Here's a number that should bother you: roughly 70% of digital transformation initiatives fail to meet their objectives. You've probably seen that stat before - it gets wheeled out at every conference and dropped into every consulting pitch deck. But what rarely gets examined is why they fail. Because when you dig into the post-mortems, the pattern is remarkably consistent. Almost never the platform.
I've been involved in enough platform recovery projects over the past 25 years to have a fairly strong opinion on this. And it's one that tends to land awkwardly during procurement processes: the implementation partner you choose will have a bigger impact on your outcome than the platform you select. By a wide margin.
But the platform is the long-term decision. The partner is just the delivery vehicle. We can always change agency.
I hear this constantly. And on the surface, it's logical. The platform is what you'll live with for five to seven years. The partner engagement might last six months. Why wouldn't you spend more time on the thing with the longer lifespan?
Because the partner shapes everything the platform becomes. The information architecture, the editorial workflows, the integration patterns, the governance habits, the training quality - all of that gets determined during implementation. And those patterns persist long after the partner has invoiced their final milestone. You're not choosing a delivery vehicle. You're choosing the architect of your operating reality for the next half-decade.
Let me describe something specific, because I think you'll recognise it.
A mid-sized professional services firm, let's say 200 fee earners across three offices, spends four to six months evaluating CMS platforms. They run demos, build comparison spreadsheets, attend vendor events, maybe commission an analyst report. Thorough, diligent work. By the time they've selected a platform, the team feels confident they've made a well-informed decision.
Then they spend about six weeks choosing an implementation partner. They issue an RFP, get three or four proposals back, sit through a couple of presentations, compare the pricing, check a few references, and pick one. Often, and I say this without any pleasure, the deciding factor is price.
I was in exactly this situation with a firm about eighteen months ago. They'd chosen Umbraco. Solid choice, genuinely - well-suited to their editorial team's needs and their integration landscape. But the implementation partner they'd selected had quoted 30% below the next cheapest option, and the discovery phase had consisted of a single half-day workshop before the proposal landed. When I asked what the sprint structure looked like, the partner contact said, "Oh, we're agile - we'll work it out as we go." Which is not agile. That's just winging it with better vocabulary.
Six months later, the platform was live but barely functional. The content model didn't match how the editorial team actually worked - they'd structured content types around what was easy to build, not around how the team needed to publish. The Salesforce integration was dropping data intermittently because it had been tested against clean sample records, never against the messy, inconsistent data the firm's CRM actually contained. (Field mapping issues, mostly. The kind of thing you only discover when real users start pushing real data through.) The training consisted of a two-hour video walkthrough that covered every feature of the CMS except the four things the marketing team needed to do every day.
The firm's conclusion? "Umbraco doesn't work for us."
Umbraco was fine. It was doing exactly what it was designed to do. The implementation had failed the platform, the firm, and the editorial team.
This isn't an isolated example. We've recovered implementations where the client was genuinely ready to abandon a perfectly capable platform - they'd just never seen it implemented properly. One firm had spent £180k on a Kentico build that their marketing team couldn't use without raising a support ticket. Not because Kentico is hard to use. Because nobody had spent any time understanding how the marketing team actually worked before designing the content management interface.
So if I'm arguing that partner quality matters more than platform choice, I should probably be specific about what quality looks like. Two things matter more than anything else.
The first is process discipline. And I don't mean "we're agile" - I mean a specific, describable delivery process with defined sprint structures, decision protocols, and escalation paths. If a partner can't tell you exactly how they run a project before they've started it, they're making it up as they go. That might work out. It usually doesn't.
The second is honesty about trade-offs. This might actually be the more important one. The partner who says "yes, we can do that" to everything in your brief isn't being responsive - they're being agreeable. And agreeable partners deliver consequences. The partner who says "you could do that, but here's the risk, and here's an alternative that gets you 80% of the outcome at 40% of the cost" - that's the one you want in the room when things get complicated. Because things always get complicated.
The other dimensions matter too - technical depth in the specific platform, sector experience, communication quality. But process discipline and honesty are the ones that are hardest to fake and most predictive of what the engagement will actually feel like on week eight when the integration is behind schedule and a key decision needs escalating.
A portfolio shows a partner's best work in controlled conditions. A pitch shows their communication capability when they're motivated to impress. Neither is particularly predictive of what the engagement will actually feel like when things go sideways.
So what does predict quality?
Ask for delivery team references, not sales references. Specifically, have conversations with clients who worked with the people who would be on your project. And ask the uncomfortable questions: what went wrong, and how did the partner handle it? Every project has moments where things don't go to plan. The quality of a partner isn't measured by the absence of problems, it's measured by how they respond when problems arrive.
Watch how they handle disagreement during the evaluation. This is a counterintuitive one, but I've come to rely on it. If you say something in a pitch meeting that a partner thinks is wrong - maybe your timeline is unrealistic, or your content migration approach is risky - do they tell you? Or do they nod along? The partner who pushes back during the sales process is showing you exactly what they'll do when the project hits a wall. The one who agrees to everything is showing you that too.
And meet the delivery team. Not the sales director. Not the practice lead who'll check in monthly. The project manager, the lead developer, the UX designer - whoever is going to be in your Slack channel and on your weekly calls. The bait-and-switch, where senior people win the work and junior people deliver it, is one of the most consistent complaints I hear about agency relationships. Entirely preventable with a direct question.
I want to be fair here, because I'm not arguing that platform choice is irrelevant.
There are real scenarios where it's the critical factor - when the partner is equally strong across the platforms you're evaluating (rare, but it happens), when the platform's technical architecture has a specific fit or conflict with your existing technology environment, or when your long-term digital strategy requires capabilities that only some platforms provide at the right cost point. Outside those scenarios? Partner quality is the dominant variable.
The conventional approach is: shortlist platforms first, then find a partner to implement whichever one you've chosen. It makes intuitive sense. Decide what you're building with, then find someone to build it.
I'd suggest reversing that sequence, or at least running them in parallel.
Start with a clear problem statement and success criteria. Not a platform wish list - a description of what you need the outcome to achieve, what your editorial team's workflow looks like, what your integration landscape requires, and what your budget and timeline allow. Then select two or three implementation partners whose quality passes the evaluation criteria above. Ask each of them to recommend a platform for your specific requirements. Their recommendation will be informed by which platforms they can implement well for firms like yours, in your regulatory context, with your integration needs. That's more useful information than a feature comparison grid, because it reflects implementation reality rather than theoretical capability.
I realise this feels backwards. It challenges a procurement instinct that runs deep - the idea that you should know what you're buying before you choose who'll supply it. And if your firm's procurement process requires a platform decision before a partner selection, I get it. Not every organisation has the flexibility to flip the sequence. But even within a conventional process, you can weight the partner evaluation more heavily than you probably are now. If you're currently spending six months on platform selection and six weeks on partner evaluation, even shifting that to four months and two months would change the outcome.
The platform decision is reversible. In three to five years, you'll replatform anyway - that's just the lifecycle. But the patterns established during implementation - the governance habits, the editorial workflows, the integration architecture, the training and capability your team builds or doesn't build - those compound. They shape how your organisation thinks about digital. They determine whether the next platform decision is an upgrade or another recovery project.
Getting the partner right matters more than getting the platform perfect. And right now, most firms have that weighting exactly backwards.
If you're in the middle of a platform selection and want an independent view on the partner options you're considering, book a partner evaluation conversation with us. Sometimes a 30-minute call saves six months of regret.