Data is at the heart of everything, but we’re still struggling to handle it

As part of the launch of MarTech, we are exploring key trends and tactics related to our six main topics. Today we look at an emerging trend in marketing Data. Click here to learn more about our mission at MarTech.

“Our relationship with data is not healthy.”

That’s the sentence which jumped out at me from Christal Bemont’s foreword to a report published earlier this month, the 2021 Data Health Survey from Talend, the data fabric and analytics platform, where Bemont serves as CEO.

How long have we been in a data-driven business environment? Remember the term “big data”? That goes back at least 10 years. If we don’t hear it used much today, I think it’s because we’ve become accustomed to data being big, and very fast too. But how are marketing teams handling that? All the time, I speak with brand marketers who haven’t yet got their data unified and under control. When I spoke to Tom Matzzie, President and CEO of CleanChoice Energy earlier this year, he was proud that his stack let him respond to customers in real-time, but admitted to having an ongoing data lake project in the background. He’s far from alone.

Some companies are still working on internal data lakes; some have abandoned that route and invested in CDPs as their single source of truth; and some are still patching things together with data from marketing automation and CRM systems. Pat Maigler, marketing strategy and operations manager at Williams-Sonoma, gave a forceful defense of the latter approach on an edition of MarTech Live.

Marketing, sales are least data-driven teams

This is all just anecdotal, perhaps, but the statistics unearthed by Talend seem to support it. 64% of executives work with data every day, but 78% report that they face challenges using data effectively. Only 40% trust the data they are basing decisions on, and 22% (a sizeable segment) don’t think their company’s investment in data management is worth it.

Perhaps most striking, though, is that sales and marketing emerge as the least data-driven departments, with almost 50% of executives in those functions saying they make most of their decisions without consulting data. That’s not the way this story was supposed to go.

I reached out to Chris Penn, co-founder and Chief Data Scientist at TrustInsights.ai, to see if he was surveying the same bleak landscape. How are marketers actually resolving the challenge of storing, managing and activating data?

Brands are not doing a good job with data

“Hey Kim,” he said, “the short answer is, they’re not. Or they’re not doing a very good job of it, at least.” If CDPs were meant to become the gold standard for data management and activation, it’s not clear that that’s actually happening. “The CDP trend seems to have peaked,” said Penn, “at least in terms of visible search interest.”

What he’s hearing from his clients is that identity resolution is “the biggest, thorniest issue, especially when it comes to unifying named entities. It turns out that identifying and resolving named entities in the wild is really difficult and the results aren’t great; companies have lots of customer data, but resolving that data with public entities (social media handles, etc.) is problematic at best.”

Legislative action on privacy is adding to the problem, ensuring what Penn called “a nice stew of incomplete data.” He expects the emerging trend of reliance on deterministic first party data to continue. “The gold standard of identity resolution is, and will continue to be for the foreseeable future, to ask customers to provide the data themselves in exchange for overwhelming value to them.”

The view from the big marketing suite

If any one of the big players in the CX and marketing space is closely identified with data, it’s Oracle. I asked Nate Skinner, Global SVP of Marketing for Oracle Advertising and CX, if he could tell me what’s going on. He started with Oracle’s own CDP, Oracle CX Unity. “We approach it more like using Unity to create segments, not as a data source, because your data is going to be everywhere. Some of our customers use CRM data that isn’t our own; some use Eloqua marketing data that is our own; and they sometimes also use finance data from an ERP.”

For example, a campaign targeting an offer to people who paid bills on time for twelve months requires data which could live in many different sources. “What our customers are doing with Unity is taking it as the segment builder.” The multi-source segments can then be loaded into Eloqua to launch a campaign. In short, this particular CDP is not being used as a single-source of truth, but, as Skinner put it, as an “interstitial.”

Does that mean it’s not really a CDP? “We had a lot of discussions at Oracle about whether Unity was a CDP or not, because a traditional CDP is a space, like you put all your stuff in there. Unity can be used that way, but the more likely use case is that you leave the data where it is.” If there’s a plateauing of interest in independent CDPs, it’s precisely because of the need to take data from all those other sources to create a persistent, unified customer database.

“You and I know that in the real world, once I’ve done the extract, transform and load, the data source I took it from has already changed. So I have to do it again, and the next thing you know you have a team of people, that’s all they’re doing every day, trying to keep the data in the CDP as the system of record. This is not a race you ever win. Data cleanliness and hygiene is not a destination where we all arrive and say, we’re done.”

From this perspective, siloed data is something we can live with, as long as there are tools to access that data and activate it. It almost turns the CDP Institute’s definition on its head: a CDP “is packaged software that creates a persistent, unified customer database that is accessible to other systems.” It was time to talk to the Institute’s founder, David Raab.

Just making do with what you’ve got doesn’t work

As it turned out, Raab had a flexible view on what might do the work of the CDP. “There’s a version of the CDP,” he said, “that sits on top of an existing data lake. That’s cool that you’ve got that data in one place, all we’re going to be is an access layer. We’re going to have the interfaces to all your marketing tools, and we’re probably going to have some identity management if it’s not part of your data lake. So there’s a level of functionality, but you don’t necessarily have to pull all the data into a separate data store – which is by definition what a CDP does.” See above. “If you think about functions, and don’t get hung up on labels, then you can get the same functionality. It’s never black and white.”

And that was before I mentioned Oracle Unity. “Oracle probably has the clearest mental picture — because they’re Oracle, they understand about data. So they’ve always drawn the picture correctly. Originally they had BlueKai, their DMP, where a CDP should have been. They realized it couldn’t handle it, so they created CX Unity, which has a technically proper kind of approach to it.”

Even the most ambitious CDP implementation, Raab admits, doesn’t pull in literally all the data. “You’re never going to replicate your inventory in real-time on all your products. There was always going to be some federated access to external systems as part of the CDP architecture. Nothing’s ever as simple as the pictures on the slides.”

What about those marketers who protest that they can make their existing systems work; that a CDP is over-reach. “A lot of companies are still trying to make do without a CDP; they’re trying to use their CRM or their marketing automation, or their DMP, and they just have to find out the hard way that it doesn’t work. That’s why people buy CDPs. It wasn’t a category that a vendor invented and just pushed onto the market, it was the marketers that had a problem and wanted something which would fix it for them.”

But if that’s the case, why are people telling me that interest in CDPs isn’t still soaring? Raab wasn’t persuaded that that’s the case. “The data we see on CDP penetration keeps going up, although the term CDP is so broadly applied that it makes it hard to look at any survey data and make sense of it. But we talk to the vendors, and almost all of them survived the pandemic, and those who survived have been doing quite well. If there’s a slowdown, it hasn’t really affected the vendors I talk to. We see a lot of interest outside the core markets; Africa is popping up, Eastern Europe too; so we do see continued interest and growth globally. What’s happening I think is there’s more interest from IT departments in building their own.”

The “build it or buy it” debate isn’t over, Raab said. “You can build it in Snowflake, but it’s not as easy as you think,” he cautioned. “There are a lot of tools which, if the IT department really did its homework, it could find the right tools to do it efficiently. Are there successful internal projects? I would hope so; but people come to see me only when their internal project didn’t work.”

Raab summed it up. “The independent vendors saw the demand, and the big vendors finally came in when they could no longer ignore it. The moral of the story is that the existing tools don’t work.”

So a CDP, whether deployed as an access and activation layer along CX Unity, or in the traditional sense of a unified database accessed by activation systems, has to be the solution. That’s what I was hearing, but I had one more conversation on my list — with a martech executive in the midst of a CDP implementation.

Dig deeper: Customer Data Platforms – a Marketer’s Guide

“It’s been literally three months”

M.H. Lines is CEO at Stack Moxie which provides automated monitoring across the tech stack to give RevOps teams the confidence that systems are working — and alerts them if they aren’t. The business’s need for a CDP emerged from the manner in which prospects were discovering them. “Martech people don’t want to talk to sales,” said Lines — with confidence, as she’s a martech person herself. The Stack Moxie product has a completely free tier, so interested users can experiment with it, without the burden of interacting with sales. As Lines puts it, “they can take a journey into our product without having to go through the dreadful product demo.”

When prospects get stuck on the journey, however, she needs them to be passed to the customer success team to address any issues they might have, or be put into segments for email re-targeting. Collecting first-party data is one thing, but Lines wants to be able to respond when some relevant piece of behavioral data can be appended to that. “The product discussion is driving it for us,” she says, “and the CDP should be the single source of truth.”

Previously, they had been trying to manipulate this data using Tableau, the data visualization software. The CDP she has now selected is, in her view, one of the best CDPs on the market. “We love it,” she says.” Problem solved? Not yet. “I know what want the reports to look like,” she told me, “but we haven’t used it yet, and it’s been literally three months.” It needs to integrate with Amplitude, the product analytics platform which tracks interactions between users and products, although Lines emphasizes that Stack Moxie is essentially a start-up, “without a crazy tech stack.”

Reflecting on the CDP category, she said: “When I was working early in CRM, we competed against Salesforce. We believed that everyone wanted the flexibility to do things their own way, we thought that Salesforce was too opinionated about the way a sales process should be run. We were wrong. CDP today feels like CRM before Salesforce. The platforms are incredibly powerful and can do anything, but it would be great to start with something. CDPs should have an opinion about what everyone needs to start with.”

Lines has an implementation consultant on the product, and has used external engineers in addition to her own team. Can she see the finish line? “I think we’re close. We just hired a bunch more engineers, so our cycle times are going faster.” She’s says they’re “cusping. That perfect moment when it will finally add value.”

Never ending journey?

To me, it felt like I had gone around in a circle. First I heard that there were brakes on CDP adoption, and that people were making a viable patchwork out of data in the systems they already had. From one of the big vendors, I heard that
— although they offered a fully functioning CDP — customers weren’t using it as a unified database but as a layer over their existing databases. The CDP Institute agreed this was one way to proceed, but told me that falling back on marketing automation and CRM no longer worked. But the path from selecting a CDP to implementing it and realizing value may be too fraught, especially for smaller, fast-moving organizations.

But I’m going to take the optimistic view, and say that this isn’t a circular story, but one of cautious advances. Yes, you need to be able to access and activate valuable customer data; but maybe you don’t have to have it all in one place. And if the traditional CDP is the correct route, you need to be realistic about how easy and fast it will be to implement and show ROI.

This is the journey marketers are on, and I’m not sure it’s the kind of journey which has an end. Perhaps it’s enough to be moving forward.