Why B2B martech integration isn’t a tech problem: it’s a trust problem – and what marketers need to do about it

Here’s a statistic that should trouble anyone responsible for marketing technology: only around 14% of martech systems are properly integrated. Some estimates suggest that 60% of martech capability goes entirely unused. Despite years of investment, countless vendor promises, and an industry that talks constantly about connected customer journeys, integration remains one of marketing’s most persistent failures.

In the most recent episode of the ‘Trust & Influence in B2B’ podcast, I sat down with Adam Sharp and Alec Weekes Weeks from Clevertouch Consulting, a B2B martech consultancy, alongside Robert Nicholson, digital marketing director at Robert Walters, to explore why this challenge refuses to go away. The conversation drew on insights from the Mosaic martech think tank, a forum Clevertouch convenes to bring together senior practitioners and explore the practical challenges facing the industry. What emerged was a reframe that should change how we think about the problem: integration isn’t primarily a technical challenge. It’s a trust challenge that happens to involve technology.

The problem keeps getting worse

You might assume that integration would get easier over time. APIs are more open than they were a decade ago. The technical barriers have lowered. Yet the gap persists—and may actually be widening.

Adam Sharp pointed to one reason: “Every year hundreds, thousands of new systems come out, millions of new ideas, new connections, new capabilities. But businesses aren’t buying that many per annum. For every one bit of software they buy or implement, there’s another 10,000 that get added. So actually the amount of technology that businesses are procuring and integrating isn’t keeping up with the pace of content and material on the market.”

The result is that the average level of integration across the industry is declining, not improving—even as the technology to integrate becomes more sophisticated. Adam also flagged a worrying trend in utilisation: “I’ve been seeing statistics suggesting that utilisation has gone down from about 55, 58 percent to around about 40 percent at the moment. There’s a lot of flirtware, a lot of stuff just not being used.”

It’s not a technology problem – it’s a trust problem

Robert Nicholson offered the sharpest articulation of what’s really going on: “We keep on looking at it as a technology problem. When it’s not a technology problem, it’s a trust problem. It’s only going to work when people trust that the data is reliable, that the sharing of systems isn’t going to create risk, and it’s not going to undermine each person’s priorities.”

This plays out in familiar ways. The CRM team doesn’t want to share data with marketing automation because they think marketing’s data is messy. Marketing doesn’t trust the numbers coming from sales. Legal is nervous about what flows where. Everyone has competing objectives and different definitions of what good looks like.

“The nirvana,” Robert explained, “is where the head of the CRM and the head of the marketing automation team are completely aligned—they’re twins. And in reality, that is never the case. They’ve all got competing priorities and objectives. And that’s the thing that causes a lot of the friction.”

Alec Weeks reinforced the point: “If you have got a data team and IT team, security, marketing—if they’re all being very protectionist and very scared about sharing, naturally you’re not going to benefit from that share. You need to break down those barriers and just go, look, for the benefit of the business, for the experience we can offer, we need to seriously all get on the same page and stop acting like we need to be precious about these things.”

The commercial case no one makes

One reason integration keeps slipping down the priority list is that marketers struggle to articulate its value in terms the business understands. Alec was direct about this: “People go, ‘Well, imagine the journeys we could do if we were more integrated, imagine the personalisation.’ Okay, but why do I care? We’re doing fine as a business.”

His challenge to marketers: get to the specifics. “You need to almost get to the pence and pounds of going—for every pound we spend on integration in this system, this system and this system, we will see back definitively a saving of this and earning of that and a profit of the end result. That’s the only thing that the board cares about.”

Without this, shortcuts proliferate. Robert observed that budget pressures lead organisations to look at integration problems and say, “There’s some errors there, it’s too expensive, let’s not invest in updating that CRM connector.” The workarounds accumulate. The plasters multiply.

Adam put it bluntly: “Integrations have always pretty much been non-negotiable. I think we’ve just tolerated. We’ve cut enough plasters over those wounds of manual processes and everything else to get past that lack of integration.”

The vendor trust gap

The trust challenge extends beyond internal functions to the vendors marketers rely on. There’s a growing disconnect between what vendors promise—particularly around AI capabilities—and what they can actually document for compliance purposes.

Robert’s advice to marketing leaders was straightforward: “Become far more comfortable at turning around to the vendors and saying, show me. Just show me how it works. Show me a real life example.” He described catching out vendors by asking them to show his own data: “They’ve pulled up data that I know has been stolen. And so I’ve sat there and done, well, we can’t use your data.”

The vendors that earn trust, Robert noted, are those that publish clear documentation about their approaches. “Adobe has a trust center. LinkedIn has a trust center where they explain their approaches to AI. That’s made us feel far more comfortable about the ways in which we’re utilising some of these technologies.”

Adam raised a broader concern about compliance standards slipping: “What’s really weird about the proliferation of new technologies coming out, particularly out of North America, is that more and more of these impactful marketing technologies are scraping data. It’s not compliant and you can’t trust it. Suddenly we’ve forgotten about GDPR and we’re happy to use data that is questionable.”

Alec acknowledged the pace makes it almost impossible to keep up: “The honest answer is that you can’t really actually keep up because the pace that these vendors are moving at is enough to give you whiplash.” His suggestion: focus on principles rather than trying to evaluate every tool. “What we should expect from these vendors is that here are our core principles. It’s not about AI tool by AI tool. It’s saying as a business, these are our three founding principles for how we handle respect, trust and integrity.”

AI will expose the problem, not solve it

There’s a temptation to believe that AI will finally force the integration issue—that the promise of intelligent orchestration across systems will make proper integration non-negotiable. The reality is more sobering.

Robert offered the standout line of the conversation: “Unless you’ve solved a lot of those fundamental problems that we talked about at the beginning around trust—who’s got weak data, who’s got strong data, who can change this data—AI is just going to simply automate confusion.”

The logic is straightforward: AI tools need integrated, trusted data to function. Without it, they’ll accelerate existing weaknesses rather than fix them. “The organisations that haven’t invested in integration are going to feel pain,” Robert said. “Whereas the ones that have invested more, saying integration is part of our core vision to build this spine of technology—they’re going to feel less pain because they’ll already be halfway down this journey.”

Adam agreed that AI is making the need more visible rather than solving it: “I think it’s making it more obvious and more clear that those integrations have always been needed and will be even more critical as we move forward.”

And he cautioned against assuming full automation is the answer: “The thing that marketers have got to think about is level-appropriate automation, because not full automation is always a good thing. You can imagine you just set these systems up and then someone leaves the business and there’s no documentation, no control. It’d be a little bit chaotic.”

What this means for marketing leaders

The path forward isn’t more technology. It’s building the trust—between functions, in your data, and with your vendors—that makes technology actually work. As Alec put it, organisations that succeed treat integration as iterative: “It’s about businesses that don’t over-plan, don’t overbake it, but also don’t be too blasé about it. It’s finding that middle ground.”

Robert summed up the leadership challenge: “Leaders are going to have to resist the temptation to shortcut the basics.”

Practical takeaways for marketing leaders

1. Assess your internal trust landscape. Before your next integration project, map where the friction actually lies. Which teams are protective of their data? Where do competing priorities create resistance? The blockers are rarely technical—they’re usually about relationships, ownership and fear of risk.
2. Build the commercial case in pounds and pence. Quantify what poor integration costs you: analyst hours spent on manual workarounds, errors from duplicate data entry, campaigns delayed by data access issues. Then articulate what integration would free up in concrete terms the board will understand.
3. Challenge your vendors directly. Ask them to show you—not tell you—how their tools work with your data. Request documentation on their AI approaches, data handling and compliance posture. The vendors worth working with will have clear answers; the others will deflect.
4. Treat integration as iterative, not binary. You don’t need to connect everything at once. Start with a single data flow, prove the value, learn what breaks, then expand. The organisations that succeed are the ones that keep pushing through initial setbacks rather than abandoning the effort.
5. Prepare for AI to expose your gaps. If your data foundations are weak, AI tools will make that painfully obvious. Audit your integration readiness now—before you’re forced to by an AI implementation that amplifies confusion rather than reducing it.