Why digital transformation fails and what we can do differently

We don’t have a technology problem in Australia and New Zealand. We have a delivery‑to‑adoption problem.

Close to home: Western Australia’s Auditor‑General reviewed 10 major public IT projects: costs are up by about A$1.6 billion, with eight of them delayed. That’s not a blip; that’s a pattern. (ABC)

Globally, the signal is the same. Across 5,400 large IT projects, outcomes ran 45% over budget, 7% over schedule, and delivered 56% less value than expected. Only 48% of digital initiatives meet or exceed their business outcome targets, and ~80% of software features in the average product are rarely or never used. Delivery is hard, but adoption is often where value goes to die. (McKinsey & Company

What “failure” looks like here (in IT & operations)

• Public safety meets IT: Network changes at Optus disrupted Triple Zero (000). The regulator fined the telco A$12 million for breaching emergency‑call rules after the 2023 outage; a 2025 incident during a network upgrade has been linked in reporting to multiple deaths. If that isn’t “IT with operational consequences,” nothing is. (ACMA)

• Ticketing tech without proportionate benefit: Victoria’s myki took >9 years (vs ~2 originally), with ~A$550m (+55%) budget increase; auditors say delays and cost increases compromised benefits. (Victorian Auditor-General's Office)

• Smart meters without behaviour change: Victoria’s AMI rollout left consumers paying billions while benefits lagged; VAGO warned of a potential A$319m net cost to consumers and “significant uncertainties” in benefit realisation. (Victorian Auditor-General's Office)

• Payroll “back office,” frontline pain: NZ’s Novopay went live and the first pay run underpaid ~5,000 staff and left 15 unpaid. Technology “delivered,” but operations and people bore the pain. (NZLII)

• Commissioning where IT and OT meet: At Fiona Stanley Hospital, the identity & access project was terminated after ~A$6m spend; the Auditor‑General warned access wouldn’t be automated at opening -> an integration + commissioning miss. (ABC)

Thread these together and you get a simple conclusion: we design for technology, not for Tuesday‑morning behaviour. Programs “finish,” but value doesn’t land.

Five patterns I keep seeing behind stalled transformations

1. Technology‑first, problem‑second Stacks get chosen before a Value Case is written. Activity looks impressive; outcomes don’t.

2. Big‑bang launches, thin early value Lots of streams, little that users can actually touch. Sponsorship fades.

3. Project funding, not product ownership Artifacts delivered; no persistent owner with decision rights to chase outcomes.

4. Vanity metrics, not value metrics We count features and training instead of time‑to‑first‑value, adoption, and cost‑to‑serve.

5. Change as comms, not behaviour Broadcast emails can’t compete with incentives, rituals, and tools.

The quiet failure mode: no one owns the outcome

This is where most transformations die. We appoint a system owner and forget the benefit owner.

• Network upgrades: The CIO owns architecture and uptime. But the benefit, employee productivity (time from boot to first productive click; meeting join time) lives with HR/Operations. If no one is on the hook for the Tuesday‑morning metric, the “upgrade” is invisible to the business.

• Facilities tech (BMS, access control, energy): Corporate Services “owns” the system. But the benefit - lower kWh/m², fewer reactive call‑outs, reduced insurance risk - sits with the CFO and risk owners. Without a benefits cadence into Finance, value stays notional.

• Critical services: When emergency call completion is the real KPI, treating change as an IT SLA is dangerous. The outcome owner is the executive accountable for public safety; governance, testing, failover and comms must be designed to that standard. (ACMA)

What works: appoint a Senior Responsible Owner (SRO) who sits where the benefit lives and make them the single point of accountability for benefits realisation. That’s standard in public‑sector guidance for good reason. (Digital.gov.au)

What “bad” vs “good” looks like (AU/NZ examples)

1) Network change & public safety

• Bad: Firewall/routing changes rolled with limited scenario testing and no clear owner for public‑safety outcomes.

• Good: The SRO is a safety/operations exec; change criteria include 000 call completion SLOs, live monitoring and rehearsed rollback, not just IT SLAs. (ACMA)

2) Smart meters & energy outcomes

• Bad: Devices deployed; households foot costs; benefits depend on behaviour that isn’t supported.

• Good: Benefits owners (CFO/regulator side) co‑own adoption - tariff design, engagement, and measured kWh/m² improvements, reported to Finance monthly. (Victorian Auditor-General's Office)

3) Payroll modernisation

• Bad: “Go‑live” equals success; first pay run exposes brittle processes; schools become the test lab.

• Good: A business SRO owns pay accuracy and admin load; go‑live criteria include defect‑escape rate, first‑pay‑run readiness, and a staffed hypercare playbook. (NZLII)

4) Commissioning IT/OT in hospitals

• Bad: Identity/physical access treated as an IT milestone, not a commissioning outcome.

• Good: The commissioning lead is the SRO; day‑one time‑to‑access for clinicians and physical‑access safety are the acceptance criteria. (Office of the Auditor General)

What to do differently 

• Value Case → Adoption Plan → Operating Model 

Value Case: Which decision, workflow, or experience improves and how we’ll know. 

Adoption Plan: Who must do what differently, by when, and with which incentives/rituals.

Operating Model: Team, budget, cadence, and decision rights to make it stick.

• One Value Owner (SRO) + One Product Owner + One Value Slice 

The SRO sits where the outcome lives (HR/COO/CFO). The Product Owner controls the backlog. Ship a thin end‑to‑end value slice that reaches real users in 4–8 weeks; then scale slices, not promises. (GOV.UK)

• Governance as speed 

Publish guardrails (security, risk, brand) up front; replace stage gates with frequent, outcome‑based reviews.

• Metrics people fight for 

A tiny set: time‑to‑first‑value, weekly active use of the intended behaviour, and a P&L proxy (e.g., kWh/m², cost‑to‑serve, claim‑cycle time).

• Behaviour by design 

Embed the new behaviour where it lives: meeting agendas, handhelds, service blueprints, bonus plans.

Quick diagnostic (five‑minute heat check)

If you’re mostly in the left column, don’t scale. Fix the setup first.

A 90‑day play that actually works

Days 1–10: Lock the foundations

• 1‑page Value Case.

• 1‑page Adoption Plan.

• Define one value slice that proves both.

Days 11–30: Wire the team and guardrails

• Name the SRO (benefit owner) and Product Owner.

• Shift funding from project to outcome (even if it’s a pilot).

• Set and publish the non‑negotiables (security, privacy, risk).

Days 31–60: Build the first slice

• Ship something end‑to‑end that real users can touch.

• Instrument usage and outcome.

Days 61–90: Prove adoption, not just delivery

• Put the new thing into the ritual where the behaviour lives (huddle/close/stand‑up).

• Run a weekly blocker court with a sponsor who can actually say “yes.”

• Decide to scale, pivot, or stop—based on usage and value, not sunk cost.

The quiet part out loud

Most “failures” aren’t failures of intention. They’re failures of design and ownership. We design for technology rather than behaviour, for projects rather than products, for artifacts rather than outcomes.

When we start with value & adoption, and give teams the operating model and single‑point benefit ownership to deliver both, momentum returns. Customers feel it. Colleagues feel it. The P&L eventually confirms it.

If you want the simple templates we use at Angsana Consulting (Value Case & Adoption Plan Canvas, Value Slice Map, Blocker Court agenda), DM me and I’ll share them.

Footnotes, Sources and acknowledgements:

• WA OAG review of major IT projects (news summary): ABC News, “SmartRider public transport upgrade among IT projects over budget…,” notes total A$1.6b over budget and 8/10 delayed. (ABC)

• Global delivery reality: McKinsey & Oxford - 45% over budget, 7% over time, 56% less value delivered (5,400 large IT projects). (McKinsey & Company)

• Adoption gap: Gartner - only 48% of digital initiatives meet or exceed outcome targets. (Gartner)

• Feature under‑use: Pendo ~80% of features are rarely/never used. (Pendo.io)

• Optus / 000: ACMA A$12m penalty (2024) and Reuters coverage of 2025 incident linked to multiple deaths. (ACMA)

• myki (VAGO): delays, ~A$550m (+55%) budget increase, benefits compromised. (Victorian Auditor-General's Office)

• Smart meters (VAGO): benefits uncertainty; potential A$319m net cost to consumers. (Victorian Auditor-General's Office)

• Novopay (NZ): first pay run underpaid ~5,000, 15 unpaid; governance findings. (NZLII)

• Fiona Stanley IAM (WA): project terminated after ~A$6m; access not automated at opening. (ABC)

• SRO/benefits ownership guidance: Australian Digital/ICT Benefits Management Policy; UK Cabinet Office guidance on the Senior Responsible Owner role. (Digital.gov.au)