Enterprise Architecture
Enterprise Architecture (EA) is the discipline of aligning business strategy with the technology landscape โ applications, data, infrastructure, and integrations โ so that change happens deliberately rather than accidentally. Done well, EA answers four questions: (1) What capabilities does the business need? (2) What systems support each capability today? (3) Where are the gaps, overlaps, and risks? (4) What is the target state and the sequenced roadmap to get there? KnowMBA POV: Enterprise architecture is a coordination function, not a documentation function. The deliverable is not a 400-page TOGAF binder โ it is a shared mental model that lets product, engineering, and finance make consistent decisions across hundreds of initiatives.
The Trap
The classic trap is the 'Ivory Tower EA' โ a central team produces beautiful TOGAF artifacts (Business, Data, Application, Technology layers), publishes them on a SharePoint nobody reads, and demands every project pass an architecture review board that adds 6-8 weeks. Engineering teams route around them. Within 18 months the EA team is dissolved. The opposite failure: 'no EA at all,' where every team picks its own database, queue, and auth provider, and within 4 years you have 14 CRMs, 9 identity systems, and integration spaghetti that costs $40M/year just to maintain. Real EA sits in the middle: lightweight standards, embedded architects, and a target-state map that gets updated quarterly.
What to Do
Build EA as a coordination layer, not a gate: (1) Define a Business Capability Model โ 30-50 capabilities the business needs (e.g., 'Order Management', 'Customer Identity'). (2) Map every major application to the capabilities it supports โ this surfaces overlaps (you have 6 apps doing 'Quote-to-Cash'). (3) For each capability, score current-state health (red/yellow/green) on cost, agility, risk. (4) Define a target-state architecture for the worst 5-10 capabilities. (5) Embed architects in delivery teams โ not in a central review board. (6) Measure EA by business outcomes (time-to-launch, integration cost, app rationalization) โ never by 'number of artifacts produced.'
Formula
In Practice
Capital One rebuilt its enterprise architecture function around 'cloud-first, API-first, data-first' principles starting in 2014. Instead of a central review board, they embedded architects in product teams and published a shared 'paved road' of approved patterns (AWS, microservices, REST APIs, specific data stores). Teams that followed the paved road got fast-track deployment; teams that deviated had to justify it. The result: closure of all 8 data centers by 2020 and the first US bank fully on the public cloud.
Pro Tips
- 01
The single most useful EA artifact is the Application Portfolio map: every app, its owner, its annual cost, its capability, its lifecycle stage (invest/maintain/sunset). Most companies don't have this โ and can't make rationalization decisions without it.
- 02
Replace the Architecture Review Board with an Architecture Guild โ a community of practice where senior engineers from each team meet biweekly. Standards emerge from peers, not edicts. Adoption rates triple.
- 03
Use the 'Pace Layered' model (Gartner): Systems of Record (slow, reliable, ERP-class) vs Systems of Differentiation (mid-pace, your competitive logic) vs Systems of Innovation (fast, disposable, experiments). Apply different governance to each layer โ one-size-fits-all kills the fast layer.
Myth vs Reality
Myth
โEA needs to document every system in detail before adding valueโ
Reality
By the time the documentation is complete, half of it is wrong. Real EA value comes from making 5-10 high-stakes decisions correctly per year (vendor consolidation, target architecture, build-vs-buy, M&A integration). The documentation supports those decisions; it is not the deliverable.
Myth
โTOGAF certification makes someone an enterprise architectโ
Reality
TOGAF teaches a vocabulary, not judgment. The best enterprise architects have shipped large systems, lived through migrations that failed, and understand the politics of getting 12 VPs to agree on a target state. Certification is table stakes; experience and influence are the actual job.
Try it
Run the numbers.
Pressure-test the concept against your own knowledge โ answer the challenge or try the live scenario.
Knowledge Check
Your company has 14 different CRM systems across business units, costing a combined $28M/year. The EA team proposes consolidating to one. The first thing they should produce is:
Industry benchmarks
Is your number good?
Calibrate against real-world tiers. Use these ranges as targets โ not absolutes.
Application Portfolio Redundancy (Large Enterprise)
Enterprises with 100+ applications, post-M&A or with decentralized procurementDisciplined
< 15%
Healthy
15-25%
Typical
25-40%
High Bloat
40-55%
Out of Control
> 55%
Source: Gartner Application Portfolio Management Research, 2024
Real-world cases
Companies that lived this.
Verified narratives with the numbers that prove (or break) the concept.
Capital One
2014-2020
Capital One's CIO Rob Alexander rebuilt enterprise architecture as a 'paved road' rather than a gating function. The EA team published reference architectures for cloud (AWS), APIs (REST + standardized auth), and data (specific stores per use case). Teams following the paved road got fast deployment; deviations required justification. Architects were embedded in product teams instead of sitting in a central review board. Over 6 years, Capital One closed all 8 data centers, became the first US bank fully on AWS, and reduced application count by ~40%.
Data Centers Closed
8 โ 0
Application Reduction
~40%
Time to Deploy New Service
Months โ Days
Architecture Review Time
Weeks โ Hours (for paved road)
EA scales by enabling teams to make good decisions quickly, not by reviewing every decision centrally. Paved roads + embedded architects beat review boards.
Decision scenario
The Architecture Review Board Crisis
You are the new VP of Enterprise Architecture at a 12,000-person insurance company. The existing Architecture Review Board has 45 reviews in queue with average wait of 9 weeks. Engineering leaders are publicly criticizing EA on internal channels. The CIO wants you to 'fix this' โ but the board was set up after a $4M security incident two years ago caused by an unreviewed third-party tool.
Review Queue
45 projects
Avg Wait Time
9 weeks
EA Team Size
14 architects
Engineering Headcount
1,800
Annual EA Cost
$4.2M
Decision 1
You can either double down on the board (more architects, faster reviews) or restructure EA entirely. The CFO will not approve more headcount. The CISO is nervous about loosening controls after the security incident.
Hire 6 more architects (offshored to reduce cost), cut review SLA to 3 weeks, keep the board as the single decision authorityReveal
Embed 8 of your 14 architects directly in product groups. Replace the board with a 'paved road' (pre-approved patterns + automated security scanning) and an Architecture Guild for novel decisions. Keep gating only for high-risk: PII data flows, third-party access, and >$1M spend.โ OptimalReveal
Related concepts
Keep connecting.
The concepts that orbit this one โ each one sharpens the others.
Beyond the concept
Turn Enterprise Architecture into a live operating decision.
Use this concept as the framing layer, then move into a diagnostic if it maps directly to a current bottleneck.
Typical response time: 24h ยท No retainer required
Turn Enterprise Architecture into a live operating decision.
Use Enterprise Architecture as the framing layer, then move into diagnostics or advisory if this maps directly to a current business bottleneck.