Why digital transformation projects fail

For best results, start small and prioritize planning and communication

Digital transformation failures
  • 70% of digital transformation initiatives fail, in part because of overly ambitious scope
  • IT is a critical ally in getting automation projects off to a good start
  • Companies need to do a better job of preparing employees for the disruption these projects cause

Last year, a top European bank asked Elena Christopher to assess how a big initiative to digitize its fraud‑management process had gone off the rails.

The bank thought it did everything right. The executive team knew of the benefits and risks of the initiative. Business and IT leaders were communicating well. Line employees were thoroughly trained on new digital tools that would relieve them of arduous manual tasks.

Yet the project dragged on one, two, and eventually eight months past its scheduled six‑week timeframe. What went wrong? Christopher is a vice president at HFS Research, a firm that studies the impact of digital process automation on business. She interviewed bank executives until she had an answer: The use case the bank had targeted was just too big and complicated for a pilot project.

A better approach, says Christopher: “Start simple, establish success, and then scale.”

The bank example highlights a key tension point for businesses. There’s intense pressure to use advanced technology to improve decision‑making and accelerate the pace of work. “Enterprises today cannot effectively operate without automation and AI underpinning it,” says Chris Brahm, an analytics specialist with Bain & Company.

Yet 70% of digital transformation initiatives fail, according to McKinsey research. Automation experts say there are three common reasons for this lamentable record. The first is poor communication between IT and the rest of the business. The second is overly ambitious project scopes. The third reason is failing to prepare employees for changing workflows and responsibilities.

Here’s how to avoid these pitfalls

Given that process change is a key outcome for digital transformation projects, it’s no surprise that the most successful transformation initiatives originate from the business and not IT, according to research from EY. After all, business employees are the ones who will have to change the way the way they work.

As a result, business teams are often tempted to leave IT out of the loop. In fact, just 31% of organizations believe their IT departments are supportive of process automation efforts, according to Deloitte.

Even so, business‑led projects need to have collaboration with IT as their first step, says Ian Barkin, co‑founder of the intelligent automation consultancy Symphony Ventures. If they don’t, IT may think the business group is “going rogue,” Barkin says.

That can lead to a host of problems. When business‑side execs shut IT out, they risk alienating a function that holds important cards in any big tech initiative. For example, if a new hire joins a company in the middle of a business‑led effort to automate the employee onboarding process, IT might restrict her access to email, CRM, and other core systems.

These tensions are best resolved at the top, Barkin says. He recommends that execs from IT and the business unit stay involved at every step of a project, ensuring that their teams stay in line by providing clear guidance from above.

“I see executive mandates as very positive,” says Barkin. “The troops tend to line up and move in the same direction.”

Start small

The individual steps toward an automation goal—say, automating your year‑end financial reporting—can look easy. But companies often mistake a part for the whole.

“It’s easy to build a bot,” Christopher says. “But to build a bot that is now running a huge chunk of your finance and accounting? That’s a lot harder than a bot that fills a form on a webpage.”

In particular, organizations tend to underestimate how complex their internal processes really are. The European bank ran into that problem when it tried to automate the various processes that it used to detect and mitigate fraud. The bank tried to transform all those processes in one fell swoop.

A better approach is to tackle one subprocess at a time, then validate and iterate with additional ones. This sort of “test and learn” approach makes it easier to spot failures and mitigate them before they cause broader problems.

Mike Salvino, managing director at Carrick Capital Partners, cautions against the temptation to awe spectators with an automation idea that may be impressive, but is ultimately just a “machine learning science experiment.”

“Who remembers AlphaGo?” he said during a recent talk, referring to Google’s experiment to crown its AI as the new world champion of the game Go. “AlphaGo had no business impact…It never got integrated into an enterprise to actually do anything with.”

Prep your people

Large‑scale digital automation projects often fail because employees weren’t properly prepared for the change. Imagine you work in accounts payable. “If an IT person you’ve never seen before shows up unannounced at your desk, saying he’s going to install a system that will eliminate 30% of your job, you’re not going to get up and thank him,” says Tony Saldanha, a former VP of IT and Global Business Services at Procter & Gamble.

Just as business units need to work with IT from the beginning, IT needs to involve HR.

“Change has to be addressed as a culture, a work process, and a broader organizational transformation,” he says. “The job of every HR leader is to have a strategy for the digital workforce of the future.”