Let me describe a meeting I've sat in more times than I can count. It's usually about six months after a digital project has gone live - or in some cases, about six months after it was supposed to go live. There's a managing partner or a COO at the head of the table. A marketing lead who looks exhausted. Someone from IT who's been quietly furious for weeks. And the conversation always starts the same way: So what actually happened?
Nobody really wants to answer that question. Because the honest answer is usually complicated, doesn't fit neatly into the "we picked the wrong agency" story that everyone's already agreed on in the corridor, and requires someone to say something that makes them look bad.
I want to have that conversation with you now. Not to lecture you - if you've been through a digital project that underdelivered, you already know it went wrong. But I've spent 25 years watching these projects succeed and fail across professional services, financial services, legal, consulting, and SaaS firms. And the patterns are so consistent it's almost eerie. The same five things come up again and again. Some of them are on the agency's side. Some of them - and this is the bit nobody writes about - are on yours.
That's not an accusation. It's an observation. And it's the most useful one I can offer you, because the firms that go on to succeed the second time around are almost always the ones willing to look at both sides honestly.
Before we get into the patterns, let me name something that doesn't show up on any balance sheet.
McKinsey's 2024 research puts the headline failure rate for large-scale transformations at 70%. Seven out of ten. That number hasn't meaningfully improved in a decade, which should tell you something about how poorly understood the problem still is. But the cost of a failed digital project isn't just the budget you spent. It's the scar tissue.
You know what I mean. It's the partner who now says "we tried that" every time someone raises a technology discussion. It's the operations lead who was burned so badly they'll only approve the absolute minimum next time. It's the marketing team whose confidence took a hit because they championed the project and it didn't land. And it's the board or executive committee that now treats every digital proposal with the same suspicion they'd give a cold call from a timeshare company.
That scar tissue compounds. Every month it sits there, it makes the next proposal harder to approve, the next investment harder to justify, and the gap between where you are and where you need to be a little wider. I've seen firms stuck in this loop for three, four, even five years after a failed project. Not because they can't afford to try again, but because nobody wants to be the person who signs off on round two.
So the commercial consequence isn't just what you spent. It's what you're not doing now because of what happened then.
Some delivered brilliantly. Some were quiet disasters. And the ones that failed almost always failed for one or more of these five reasons. They're not exotic. They're not surprising. That's what makes them so frustrating - they're entirely recognisable, and yet they keep happening.
This one's the silent killer. Someone in a leadership meeting says something like "we need a modern digital presence" or "we should be using AI" or "the website needs to reflect who we are now." Everyone nods. Budget gets approved. An agency gets briefed.
But nobody ever wrote down what success actually looks like in terms you could measure. No specific number. No behaviour change. No "we'll know this worked when..." statement that everyone's signed up to.
I was involved in a post-mortem for a mid-sized consulting firm a couple of years ago. They'd spent north of £300k on a website rebuild. The site looked beautiful. It worked well technically. And yet the partners considered it a failure. Why? Because the partners expected it to generate new business enquiries, the marketing team thought the goal was brand positioning, and the agency had been briefed on "a modern digital experience." Everyone was right according to their own definition of success. Everyone was disappointed.
BCG's 2024 research found that 60% of transformation failures trace back to the absence of a clear problem definition and master plan. Sixty percent. That's not a technology failure. That's a briefing failure.
Here's how it usually works. A steering committee of four or five senior people approves the investment. They meet monthly - or at least, they're supposed to. An agency is appointed. A project manager is assigned internally, usually someone who already has a full-time job doing something else.
Three months in, the steering committee meetings start getting rescheduled. The internal project manager is firefighting on other priorities. Decisions that need sign-off sit in someone's inbox for two weeks. The agency sends emails that don't get replies. And when something goes sideways - a scope question, a design disagreement, a data integration that turns out to be more complicated than anyone expected - there's nobody with the authority and the accountability to make a call.
I sat with a law firm COO last year who told me, quite candidly, that their failed CMS migration could be traced to a single moment: when the senior partner who'd championed the project went on a three-month sabbatical and nobody formally picked up the baton. The project didn't stop - it just drifted. By the time the partner came back, it had gone so far off course that the only option was to descope half the features and launch something nobody was particularly proud of. The partner, to her credit, was mortified. But by then it was too late.
Ownership doesn't mean attending meetings. It means one named person with genuine authority who is responsible for the outcome and empowered to make decisions when things get complicated. Which they always do.
This is one I feel quite strongly about, because it's so avoidable.
You commission a twelve-month digital programme. The agency presents a Gantt chart. Discovery in months one and two. Design in months three and four. Build in months five through nine. Testing in month ten. Launch in month eleven. Month twelve is "optimisation" which, let's be honest, usually means fixing whatever broke.
The problem? The organisation doesn't see any value until month eleven. For ten months, all anyone experiences is cost, disruption, and requests for feedback on wireframes. The partners who were sceptical from the start have ten months of evidence that this was a waste of money. The finance team has ten months of invoices with nothing to show for them. And by month seven or eight, the patience runs out.
I've seen this happen so many times it physically pains me. A project that would have delivered genuine value if it had run to completion, killed at 70% because the firm ran out of political capital.
The fix is boringly obvious: break the programme into phases that each deliver something visible and useful. Not just a milestone on a plan - actual working capability that people can see and touch. A rebuilt enquiry form that goes live in week six. A new service page template published in month two. Something, anything, that gives the sceptics a reason to stay on board.
You know the pattern. A project kicks off with a big announcement. Maybe an all-hands email from the managing partner. "Exciting news - we're investing in our digital future." Then silence for three months. Then another email: "Here's the new system, training is next Tuesday."
The people whose daily work is about to change - the fee earners, the business development team, the client services coordinators - were told about the project. They weren't involved in it. Nobody asked them what was broken. Nobody showed them early prototypes and asked "would this actually help you?" Nobody gave them a reason to care.
And then everyone's surprised when adoption is poor.
I remember a financial services firm - mid-sized wealth manager, about 200 people - where the new client portal had been designed entirely by the digital team and the agency. The relationship managers who'd be using it daily were consulted exactly once, in a 45-minute session that was really more of a presentation with a Q&A bolted on. When the portal launched, it technically worked perfectly. But it didn't match how the RMs actually managed their client relationships, so they kept using the old spreadsheets and email threads. Six months later, usage was at 30% and the COO was wondering why they'd bothered.
The people who need to change their behaviour need to be part of shaping the change. Not informed about it. Part of it.
We need a new CRM. We need to move to a headless CMS. We need an AI strategy.
Every one of those statements might be true. But they're all solutions. Not one of them describes the problem that needs solving.
I had a conversation with a managing partner at a 150-person consulting firm who told me they'd spent £180k implementing a particular CRM platform because a peer at another firm had recommended it at a conference. When I asked what specific business problem the CRM was supposed to solve, there was a long pause. "Well... we needed a CRM." That was it. The platform was technically excellent. The implementation was clean. And nobody used it properly because it had been configured around the vendor's default workflows rather than the firm's actual sales process.
It's like buying an expensive suit without trying it on. Looks great on the hanger. Doesn't quite fit anywhere that matters.
This happens more than you'd think. A technology gets selected - sometimes because of a conference demo, sometimes because a board member read an article, sometimes because the IT team has a preference - and then the project becomes about implementing that technology rather than solving the original problem.
Right, here's where I need to say something that most agencies won't say, and most clients don't want to hear.
The agency let us down. We had the wrong partner. If we'd chosen differently, we'd have got a better result.
Maybe. And look, there are genuinely bad agencies out there. Ones that overpromise on timelines they know are unrealistic. Ones that put the senior team in the pitch and then hand the work to juniors who are learning on your budget. Ones that go quiet when problems emerge instead of raising them early. I've seen all of that, and when it happens, it's fair to point the finger.
But in the vast majority of failed projects I've examined - and at Distinction we've been called in to unpick quite a few over the years - the failure wasn't entirely on one side. Usually it was a combination. The agency underestimated the complexity. And the client moved the brief after key decisions had already been made. The agency didn't push back hard enough on an unrealistic timeline. And the executive sponsor quietly disappeared three months in because something else became more urgent. The agency's communication was poor. And the internal IT team were hostile to the project from day one because nobody thought to involve them in the selection process.
I spoke to a COO recently who'd been through exactly this. She was very honest about it. "We blamed the agency for about eighteen months," she said. "Then I went back and looked at the project files and realised we'd changed the scope four times after the design phase, our project lead had been pulled onto another initiative halfway through, and we'd ignored two escalation emails because we were too busy." She paused. "The agency wasn't great. But we weren't great either."
That's not easy to say. And I'm not asking you to absolve a partner who genuinely let you down. But consider whether the failure was 100% on their side, or whether some of those five patterns were present on yours too.
Because if you change agency but don't change approach, you'll get a different flavour of the same result.
Here's something that took me a long time to properly articulate, even though I'd been seeing it for years.
In most failed digital projects, the technology worked as specified. The CMS did what the CMS was supposed to do. The CRM stored records correctly. The website loaded. The integrations functioned.
The project failed because the conditions required for that technology to create value weren't in place. Clear goals. Genuine ownership. Early visible wins. Honest communication. A problem-led brief. Those aren't technical requirements - they're governance requirements. And they're a leadership responsibility.
Most professional services firms have well-developed governance for their client work. Partners know how to manage client relationships, how to escalate issues, how to keep complex engagements on track. But they apply a completely different - and usually much weaker - governance model to their own internal projects. I've never quite understood why. The stakes are comparable. The complexity is often greater. And yet the rigour just... isn't there.
The shift isn't about learning project management methodology. It's about treating internal digital investment with the same seriousness you'd apply to a significant client engagement.
The Standish Group's research, supported by McKinsey's 2024 findings, shows something worth sitting with: firms that properly identify the root causes of a failed digital project and address the governance conditions that caused it succeed on the second attempt at a rate of around 72%. That's not a marginal improvement. It's a dramatic one, especially against a baseline where 70% of first attempts fail.
But the key phrase there is "properly identify the root causes." Firms that simply change agency and run the same kind of project with the same governance model? Their success rate barely improves at all.
The difference is honesty. The firms that succeed the second time are the ones who looked at those five patterns and asked "which of these were ours?" Not because they wanted to feel bad about it, but because they wanted to fix the actual problem rather than just change the name on the invoice.
I've seen this play out so many times. A firm comes to us after a failed project with another partner. And the first conversation isn't about what we'd build - it's about what happened last time. What went wrong. What they'd do differently. What governance they'd put in place. That conversation is, honestly, more valuable than any technical discovery session. Because if the conditions aren't right, even the best agency in the world - and I'm not claiming that's us, by the way - will struggle to deliver something that sticks.
There's a companion piece on how to build momentum for change in risk-averse industries that's worth reading if your firm operates in a regulated environment where some of these patterns are baked into the culture. The structural resistance is real, but it's not immovable.
If any of what I've described sounds familiar, you're not alone. And you're not stuck.
The next step isn't to commission another project. It's to understand - honestly, without the comfortable narrative - what went wrong with the last one. Which of those five patterns were present? Which were on the agency's side and which were on yours? What would need to change for the next attempt to land differently?
The next piece in this series has a diagnostic tool you can use to assess whether your current project is heading the same way - early enough to do something about it. It's called 'The five warning signs a transformation is about to stall,' and it picks up exactly where this piece leaves off.
Because the worst thing you can do right now isn't to try again. It's to not try again, for the wrong reasons.