I’ve spent the past twelve months investigating, critiquing and slowly understanding the role of the cloud data warehouse for the future of Martech and Adtech. And now, some of these investigations are starting to solidify.
But one practical question has been eating away at me – do marketers want to embrace the cloud data warehouse in how they use various technologies to do marketing and advertising? In other words, should the warehouse become the home of marketing data? The question seems obscure, but as we will see, the data warehouse may be one of the most impacting technologies of modern marketing.
To help find answers, I hit up Hightouch, one of the breakout companies right at the center of this emerging trend. This company is also asking the question, so we went to talk with a variety of brand-side people leading Martech teams.
This essay is a little different because it’s a collaboration of sorts – two groups figuring out the answer to a question that has ramifications for the entire marketing technology industry, a research approach I'll be experimenting with in the future.
Hightouch’s Co-CEO Tejas Manohar, wrote his reflections here. Mine are below.
The value of abstraction
It depends on what side of marketing you’re talking about when you’re trying to figure out if there’s value in warehouse-native marketing technologies. With a broad stroke, you can say that all marketing technology relies on some kind of database for it to function. This includes categories like CRM, marketing automation, identity and authentication, analytics and BI, advertising, content management, customer data platforms, experimentation and personalization, app marketing, and mobile and SMS.
All of these categories rely on things like an email address, characteristic data like transaction records, behaviors, or things like a customer’s content preferences. But these tools also create their own kinds of data, like sales interaction events, email open and click data, or the results from AB tests.
In a way the data warehouse seems to provide the core kinds of data that are required for marketing, much of which is hard to access; but in other ways, Martech tools create their own kind of data - data that's hard for the warehouse to collect.
The vast majority of B2B SaaS that we see today relies on the cloud data warehouse because it’s more secure, scalable, and stable. It’s also incredibly cheap, and increasingly so. On a theoretical level, most marketing technologies are already running on some kind of data warehouse, so why not just cut out the double handling of storage space and run the software on a company’s existing instance of AWS or Snowflake?
There’s a reason for this – abstraction. There was once a joke that all B2B SaaS can be reduced down to a fancy excel spreadsheet with an API. And there’s some truth to this. The B2B SaaS category exists because it’s hard to build the processes and software required on the brand’s side. It’s cheaper and easier to subscribe to a platform, start sending it data and allow the marketers to go in and use WYSIWYG or drag-and-drop editors to do all the work.
But that’s changing. The data warehouse was a domain for the engineers and analytics folks – people who are well-versed in SQL, Python, and JavaScript. But there’s a new generation of tools that bring about a marketer-friendly level of abstraction, but respects that there are greater technical challenges.
The data warehouse represents a new paradigm, the Martech solutions are now the tools, the cloud warehouses the raw materials, and by breaking these into two formats you have a lot more flexibility to build the kind of marketing practice that’s right for the various types of businesses and their goals.
This is motivating marketers like Darrell Alfonso, the Director of Marketing Strategy and Operations at Indeed, to see the data warehouse as a central pivot point for making marketing technologies work today. In our interview research, Darrel speaks to the reality of creating data-driven marketing from the warehouse as a non-negotiable for cost and flexibility with modern teams:
“The lack of a single source of truth is the problem that many have come across. The first problem is disparate data, right? Data is in different places, from different sources. There is no single source of truth. So that's problem number one. And you want to solve for that by bringing the data together in one place.
The data warehouse is a great way to do this because it's cost effective. A CDP could store some of this data, too, but it's expensive to do so. The data warehouse is one of the most scalable and cost effective ways to aggregate all your sources of data, but you do need data engineers to manage it.”
We’ll always need data engineers
The reality is that for most marketing technologies, you need far more than a technically interested marketer to get value out of the tools and services. The reliance on data engineering and cloud skills is an ever-constant need for marketing automation and customer data platforms, precisely because with most other Martech you still need to store and monitor the data. Even if it’s not in your data warehouse.
Because of this, marketing departments are investing in technical skills more than ever to enable the better utilization of data in tools and platforms. According to a Gartner 2022 CMO spend and strategy survey, there are now a variety of ways that marketers are investing in data-centric functions like customer analytics, marketing insights and loyalty program management.
The reality is that as marketing becomes increasingly mature in its digital sophistication, so do investments in data tools and platforms. But it’s not the marketers that need to skill up, it’s the data engineers that enable the services to work for marketers.
Jessica Kao, the Senior Director of Marketing Operations and Analytics at F5 suggests this is the case – when it comes to building out analytics solutions on top of the data warehouse it’s always going to be an ROI equation. Is investing in a team to build and monitor a specific analysis process or a dashboard going to provide an efficiency? Or are there ways to not do it in the warehouse that saves time?
Like any technical decision, there’s always a value and effort conversation, but the warehouse breaks the existing train of thought to suggest that there’s more to gain through enabling the wider organization with data and insights from marketing. In a way the data warehouse creates flexibility not just for marketing teams, but the wider business to see the same data from a source of truth, not just another Martech tool with its own specific lens on the world:
“When you use a data lake or a data warehouse, the data is in free-form, which means you can manage your analytics however you like. You can then assemble it like a jigsaw puzzle to answer any questions you might have. For example, if you have a customer question to answer or a campaign goal to achieve, you can assemble the data in whichever way you need from a data warehouse. The flexibility is there unlike with other tools. You may have different challenges, but that's part of the data maturity curve.”
Yet even a seasoned analytics leader like Jessica recognizes that we all need a data analytics engineer to keep things working, but not because of the technical skills to manage data warehousing for insights, it’s a different viewpoint from which helps to strengthen marketing insights. Here's Jessica again:
“It is rare to have an analytics and performance marketing team sit under the same umbrella as the marketing operations team. But I think that is actually our competitive advantage. Because if you think about it, the team that owns the MarTech stack is also part of the source of data, right? We know what we want to do with the data. With my two teams, MarTech and analytics, under one roof, when a stakeholder has a question, the data analyst and MarTech analyst working together can more effectively help the stakeholder get to the answers they need.”
This is why I’m seeing an increased focus on shared-use platforms for reverse ETL and composable CDP. While the tooling for making the data available for activation across channels is managed by data engineering teams, marketers can use no-code tools to design, segment, and send data to downstream apps.
What I like about this approach, more than anything is it doesn’t pretend that marketers can take care of the entire pipeline of data storage, ETL, enrichment and activation; it respects that marketers and technical folks will need to co-exist, and they can probably co-exist better together. If you're investing in a CDP or marketing automation platform you’re going to need data engineers and skilled analytics folks to set up and use the tools; why not extend that to the data warehouse instead?
Data is an experience problem
If there’s already a theoretical argument for Martech to embrace the data warehouse driven by greater technical sophistication in the four walls of brands, and the reality that most Martech requires data engineers anyway, there’s a strong reason that embracing the data warehouse is an important way to create better customer experiences.
For example, if you don’t have your product interaction data in Salesforce Marketing Cloud, it’s going to be hard to trigger communications based on the next interaction you want a customer to take. But it will also be hard to suppress them from remarketing audiences. It’s a similar challenge for identifying churned customers – what are the characteristics that lead to a customer leaving a brand?
You’d only know these things with a complete picture of the product utilization, transactional data, and their behaviors across email, web, and app. At best this kind of data can be stored and managed in a CDP, but then you’d need to stream a huge volume of data from a data warehouse anyway in most situations.
It’s for this reason that Agness Allagh, the former Head of Marketing Operations and Martech at Meta, argues for better availability of data to inform and enable the lifecycle of a customer:
“When marketers can access insights that inform the customers' unique interactions across the customer journey, they can speak directly and more effectively to customer needs at each lifecycle stage.”
The value of the data warehouse in this case is the ability to pull from all kinds of data to create the experience based on a marketer’s own insights, strategy and unique business situation, not just the use cases that a vendor might give them.
This is why sophistication around the data warehouse in Martech teams is going to be a differentiator for companies that compete primarily on experience. Streaming data from a warehouse is more than just having access to a larger pool of data resources; it’s a way to unlock more experiences.
Recent research from Learning Experience Alliance, a marketing and sales education platform, also supports this idea citing a lack of integration and challenges to accessing metrics as the main barriers to making marketing technology investments valuable in companies.
The data warehouse plays a role here, but it’s still obfuscated by a lack of integrations, data quality, technical skills, and coordination to fully access customer data across the customer journey. But there’s a way to build for scalability and flexibility, and there’s a way to build for your team’s level of skills and knowledge. Like most things in life, you’ll do better if you rise to a higher level of mastery, not fall to a tolerable level of incompetence.
The forcing function: Advertising
The other adjacent trend here for marketers is that there’s increasing activity across the AdTech sector leveraging the data warehouse for first-party data as third-party tracking goes away.
The most recently announced move in this direction was LiveRamp’s partnership for identity enrichment and activation with Snowflake, following on from The Trade Desk’s and Lotame’s announcements of their own flavor of CDP. There’s a big push for first-party data as the replacement for targeted advertising.
And as the utilization of data warehouses like Snowflake goes up, so do venture capital dollars go into reverse ETL startups that are attempting to bridge the gap between warehouse and Martech.
I’ve argued at length with the day customers stop sharing data that despite the shift to first-party data being a potential mistake for the advertising industry, all the capital is flowing there as more and more announcements are made about AdTech platforms, data warehouses, and customer data platforms moving into first-party data for advertising.
The corollary is not just replacing third-party cookies with something like the trade desk’s UID 2.0 or alternative solutions, it’s also represented in the shift towards retail media – a behavioral and transactional well of consumer data, along with major walled gardens such as Apple and AWS, both of which are building AdTech solutions.
If anything, the crushing of the third-party cookie is the biggest trigger point for marketers, and by extension their agencies and AdTech partners to embrace the warehouse. There’s just so much latent opportunity in the product, transactional and preferential zero-party data stored in the data warehouse - advertisers can’t resist!
The Achilles Heel
The one challenge I have with more enterprises embracing the data warehouse for marketing use cases is the political situation of most companies. The divide between marketing and data teams is deeper than just the difference between them using PowerPoint and Excel – it’s a cultural and values divide.
If marketers can see an opportunity in accessing data from a data warehouse, then it’s often the case that data teams will see risks and challenges. Conversely, data teams can be great at building things that marketers don’t need. The political alignment in most organizations can feel impossible to some as Tejas Manohar suggests in what he calls the “Gordian Knot”:
“The people that tend to drive these initiatives are marketers - not technical experts. That's not an indictment of marketers; instead, it means that there are ways of doing things outside their domain knowledge that could help them succeed but that they simply don't know. The result is marketers embracing approaches for untying the knot that have repeatedly failed in other contexts.”
Clearly, there’s a language and education challenge to overcome if marketers are to embrace the data warehouse for Martech. But there is something shifting here; marketers don’t need to learn SQL to understand the possibilities of using the warehouse as a rich source of customer data, but they will need to learn how to make small talk with an engineer on their hundredth data query request.
So the answer to the question is that yes, it seems as though marketers are increasingly wanting to understand the data warehouse, but the technology to make it work for marketing is still nascent, ways of collaborating between data and marketing are underdeveloped, and there’s so much data that Martech tools create that it represents new challenges altogether.
But if there’s anything that stops progress in this space, it’s the Achilles heel of misalignment between teams.
The modern infrastructure for sending, analyzing, and enriching data from the warehouse is impacting a variety of marketing departments. If it’s to send email, build advertising audiences, power a customer journey or analyze marketing effectiveness, there’s a home for marketers with the data warehouse, but like most new technological movements the problems are always the people.
