An old proverb states that it is “Better to be hurt by the truth than comforted with a lie.” A roundtable of martech experts I recently hosted proved the truth of that, validating many things I had long feared to admit to myself. Here they are: Five truths we must address to move forward more effectively.

1:B2C brands aren’t comfortable saying who their customers are

The first question I asked at the roundtable was, “Who can tell me who your best customers are?” I then looked around at participants who avoided eye contact with me, perhaps embarrassed that they couldn’t identify their best customers. 

How is this still possible? We have spent decades making enormous investments of time and resources into data, marketing technology, high-priced consultants and large teams. Yet, not one of the people working at leading brands felt comfortable saying they knew who their customers were.

Recommendation to address Truth 1: Customer analytics is paramount. Most of our investments have been in systems that activate data. The need for the requisite data is a given, but these systems – CDPs, messaging tools and journey orchestration – have focused on the decision, not the understanding. Of course, brands must continue to collect valuable data about their customers. However, it’s time for brands to invest more in their customer analytics capabilities to better understand their customers through:

• Explainable predictive analytics
• Relevant data enrichments
• Ongoing segmentation re-evaluation
• AI-enabled classification of content interactions and corresponding affinities
• And much more.

2: No one measures ROI on their marketing technology investments.

Martech vendors love to promote ROI. Treasure Data shared an 802% ROI for working with their solution. HubSpot has a handy calculator with assumptions for lift in key marketing and sales metrics. Yet, having worked with hundreds of brands on customer data-related and martech projects, ROI measurement is the rare exception and far from the rule. 

At the roundtable, not a single brand had a strong story for ROI on their customer data. One CMO said to me, “Craig, I appreciate that you’re always pushing for ROI, but I don’t see it that way. I prefer to evaluate this investment based on the capabilities it unlocks.” They were talking about a CDP recently rolled out across the entire organization.

Dig deeper: The marketing ROI problem has its roots in marketing culture

Recommendation to address Truth 2: Don’t commit to perfect measurement on your customer data and marketing technologies. Commit to outstanding MOps processes so that you have options to measure their impact. All too often, I encounter brands that can’t see their segments in their web analytics, can’t report on their customer universe and have no idea how their segments are performing. 

3: Your measurement sucks if you’re not using customer data

Signal loss in digital is real, regardless of the reprieve Chrome granted hopeful marketers whose heads were buried in the sand during the time of impending cookie apocalypse. The multi-touch attribution industry experienced a serious reckoning, which is why brands are turning to the more complex marketing mix modeling. 

Marketing mix modeling doesn’t necessarily require customer data. Oftentimes, MMM solutions use aggregated spend and exposure data across various markets to approximate media impact. It’s more reliable than MTA guesswork, but it’s still insufficient for day-to-day optimization — at least without great customer data. The best-in-class brands are looking top-down, using MMM to guide spend allocations, but optimizing spend based on its impact on their customer file. This requires great customer universe reporting – basically knowing where all their customers and users are within their journey with the brand.

Dig deeper: Why MMM makes marketers nervous — and why you should use it

Recommendation to address Truth 3: Simplify your customer data to the key inputs for journey reporting. It doesn’t have to be every detail, but it must include key details such as:

• Known / Unknown
• Registrant
• Customer / Subscriber / etc.
• Number of Purchases
• LTV
• Recency

You must understand which customers are converting from which campaigns. Good marketing operations in the recommendation for Truth 2 will help you tremendously here.

4: Brands don’t know why their data isn’t ready

Composable CDP has been all the rage for the last few years. Even so, many brands don’t think they are ready for composable. They often say, “Our data isn’t ready.” It has been my observation that brands’ data is much closer to ready than they realize. What really blocks them from adopting the simpler, composable pattern for CDPs is that their tech teams struggle to make key data available. This is usually for one or more of these reasons:

1. Not a priority to IT.
2. IT is willing to provide data, but doesn’t understand the requirements, and marketing has trouble stating them clearly or concisely.
3. Data is provided, but it is either overly simplified to control costs or is too raw, placing high demands on data literacy among non-technical teams.

Recommendation to address Truth 4: Move to a modern data stack, but do it quickly and with agility. My colleague Craig Howard advocates for the “Customer 101″ approach over the prohibitively expensive and time-consuming Customer 360.

5: Your team can’t use AI at scale until you get your data right

Everyone’s talking about scaling AI, and many of us are already using it—whether in our personal lives, inside parts of our tech stack, or even through company-wide AI policies. However, the reality is that most AI initiatives fail to deliver. That stat from MIT—that 95% of AI projects fail—gets thrown around a lot, and for good reason. A big chunk of those failures is because of messy data.

Dig deeper: Before scaling AI, fix your data foundations

I’ve seen this play out myself. We attempted to set up a basic context agent to gather information from Fireflies, SharePoint, Google Drive and Slack. The goal was simple—help new team members or consultants juggling multiple clients get up to speed faster. But we hit a wall. Different naming conventions and no standard taxonomy for client or meeting names meant the agent couldn’t make sense of it all. It had the potential to save hours of work, but without clean, consistent data, even a simple AI tool got tripped up. It turns out that you can’t scale AI until your data house is in order.

Recommendation to address Truth 5: Develop a use case-focused task force for how your organization can use AI. Follow through with tactical actions for operational protocols that will enable AI agents to make your teams’ lives easier and unlock incremental productivity.

These five truths may be uncomfortable, but they’re also clarifying. They reveal the gaps we’ve normalized—and the opportunities we’ve yet to fully seize. Martech doesn’t need more tools; it needs better practices, clearer priorities and a renewed focus on understanding the customer. Facing these realities head-on is the first step toward making your technology, data and teams actually work together. Let’s stop pretending and start fixing.