There’s a number that gets thrown around a lot in consulting circles: 70% of digital transformation projects fail. Whether or not you believe that exact figure, anyone who’s worked in business technology has seen money disappear into projects that never delivered what was promised.

The usual suspects get blamed. Poor leadership. Resistance to change. Bad vendor selection. But there’s a more fundamental issue that rarely gets discussed honestly: most digital transformation budgets are fiction from the moment they’re approved.

How Budgets Get Built (and Why They’re Wrong)

Here’s how it typically goes. A senior leader decides the business needs to “modernise.” Maybe the ERP system is 15 years old. Maybe customer complaints about the website are piling up. Maybe a competitor launched something flashy and the board is nervous.

A consulting firm gets brought in to assess the situation. They produce a document — usually with a lot of quadrants and arrows — that recommends a phased transformation costing somewhere between $2M and $10M, depending on the size of the business.

The budget gets approved based on projected ROI that includes numbers nobody can verify. “We’ll reduce operational costs by 30%.” “Customer satisfaction will increase by 25%.” These figures are educated guesses at best.

Then reality hits.

Where the Money Actually Goes

The software itself is often the smallest cost. Licensing for platforms like Salesforce, SAP, or Microsoft Dynamics is expensive, sure, but it’s predictable. The unpredictable costs are everything around it.

Integration work. Connecting new systems to existing ones always takes longer than expected. That legacy database from 2009 doesn’t have an API. The HR system uses a file format nobody’s heard of. Every integration is a custom job, and custom jobs blow budgets.

Data migration. Moving data from old systems to new ones sounds simple. It isn’t. The old data is messy, inconsistent, and full of duplicates. Cleaning it up takes months, not the two weeks that was in the project plan.

Change management. Training people on new systems costs money and time. More importantly, it costs productivity. The three months after a major system change are always slower than the three months before. Nobody budgets enough for this.

Scope creep. Once a transformation project starts, everyone wants a piece. The marketing team wants a new CMS. Finance wants automated reporting. HR wants a self-service portal. Each addition seems small in isolation but adds up to months of extra work.

A McKinsey study found that large IT projects run 45% over budget and deliver 56% less value than predicted. That’s not a rounding error. That’s a structural problem with how these projects are planned.

A More Honest Approach

So what should businesses do differently? For starters, they need to stop treating digital transformation as a single project with a start and end date. It’s ongoing. Budgeting for it like a construction project — where you pour the foundation, build the walls, and hand over the keys — doesn’t work.

Start smaller. Instead of a $5M multi-year roadmap, pick one process that’s genuinely broken and fix it. A company we spoke with recently worked with AI strategy support to identify just two workflows causing the most friction. They automated those first, proved the value, then used that momentum to fund the next phase.

Build in contingency — real contingency. Not 10%. More like 30-40% for anything involving legacy system integration. If you don’t use it, great. If you do, you won’t be scrambling for emergency funding halfway through.

Separate the technology budget from the people budget. The technology is the easy part. The hard part is getting people to actually use it. Budget separately for training, support, and the inevitable productivity dip during transition.

Measure outcomes, not activity. “We migrated 50,000 records” isn’t an outcome. “Customer response time dropped from 48 hours to 6 hours” is. Define what success looks like in terms the business actually cares about, and measure against that.

The Australian Context

Australian businesses face some specific challenges here. The talent pool for complex system integration work is smaller than in the US or UK. Bringing in overseas consultants adds cost and timezone friction. And the Australian Information Industry Association has noted that mid-market companies — the $50M to $500M range — are particularly vulnerable to over-scoped transformation projects.

The businesses getting this right aren’t spending more. They’re spending more carefully, on smaller projects with clearer goals. That’s not as exciting as a boardroom presentation about “enterprise-wide digital transformation,” but it’s a lot more likely to actually work.