption stalls. The finance director stays in Keep satisfied only until clarification Q2 is answered: if Billed needs a real integration, finance becomes a key player.

RACI for the whole engagement

Key: Sched = Scheduling Manager; Clin = clinical lead; DPO = Data Protection Officer; Desk = referral desk; Ptnr = partner practices; Id SME = patient-identity and SSO SME; Ref SME = referral-system SME; BA = Business Analyst; UX = UX designer; IT Sup = IT Support.

ActivitySponsorSchedClinDPODeskPtnrId SMERef SMEPMSABAUXDevTesterDevOpsIT Sup
Project initiationACICRC
Discovery and baseline measurementICCCACRC
Project planningCCA/RCCCCCCC
RequirementsCCCCCCCARRCCCC
Clinical rules (match, urgency, consolidation)CACICRII
Prioritisation rulesA/RCCICRII
UX design and prototypesCCCACCRC
ArchitectureICCCCCCA/RCCCCCC
Data-protection impact assessment (per region)ACCCRCC
DocumentationICIIIIARRCRRRC
DevelopmentIIIACCCRCR
Integration — referral systemIACCCRRR
Integration — SSO, patient identity and partner APIICCCACRRR
System and performance testingIACCCRCI
StabilisationCCACCRRRC
UAT (user acceptance testing)IACRRCCCCICC
Training and cutover (per region)ICCCCACRCRR
Go-live decision (per region)ACCCCRCCC
Support (hosted service)ICIIIIIACIIRR

The PM is accountable for most delivery activities — but not for everything. The Sponsor is accountable for starting the project and for each region's go-live decision; the clinical lead for the clinical rules and the Scheduling Manager for the prioritisation rules; the Sponsor for each region's impact assessment, on the DPO's advice; the SA for the architecture itself. UAT sign-off sits with the Scheduling Manager, while the people who will use the system — referral desk and partner practices — execute the tests, with the supplier's team in support. Integration of the referral system has a customer accountable (the Referral SME), so it sits in its own row, apart from the SSO and partner-API integration the PM owns. Support's A sits with the PM through the warranty period and passes to the supplier's service manager when the service contract starts; add that column when it does.

Capability map (level 1)

Eleven business capabilities cover the scope. Each is a noun, each traces to an objective, and each is rated today and at target. The ratings are assumptions to validate in discovery; the Low → High rows are where the investment goes.

CapabilityWhat it coversToday → targetTraces to
Referral managementReceiving, validating and tracking referrals, amendments and cancellationsMedium → HighO2.1
Clinical protocol managementClinical-match criteria, urgency, which studies may be combinedLow → High (rules live in people's heads)O3.1
Scheduling policy managementThe Scheduling Manager's prioritisation rules and preferred locationsLow → HighO3.1
Capacity managementAvailability of Lumen centres and partner practices in one viewLow → High (partner capacity invisible today)O2.3, O1.2
Appointment managementOffering, confirming, rescheduling, reassigning and combining appointmentsLow → High (phone and email)O1.2, O2.2
Partner managementOnboarding partners; their pricing, capability and performanceLow → MediumO3.3, O2.3
Patient communicationKeeping the patient informed at every stage; capturing feedbackLow → HighO2.2
Billing and settlementCharging for completed studies and settling partner feesMedium → Medium (pending clarification Q2)O1.3
Performance reportingUtilisation, partner performance and financial reportingLow → High (no reliable data)O1.1, O1.3
Information governanceLawful processing, access control and audit, residency, retentionMedium → High (new jurisdiction)O4.1, O4.2
Service continuity managementKeeping bookings flowing through centre outages and regional surgesLow → MediumO5.1, O5.2

Value stream — Fulfil imaging referral

Triggered by a referring clinician's referral; the beneficiary is the patient. The stages follow the value the patient and clinician receive, not the workflow states, although the two line up.

StageValue itemWorkflow stagesEnabling capabilities
Accept referralReferral registered, patient identified, requirements recordedRequestedReferral management, Information governance, Patient communication
Triage referralClinical requirements and urgency agreedTriagedClinical protocol management, Referral management, Patient communication
Match capacitySuitable slots ranked by clinical match, location, price and rules(before Offered)Capacity management, Scheduling policy management, Partner management, Patient communication
Agree appointmentAn appointment the patient has accepted and the provider has confirmedOffered → Accepted/Declined → ConfirmedAppointment management, Patient communication
Perform studyStudy completed, combined into one visit where clinically permittedCompletedAppointment management, Clinical protocol management, Service continuity management, Patient communication
SettleStudy charged and partner paidBilledBilling and settlement, Partner management, Performance reporting, Patient communication

Value proposition: the patient is seen at the right place, at the right time, for the right study at the right price — and the referring clinician knows it. Patient communication appears on every stage: each stage change notifies the patient, which is why O2.2 sits on it.

Functional decomposition (system functions)

These are the functional parts the platform must have — system functions, portals and adapters, not business capabilities. At this stage do not pick a stack, and do not produce a WBS. Each function traces to the capability it supports:

  • Identity and access — corporate SSO for staff; patient-portal sign-in through the patient identity service → Information governance
  • Availability search and prioritisation — partner slots plus uploaded slots, ranked by clinical match and Scheduling Manager rules → Capacity management, Scheduling policy management
  • Offer and confirmation workflow — the Requested → … → Billed stages, including manual booking by the desk → Appointment management
  • Change management / referral sync — inbound updates from the 15-minute extract, merge policy, alerts, manual reassignment → Referral management
  • Partner slot management — manual upload of availability, pricing and capability → Capacity management, Partner management
  • Partner scheduling adapter — the one automated partner group (~15% of partner bookings); the rest stay on the manual path → Capacity management, Appointment management
  • Centre availability publishing — how Lumen's own 180 centres publish slots; not in the RFP: clarification Q1 → Capacity management
  • Patient portal — view, confirm or decline, print, feedback → Patient communication, Appointment management
  • Administration portal — configuration, status, reporting → Scheduling policy management, Clinical protocol management
  • Notifications — email and SMS on stage change → Patient communication
  • Reporting — appointment detail, partner performance, financial → Performance reporting
  • Billing hand-off — passing Billed studies to finance; not in the RFP: clarification Q2 → Billing and settlement
  • Appointment consolidation — combine studies into one visit where clinically permitted → Appointment management, Clinical protocol management

Each function traces through a capability to an objective. Without the compliance driver (D4) and the cost objective (O1.2: one booking instead of several), identity and consolidation would trace to nothing and look like invented scope — a function with no trace can mean a missing driver, not just extra scope. The trace also runs backwards: Capacity management covers Lumen's own centres, but no requirement says how their availability arrives. That is how the capability map surfaces clarification Q1.

Constraints and assumptions

Constraints — from the case; design within them:

  • Delivered as a service, hosted and supported by the supplier
  • Integrate with the existing referral management system (a change extract every 15 minutes, no event or webhook interface), the patient identity service and corporate SSO
  • All authenticated traffic over TLS 1.3 or better; patient data handled under the health-data regulation of each region
  • Currently supported technology versions, with planned upgrades
  • First release live within 12 months of contract signature; a phased roll-out by region is allowed
  • One partner group integrated through its scheduling API; all other partners manual

Assumptions — ours; each needs an owner and a confirm-by date:

  • Lumen's centres publish availability by upload through the platform, as partners do (clarification Q1)
  • Billed studies leave the platform as a daily export to finance, with no real-time finance integration in the first release (clarification Q2)
  • Lumen's DPO provides the regulatory position for each region, including the new jurisdiction, before that region's design is fixed
  • The automated partner group's API has a published contract and a test environment the supplier can use from the start of construction
  • Lumen's clinical lead, Scheduling Manager and desk representatives are available for rule workshops and UAT when the plan needs them

RAID log (seeded)

IDTypeDescriptionOwnerResponse / next step
A1AssumptionLumen's 180 centres publish availability by upload through the platform (clarification Q1). If they run their own scheduling system, that is a new integration and a re-estimateScheduling ManagerRaise in the clarification round; confirm by the end of discovery
A2AssumptionBilled studies reach finance as a daily export, with no real-time integration in the first release (clarification Q2)Finance directorRaise in the clarification round; confirm by the end of discovery
R1RiskLumen may read "frequent synchronisation" as "within 15 minutes of the change", which polling a 15-minute extract cannot guarantee (worst case ≈ 15 min + processing ≈ 20 min). Likelihood high, impact mediumSolution ArchitectAgree the target as "within 5 minutes of the extract that carries them" (Module 3, QA11); ask the referral-system owner whether a shorter interval or a change feed is possible
I1IssueLumen has no reliable utilisation or status-call data, so O1.2, O2.2 and O2.3 cannot be measured yetBusiness AnalystRun the baseline measurement (O1.1) in the first 3 months
D1DependencyThe deployment topology for the new-jurisdiction region depends on the DPO's regulatory position (residency, cross-border transfer)DPORequest the position before that region's design starts; until then, design for in-region data storage